Variations in Gas Content in Organic Matter-Rich Low Maturity Shale; Example from the New Albany Shale in the Illinois Basin
Corresponding author: Dr. Maria Mastalerz, Indiana Geological Survey, Indiana University, Bloomington, IN 47405-2208, USA, Tel: 812-855-9416, Email: email@example.com
The New Albany Shale (NAS) in the Illinois Basin is a Middle and Upper Devonian to Lower Mississippian unit, correlative with the Antrim Shale of the Michigan Basin and the Ohio Shale of the Appalachian Basin. These shale units are thought to be part of an epicontinental succession deposited in response to a sea-level rise over large areas of the North American craton [1,2]. In Indiana, the NAS is underlain unconformably by Middle Devonian North Vernon Limestone and is lithostratigraphically subdivided into the following six member units in succession from the oldest to the youngest: the Blocher Member, the Selmier Member, the Morgan Trail Member, the Camp Run Member, the Clegg Creek Member, and the Ellsworth Member . In most of Indiana, the NAS is overlain by the Rockford Limestone (Figure 1a).
The NAS is an organic-matter-rich formation, with organic matter content often exceeding 10% and representing dominantly oil window maturity . The presence of gas in the NAS sparked continued industrial interest since the late 1800s. Gas production from the NAS commenced in Harrison County, Indiana, in 1885 [4-6]. Over time, limited early production in Harrison County expanded into other areas of Indiana. Initial production test rates (IPs) of wells typically ranged from 567 to 11,327 m3 of gas per day (20 to 400 Mcf/day), although some wells in northern Daviess and southern Sullivan Counties are believed to have produced more than 28,316 m3/day (1 million ft3/day, MMcf/day). Gas-in-place estimates of the Illinois Basin NAS gas re erves vary from 86 to 160 Tcf  from which technically recoverable resources are estimated at between 1.3 to 8.1 Tcf (mean 3.8 Tcf; ).
With regards to origin, the NAS gas appears to fall into three general gas play categories, (i) thermogenic, (ii) mixed thermogenic- microbial, and (iii) microbial [9,10]. The most prominent gas production so far has come from a dominantly microbial (i.e., biogenic) gas play at shallow depth in Indiana. Strąpoć et al.  analyzed petrography, gas desorption, geochemistry, and micro- and mesoporosity of the NAS in the eastern part of the Illinois Basin at two locations in Indiana, in Owen and Pike Counties. The gas content in these locations was primarily dependent on the total organic carbon (TOC) content and the micropore volume of shales. The depth of burial, thermal history, and the salinity of formation water dictated the origin of the gas-in-place (i.e., microbial, thermogenic, or mixed).
Figure 1a. Lithostratigraphy of the New Albany Shale in Indiana. After Strąpoć et al. (2010) and references therein.
This paper further investigates controls on gas content in the NAS, specifically focusing on the influence of organic matter content and porosity characteristics on the desorbed and residual gas. The main purpose of the study is to understand and predict the fraction of gas that can be easily extracted (desorbed gas) from gas that is very difficult to impossible to extract without opening additional pores (residual gas). The location studied is in Davies County, Indiana, where the drilling intersected the entire NAS thickness (Figure 1b).
Of the 23 shale samples originally selected for canister desorption, 12 were selected to closely examine the relationship between gas content, porosity characteristics, and mineralogical composition. The samples are from one borehole locat ed in Daviess County, Indiana; this well penetrated the entire thickness of the NAS, and the selected samples represent all the members of the NAS, from the oldest, Blocher, at the bottom part of the section to the youngest, Ellsworth, in the topmost sample (Table 1).
Canister desorption measurements were initiated in the field using a volumetric displacement apparatus and standard gas desorption measurement procedure described in “A Guide to Determining Coalbed Gas Content” (Gas Research Institute, 1995). Briefly, the volume of gas that displaced fluid in the displacement apparatus (called “desorbed gas”) was recorded first at field temperature, and later, after the samples were transported to the laboratory at the Indiana Geological Survey, at room temperature (~21oC). The initial gas measurements were taken every 10 minutes for 1 hour, followed by 1-hour intervals for 7 hours; after that, gas readings were taken every day.
Recording of the desorbed gas was discontinued after several measurements did not record any gas. After gas desorption was complete, canisters containing samples were weighed, and the weight of the shale was obtained by subtracting the empty canister weight. Lost gas content, desorbed gas, and total gas contents were calculated using Desorption Data Analysis Software of the Gas Research Institute (1995).
After canister desorption was complete, residual gas content was determined on a split of ~15 g of each sample. The sample, after being placed in a container with a septum, was ball milled for 5 minutes. The gas volumes were measured on a small volumetric apparatus. In addition to the analysis of the representative sample, from each canister two samples were selected, one of a relatively dark color and one of a relatively light color, to see variations within the same canister. To test the reproducibility of the technique, two residual gas determinations were done on each sample. The same samples that were used
for residual gas determinations were analyzed for TOC using ELTRA instrumentation.
Dark and light varieties of each canister sample were crushed to 60 mesh (250 μm) and analyzed for surface area and mesopore
characteristics using low-pressure nitrogen adsorption; micropore characteristics were tested using low-pressure carbon dioxide adsorption. Surface area, mesopore- and micropore characteristics were determined on an ASAP 2020 porosimeter. More details about these techniques are available in our previous publications [11,12]. The pore classification into micropores (<2 nm), mesopores (2–50 nm), and macropores
Figure 1b. Location of the well in Daviess County, Indiana.
Table 1. Stratigraphic unit, depth, vitrinite reflectance (Ro), total organic carbon (T0C), total sulfur (S) of the shale samples studied.
(>50 nm) used in this study follows the classification system of the International Union of Pure and Applied Chemistry .
Total porosity was measured on splits of samples of approximately 7 mesh (particle size ~2.8 mm) size and calculated by subtracting the grain volume from the sample volume. The former was determined on a helium pycnometer ULTRAPYC 1200e,whereas GeoPyc 1360 was used to determine the sample volume.
The mineralogy of selected samples was identified using X-ray diffraction (XRD) techniques. The XRD analyses with 2θ 2–70° were carried out on nine bulk powdered samples (canister numbers 17, 19, 21, 25, 29, 33, 34, 35, 40) (Table 2). Selected samples were also analyzed using SEM techniques on carbon-coated polished blocks to determine chemical composition of minerals. Specifically, eight polished blocks representing dark varieties of canister samples (17, 19, 25, 29, 35, and 40) and light varieties (33, 34) were analyzed with SEM at acettepe University and the General Directorate of Mineral Research and Exploration.
Organic Petrographic Characteristics
Random vitrinite reflectance Ro values of the samples studied are around 0.67 % (Table 1), placing the source rock into the early mature zone but close to the mid-mature transition. Amorphinite and alginite macerals of the liptinite group dominate the organic petrological composition of the samples throughout the section. Alginite occurs both as small bodies and large well-preserved algae spores such as Leiosphaeridia and Tasmanites. Alginite fluorescent color is golden yellow, but greenish-yellow and dull yellow alginite is also present. Contributions of terrigenous vitrinite and inertinite macerals are low, rarely exceeding 5 % of the total volume of organic matter. Solid bitumen is present in all the samples. Organic petrographic observations (e.g., abundance of marine algae) suggest a predominantly marine source of organic matter in the NAS, which is consistent with the previous studies .
TOC ranges from 1.22 to 15.61 % for the lighter color samples and from 4.94 to 15.23 % for the darker color variety (Table 1). For the majority of samples, TOC of the lighter varieties is very similar to those of the darker ones (Figure 2A), but five samples (25, 29, 33, 34, and 35) had significant differences between the lighter and the darker counterparts, indicating internal canister sample heterogeneity. Total sulfur contents range from 1.1 to 3.66 % for the light-colored samples and 0.95 to 4.61 % for the darker samples. Similar to TOC, the majority of the light and dark samples had similar S contents (Table 1).Collectively, the C/S values for the lighter and darker varieties are very similar; only two samples, the shallowest and the deepest, showed significant differences. All but one of the C/S ratios are below 4; most are within the 2 to 4 range (Table 1), suggesting conditions close to normal marine [14,15].
Figure 2. A – Histograms of total organic carbon (TOC) in lighter and darker shale; B – Residual gas content in lighter and darker shale.
There is a distinct positive correlation between TOC and S, with only the shallowest sample (can 17) differing (Figure 3). With this sample omitted, the coefficient of determination (R2) is 0.87 for the dark samples and 0.62 for the lighter samples (Figure 3). This positive correlation with a near-zero sulfur intercept also suggests that the sediments were deposited under normal marine conditions . Recently Alrowaie (2015) documented a positive correlation between C and S for the NAS in Pike County, Indiana, but showed that this relationship varied between different members, which would explain the scatter observed in this study, which includes samples from all members.
Mineralogically, the shales are dominated by quartz (30.5–51.5 %), muscovite (illite, 31.1–48.6 %), and albite (2.2–13.0 %), with other minerals varying within minimal ranges (Table 2). There is a weak positive correlation between quartz and TOC (R2=0.45, Figure 4A), suggesting a biogenic source of some silica .There is also a negative correlation between muscovite (illite) and TOC (R2 =0.65, Figure 4B).
Figure 3. Scatter plot showing relationships between TOC and S content for lighter and darker shale. Sample 17 (can 17) shows unusually high TOC for a relatively low S content.
No correlation between mineralogical composition and residual gas content was observed. Brittleness index calculated as the ratio of (quartz + dolomite)/(quartz + dolomite + calcite + clay + TOC)  varies within a range of 38 to 53 (Table 2). Brittleness index serves as an indicator of rock affinity to break under stress of hydraulic fracturing, with high values being desirable for unconventional systems. For example, the best-producing intervals of the Devonian Barnett Shale consist of ~45 wt % quartz and as low as 27 wt % clays . The brittleness index range for the shales studied is comparable to other shales considered suitable for hydraulic fracturing, for example, the Cretaceous Second White Specks Formation .
Shale-forming minerals identified by XRD in the samples were also analyzed with SEM to determine chemistry of minerals and their distribution (Figures. 5 and 6). SEM work indicates that some siderites in the samples are Mg siderites with Ca and Mn traces, and some dolomites have light-color rims of Fe-bearing dolomite (or ankerite) composition.
Figure 4. Scatter plots showing relationship between TOC and (A) quartz content and (B) illite content.
Figures 5 and 6 present examples of mineral associations in clayrich samples 33 and 17 and quartz-rich sample 19. Quartz-rich (51.5%, Table 2) sample 19 shows associations of silica with illite, K-feldspar, Na-feldspar, dolomite, pyrite, and organic matter (Figure 5A, B). A clay-rich (48.6 % illite, Table 2) sample 33 shows association of clays with silica, K-feldspar, in addition to framboidal pyrite (Figure 5C, D). Micron-sized minerals such as sphalerite, sphalerite with Fe traces, Ti-oxide with Fe traces, rare-earth phosphate (monazite), fluor-apatite, zircon, and pentlandite were also identified in some samples.
Table 2. Mineralogical composition (weight %) of the samples studied. BI = brittleness index.
Figure 5. SEM images of quartz-rich sample can 19 (A and B) and clay-rich sample can 33 (C and D). Si = silica, KF = K-Feldspar, NaF = Na-Feldspar, An = ankerite, Al = illite, Fe = pirite, OM = organic matter, Sd = siderite, D1 = dolomite, D2 = dolomite with Fe traces, D3 = dolomite with Fe and Mn traces, D4 = dolomite with Fe traces and clay minerals.
Total porosity of the individual samples studied varied from close to zero to 10.3 %, and expectedly for the majority of samples was higher for light than for the dark variety (Table 3). On average, this corresponds to a range of 0.0137 cm3/g to 0.0426 cm3/g pore volume. As determined by N2 adsorption, ET surface area (averaged between light and dark samples) ranges from 4.1 to 9.1 m2/g, and BJH mesopore volume (determined on the adsorption branch) ranges from 0.0128to 0.0243 cm3/g (Table 4). On average, lighter samples havea larger BET surface area and BJH mesopore volume, with anaverage mesopore size of 12.4 nm.
As determined by CO2 adsorption, the micropore surface area ranges from 8.6 to 21.3 m2/g, the monolayer capacity from 1.9to 4.7 cm3/g, and the micropore volume from 0.008 to 0.0197 cm3/g, with an average micropore size of 1.1 nm (Table 5). Incontrast to mesopore characteristics, micropore surface area, monolayer capacity, and micropore volume on average are larger in the darker samples.
The ranges of surface area and mesopore and micropore volumes obtained in this study are comparable to those recordedearlier for the NAS in Pike and Owen Counties, Indiana . There is a strong positive correlation between micropore volume and TOC (Figure. 7A) and a negative correlation between mesopore volume and TOC (Figure. 7B).
The total gas content (on as-received basis) varies from 18.8 to 67.8 scf/ton (0.6 to 2.1 cm3/g) (Table 6). The desorbed gas content ranges from 11.3 to 56 scf/ton (0.4 to 1.8 cm3/g), whereas the residual gas content from 3.13 to 24.33 scf/ton (0.1 to 0.8 cm3/g). These values are comparable to gas content from other locations of the NAS in Indiana .
Figure 6. Secondary electron images and elemental maps indicating minerals identified in sample can 17 (dark variety). Zn = sphalerite, Fe = pyrite, Si = silica, Al = illite, P = apatite. Organic matter (OM) including sulfur traces occurs as black areas. Note that the sample is rich in illite.
Figure 7. Scatter plots and regressions showing relationships between TOC and (A) micropore volume and (B) mesopore volume.
Organic matter content is the main parameter controlling total gas content and desorbed gas content. As TOC increases from close to 3 % to more than 15 %, total gas content increases from 20 scf/ton to close to 70 scf/ton (0.62–2.2 cm3/g, Figure8), and the desorbed gas content increases from close to10 scf/ton to almost 60 scf/ton (0.31–1.86 cm3/g). The coefficientsof determination for these relationships are 0.77 and0.78, respectively. These relationships suggest that the majorportion of the gas resides in pores within the organic matter.
Desorbed gas shows positive correlations with both micropore surface area and micropore volume (Figure 9A, B), suggesting that micropores (<2 nm) are favorable adsorption and storagesites for desorbed gas. Stronger correlation with micropore volume than micropore surface area points to a volume filling of micropores rather than monolayer adsorption as the mainstorage mechanism [20, 21].
Table 4. Surface area and mesopore characteristics determined by low-pressure nitrogen adsorption. BET – Brunauer-Emmett-Teller, BJH – Barrett-Joyner-Halenda.
In contrast, desorbed gas shows a negative correlation with BET surface area and mesopore volumes (Figure 10A, B), indicating
that mesopores (2–50 nm pores) are ineffective sites for storing desorbed gas.
Residual gas is also strongly related to TOC, showing a positive correlation both for lighter and darker shales, with R2 above 0.70 (Figure 11). The only sample that is outside of this trend, having high TOC and low residual gas content, is the uppermost
sample 17. It is possible that gas from this sample migrated out to the overlying rocks.
Figure 9. Scatter plots and regressions showing relationships betweendesorbed gas content and (A) micropore volume and (B) micropore surface area.
Table 5. Micropore characteristics determined by low-pressure carbon dioxide adsorption. D-R Dubinin-Radushkevich, D-A – Dubinin-Astakhov
Table 6. Desorbed, residual, and total gas content of the samples studied.
Residual gas content shows a good positive correlation with micropore volume (Figure 12A) and micropore surface area(Figure 12B). In contrast to micropores, residual gas shows a negative relationship with mesopore volume and BEF surface area measured by nitrogen (Figure 13A, B). This clearly demonstrates that micropores are the host of residual gas. These micropores are likely closed, which prevents liberation of this gas during canister desorption; crushing the samples prior to residual gas measurements opens up these pores.
Comparison of residual gas contents between lighter and darker varieties emphasizes that it can vary significantly in different portions of the sample (Figure 2B) and that these differences are greater where there are greater differences in TOC (Figure 2A). This reinforces the crucial role of organic matter in residual gas storage. Differences in residual gas contents between two splits of the same shale fragments were very small, indicating that the differences between lighter and darker shales of the same canistered sample are real and not related to
Figure 10. Scatter plots and regressions showing relationships between desorbed gas and (A) BET surface area and (B) mesopore volume.
Figure 11. Scatter plots and regressions showing relationship between residual gas content and TOC. Sample 17 (can 17) show s unusually low gas content for its TOC.
technique reproducibility. These differences also indicate that, for precise residual gas measurements, it is crucial to have a representative canister sample to avoid internal inhomogeneity.
In contrast to the positive or negative relationships between gas content and micropore or mesopore characteristics, the total porosity does not show a relationship with residual gascontent, and there is a trend of decreasing desorbed gas contentwith increase in total porosity (Figure 14A). This negativetrend results from the fact that mesopores (Figure14B) and not micropores (Figure 14C) dominantly contribute to thetotal porosity and that gas is hosted in micropores (Figure 9)and not in larger pores (Figure 10). Total porosity also shows anegative trend with TOC (Figure 14D).
Figure 12. Scatter plots and regressions showing relationships between residual gas content and (A) micropore volume and (B) micropore surface area.
All of these observations suggest that organic matter content (expressed here by TOC) is the main host of the gas, supporting
our previous observations of the NAS from Pike and ClayCounties. Strąpoć et al.  showed that the relationship betweenTOC and gas content, although distinct, was not universal and varied between locations, suggesting that gas contentis dependent also on the maturity of the organic matter and the contribution of thermogenic gas to the dominantly biogenicsystem.
Our TOC versus gas content data from Daviess County (with Ro 0.67 %, Table 1) are closer to those from the Pike County location (Ro 0.68–0.72 %) than from the Clay County location (Ro 0.49–0.57 %) (Figure 15. Although the character of the relationship between gas content and TOC changes between locations, the overall relationship for the NAS samples is strong,
and TOC can be used to predict gas content of this formation. Combining the information provided by  and the current data, relationship between gas content and TOC can be expressed by a power equation where gas content (scf/t) = 5.5863xTOC0.9441, with R2 = 0.80 (Figure 15).
In addition to the TOC/gas content relationship, our results suggest that micropores present in organic matter are the main sites for gas storage. The majority of these micropores are likely connected as gas is liberated during canister desorption. However, residual gas held by closed pores can only be liberated after crushing. It is likely that the proportion of closed micropores to connected micropores influences the proportion of the residual gas to the desorbed gas in shale. Forthe samples studied, a ratio of desorbed to residual gas ranges widely from 1.1 to 7.4 (Table 6), suggesting also a wide range of interconnectivity of micropores, from less than 15 % to almost 50 %. Small-angle neutron scattering (SANS) data on a NAS sample suggested that about 30 % of the pores within 2.5 to 4 nm diameter range were connected , which is within the range suggested by the ratio of desorbed to residual gas. It has been demonstrated for coals that organic matter can have a very wide range of accessible versus inaccessible pores [22, 23], which can also explain large variations in residual gas contents.
Figure 13. Scatter plots and regressions showing relationships between residual gas and (A) BET surface area and (B) mesopore volume.
Figure 14. Scatter plots and regressions showing relationships between total porosity and (A) desorbed and residual gas content, (B) mesopore volume, (C) micropore volume, and (D) TOC.
The main conclusions from this study are as follows:
1) There is a large mineralogical variability throughout the section of the NAS. Variations in quartz content of 30.5 to 51.1 % and carbonate content from zero to more than 10% result in a variable but generally high brittleness index, suggesting these shales are suitable for hydraulic fracturing.
2) Organic matter composition showing the dominance of marine organic matter (alginite and amorphous organic matter) and rare terrestrial organic matter (vitrinite and inertinite) indicates a typical Type II kerogen sequence, similar to other locations in Indiana. A positive correlation between TOC and S contents suggests close to normal marine depositional conditions. An early mature stage is also consistent with the previous maturity determinations.
3) Porosity of the shales varies, and this variability is expressed in total porosity as well as in micropore and mesopore characteristics. Total porosity ranges from less than 3 to ~10 %, corresponding to the total pore volume of 0.0137 to 0.042 6
cm3/g. Average BET surface area of 5.5 m2/g and BJH mesopore volume of 0.01626 cm3/g, and average micropore volume of 0.01296 cm3/g and the ranges of these parameters are consistent with these types of data from previously studied NAS locations.
4) Micropores are the major sites of gas storage in shales having high organic matter content. They are the major sites both for desorbed gas and residual gas. Larger pores are not favorable sites for gas storage. An increase in mesopore (pores 2–50 nm in diameter) volume results in a decreases in gas content. Pores larger than 50 nm were not measured, but they are likely uncommon, as demonstrated for the NAS samples in previous studies . In addition, a negative relationship or no relationship
between total porosity and gas content indicates an insignificant role of larger pores in gas storage for these shales.
5) Organic matter is the main host of gas-filled micropores, as indicated by a strong correlation between TOC and microporevolume and between TOC and desorbed and residual gas contents.
6) The contribution of residual versus desorbed gas to the total gas content varies in the samples studied; we suggest that it is dependent on the proportion of closed versus open micropores. This suggestion, however, requires further testing. If confirmed, the proportion of open micropores within the total micropore volume could help to predict the proportion of desorbed
(extractable) versus residual (nonextractable gas).
This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division under Award Number DE-SC0006978 (formerly DE-FG02- 11ER16246). We thank the Riverside UMBONO Company for access to the samples. We thank Cortland Eble, University of Kentucky, and the Indiana Geological Survey Publication Review Committee for valuable suggestions to improve the paper
2. de Witt WJ. Roen JB. Wallace LG. Stratigraphy of Devonian black shales and associated rocks in the Appalachian Basin in: Roen J.B. Kepferle R.C. (Eds.) Petroleum geology of the Devonian and Mississippian Black Shale of eastern North America. United States Geological Survey Bulletin. 1909, B1-B57.
5. Hamilton-Smith T, Hasenmueller NR, Boberg WS, Smidchens Z, Frankie WT. Gas production in: Hasenmueller N.R. Comer J.B. (Eds.) Gas Potential of the New Albany Shale (Devonian and Mississippian) in the Illinois Basin. Gas Research Institute GRI- 00/0068 Illinois Basin Studies. 1994, 2: 23–40.
6. Partin T. Demand for high price spurs a new era of drilling for new Albany Shale gas wells in Harrison County Indiana in: Schieber J. Lazar R. (Eds.) Devonian black shales of the Eastern U.S.–new insights into sedimentology and stratigraphy from the subsurface and outcrops in the Illinois and Appalachian basins: Field Guide for the 2004 Annual Field Conference of the Great Lake Section of SEPM Indiana Geological Survey Open- File Study. 2004, 4(5): 68–76.
9.Martini AM, Walter LM, McIntosh JC. Identification of microbial and thermogenic gas components from Upper Devonian black shale cores Illinois and Michigan basins. AAPG Bulletin. 2008, 92(3): 327-339.
10. Strąpoć D, Mastalerz M, Schimmelmann A, Drobniak A, Hasenmueller NR. Geochemical constraints on the origin and volume of gas of the New Albany Shale (Devonian-Mississippian) in the eastern part of the Illinois Basin. AAPG Bulletin. 2010, 94: 1713-1740.
12. Mastalerz M, Schimmelmann A, Drobniak A, Chen Y. Porosity of Devonian/Mississippian New Albany Shale across a maturation gradient–insights from organic petrology gas adsorption and mercury intrusion. AAPG Bulletin. 2013, 97(10): 1621-1643.
13. Orr C. Surface area measurement in: Kolthoff I.M. Elving F.H. Stross (Eds.) Treatise on analytical chemistry: Part III: Analytical chemistry in industry. New York John Wiley and Sons. 1977, 4: 21-358.
15. Adegoke AK, Abdullah W.H, Hakimi MH, Yandoka BMS. Geochemical characterisation of Fika Formation in the Chad (Bomu) basin northeastern Nigeria: Implications for depositional environment and tectonic setting. Applied Geochemistry. 2014, 43: 1-12.
19. BridgeFurmann A, Mastalerz M, Bish D, Schimmelmann A, Pedersen PK. in press. Porosity and pore-size distribution in shales from the Belle Fourche and Second White Specks Formations in Alberta Canada. AAPG Bulletin.
21. Jaroniec M, Choma J. Theory of gas adsorption on structurally heterogeneous solids and its implication for characterizing activated carbons in: Thrower P.A. (Ed.) Chemistry and physics of carbon. New York Marcel Dekker. 1989, 22: 197-243.