The Mount Simon Sandstone and Eau Claire Formation represent a potential reservoir-caprock system for wastewater disposal, geologic CO2 storage, and compressed air energy storage (CAES) in the Midwestern United States. A primary concern to site performance is heterogeneity in rock properties that could lead to nonideal injectivity and distribution of injected fluids (e.g., poor sweep efficiency). Using core samples from the Dallas Center domal structure, Iowa, we investigate pore characteristics that govern flow properties of major lithofacies of these formations. Methods include gas porosimetry and permeametry, mercury intrusion porosimetry, thin section petrography, and X-ray diffraction. The lithofacies exhibit highly variable intraformational and interformational distributions of pore throat and body sizes. Based on pore-throat size, there are four distinct sample groups. Micropore-throat–dominated samples are from the Eau Claire Formation, whereas the macropore-dominated, mesopore-dominated, and uniform-dominated samples are from the Mount Simon Sandstone. Complex paragenesis governs the high degree of pore and pore-throat size heterogeneity, due to an interplay of precipitation, nonuniform compaction, and later dissolution of cements. The cement dissolution event probably accounts for much of the current porosity in the unit. Mercury intrusion porosimetry data demonstrate that the heterogeneous nature of the pore networks in the Mount Simon Sandstone results in a greater than normal opportunity for reservoir capillary trapping of nonwetting fluids, as quantified by CO2 and air column heights that vary over three orders of magnitude, which should be taken into account when assessing the potential of the reservoir-caprock system for waste disposal (CO2 or produced water) and resource storage (natural gas and compressed air). Our study quantitatively demonstrates the significant impact of millimeter-scale to micron-scale porosity heterogeneity on flow and transport in reservoir sandstones.
The Cambrian Mount Simon Sandstone, and its stratigraphic equivalents, occur throughout the Midwestern United States, where it is a target injection horizon for wastewater disposal, geologic CO2 storage, and compressed air energy storage (CAES) operations (Medina and Rupp, 2012; Heath et al., 2013). The Mount Simon Sandstone is overlain by the upper Cambrian Eau Claire Formation, a generally low-permeability mudstone and regional seal (Neufelder et al., 2012; Lahann et al., 2014). Of primary importance are reservoir-caprock properties that govern multiphase flow, because injectivity, sweep efficiency, and capillary trapping greatly affect site performance. Geologic controls on heterogeneity of pore structure and flow properties can be complex and difficult to characterize due to the interplay of textures from primary depositional environments and a variety of potential postdepositional processes, including precipitation-dissolution with a range of textures, mechanical compaction, pressure solution, and fracture porosity or mineralization (Hoholick et al., 1984; Bowen et al., 2011). Despite the regional extent and storage potential of the Mount Simon–Eau Claire system, few conventional cores are available that cut the entire thickness of the Mount Simon Sandstone and the Eau Claire–Mount Simon contact (Bowen et al., 2011), and thus detailed studies of the major lithofacies and their associated multiphase flow properties have been heretofore limited.
In this study we use mercury intrusion porosimetry (MIP) to quantify pore size distribution and evaluate the impact of porosity heterogeneity on multiphase flow. MIP is a readily available tool for examining capillarity and pore structure of porous media. The technique has been applied extensively for assessing sealing capacity of caprock in order to understand hydrocarbon traps (Almon et al., 2005). It is also applied, but less commonly, to the study of sandstones and siltstones, including argillaceous sandstones and tight-gas sandstones. Wardlaw and Cassan (1979) examined 27 samples of coarse-grained siltstone to medium-grained sandstone from a wide variety of units of variable ages and a wide range of permeability; they examined relationships among pore-throat aperture, grain size, pore size, and mercury recovery efficiency, which is a proxy for residual trapping. They noted that the presence of carbonate cement increases heterogeneity in pore-throat aperture along with a decreased recovery efficiency. Wardlaw and Cassan (1979) also found that high mercury recovery efficiency is found in samples that have high porosity, small pore to throat size ratios, and small mean particle sizes, the latter being somewhat counterintuitive. Pittman (1992) used an unpublished industry data set of 196 sandstone samples for which conventional porosity and permeability data as well as mercury intrusion data were available; he used these data to develop empirical relationships to allow determination of pore-aperture size parameters. Nelson (2009) provided a summary of a number of MIP studies on sandstone. In this study we build upon this prior work by using MIP data to evaluate the potential impact of millimeter-scale to micron-scale pore-aperture heterogeneity on multiphase flow in reservoir sandstone.
We examined a conventional core from the Dallas Center domal structure (area of Dallas Center, Iowa) that was obtained for site evaluation of a planned CAES project (Heath et al., 2013; Dewers et al., 2014). This continuously cut core includes both a complete section of the Mount Simon Sandstone (∼28.3 m) as well as a portion (∼36.1 m) of the overlying Eau Claire Formation, including the reservoir-caprock interface. Using gas porosimetry and permeametry, MIP, thin section petrography, and X-ray diffraction (XRD), we characterize pore types and rock textures and quantify the range of pore-throat sizes in major lithofacies of the Mount Simon Sandstone and Eau Claire Formation. These data then allow us to make direct comparisons of pore characteristics of the reservoir-caprock system. The large number of MIP samples (n = 30) and detailed petrographic observations provide a unique data set for characterizing the nature and origin of the pore-size heterogeneity within a sandstone reservoir (saline aquifer) and across and into the overlying caprock. Because the petrophysical characteristics and sealing capacity of the Eau Claire Formation caprock were discussed in detail elsewhere (Neufelder et al., 2012; Lahann et al., 2014; Swift et al., 2014), we mainly provide detailed analysis for the Mount Simon Sandstone.
Regional and site-specific geologic information are presented to facilitate comparison between properties of the Mount Simon Sandstone and the overlying Eau Claire Formation at the Dallas Center structure (Fig. 1), where the core was collected, with other locations in the Midwestern United States. For example, Decatur, Illinois, is the site of the Illinois Basin–Decatur Project by the Midwest Geological Sequestration Consortium, where CO2 is injected at a rate of 1000 t/day in the Mount Simon Sandstone (Finley, 2014). The Cambrian Mount Simon Sandstone, a major regional aquifer, extends broadly throughout the Midwestern United States and overlies Precambrian crystalline basement or sedimentary rocks in Iowa, Illinois, Indiana, Ohio, Michigan, Minnesota, Wisconsin, and Nebraska, reaching thicknesses of >305 m in eastern Iowa, Illinois, Indiana, and Michigan (Anderson, 1998), and >800 m in north-central Illinois (Medina et al., 2011). In Ohio, several lithofacies (as many as eight) have been identified and interpreted as indicating deposition as a tidally influenced transgressive barrier sequence, with environments including swash and surf zones; mud, sand, and mixed flats; sand flats to tidal channels; tidal inlet channels, and bioturbated sand flats (Saeed and Evans, 2012). Lithologies include conglomerate to pebbly-to-fine sandstone, siltstone, mudstone, and heterolithic sandstone-mudstone (Saeed and Evans, 2012). Bowen et al. (2011), focusing on the Illinois Basin, stated that depositional environments may include shallow-marine, deltaic, fluvial, eolian, and possibly sabkha settings, with lithofacies including cobble conglomerate, stratified conglomerate, poorly to well-sorted sandstone, and interstratified sandstone and shale; and shale. The Mount Simon Sandstone thus exhibits strong heterogeneity in lithofacies (Bowen et al., 2011); such heterogeneity motivates an understanding of intraformational variability in flow properties on site-specific to regional scales.
Overlying the Mount Simon Sandstone for much of its regional extent is the Cambrian Eau Claire Formation. The formation varies, north to south, from siliciclastic dominated in its type area near Eau Claire in western Wisconsin to mixed siliciclastic-carbonate to carbonate dominated in northeastern Missouri (McKay, 1988; Lahann et al., 2014). McKay’s (1988) descriptions of three (upper, middle, and lower) informal units of the Eau Claire Formation in eastern Iowa include an upper portion of northeastern Iowa with finely and coarsely interlayered feldspathic and glauconitic sandstone and shale that contain inarticulate brachiopod shells and trilobite and hyolithid molds. In southeastern Iowa, the sandstone-shale facies grades into a variably glauconitic, dolomitic, feldspathic siltstone-shale facies, with sparse inarticulate brachiopod shells and fragments of trilobite molds. The middle Eau Claire Formation, from northeastern to southeastern Iowa, grades from siltstone-shale facies to a shale and dolomite facies. The dolomite facies includes skeletal dolomite packstones to grainstones, bioturbated dolomites, dense crystalline dolomite, and dolomite algal thrombolites and stromatolites. The lower Eau Claire is a fine sandstone and shale facies in eastern Iowa. Lithologies range from laminated to ripple laminated to bioturbated shaly sandstone and interlayered sandstone and shale, shale, flat-pebble conglomerate, and dolomite echinoderm packstones. In summary, the Eau Claire Formation has significant large-scale facies changes from siliciclastic to carbonate, and includes a large variety of lithologies (Neufelder et al., 2012).
Our study site, the Dallas Center domal structure, Iowa, is near the southeastern edge of the Iowa horst of the Mid-Continent Rift System, next to the Thurman-Redfield structural zone, where domal structures occur due to deformation near the horst and the structural zone (Anderson and McKay, 1989; Heath et al., 2013). The Dallas Center structure is 9.6 km west of the Redfield gas storage field. The Hydrodynamics Group oversaw seismic mapping and coring of the Keith No. 1, Mortimer No. 1, and Mortimer No. 3 wells to confirm and update previous structural interpretation of seismic reflection data collected by Bay Geophysical, Inc. to support assessment of a planned CAES facility (Heath et al., 2013). We used the core from the Keith No. 1 well for this study.
Using electric log and core attributes, Dewers et al. (2014) subdivided the Mount Simon Sandstone in the Keith No. 1 well into upper, middle, and lower units following the approach of Bowen et al. (2011) and Saeed and Evans (2012). Dewers et al. (2014) noted gamma ray and density-porosity signatures strikingly similar to those reported by Saeed and Evans (2012) for a well in Ohio, and concluded that this probably reflects a similar depositional setting of transgressive sheet sands overlying basement highs. Dewers et al. (2014) indicated that the upper unit in the Keith No. 1 well is probably equivalent to the middle unit of Barnes et al. (2009) and Bowen et al. (2011) in the Michigan and Illinois Basins, respectively, and lacks the B-cap unit of those authors. The middle and lower units of Keith No. 1 from Dewers et al. (2014) likely correspond to the lower unit of Bowen et al. (2011).
Heath et al. (2013) and Dewers et al. (2014) described the lithofacies present in the Mount Simon Sandstone in the Keith No. 1 well. The upper unit consists principally of fine-grained quartzose sandstone that is locally fossiliferous and glauconitic. Planar bedding, small-scale trough cross-stratification, and bioturbation are present locally. The middle unit is heterolithic, consisting dominantly of subarkosic fine- to coarse-grained sandstone, with interbeds of mudstone and conglomerate. The lower unit consists mainly of quartzose fine- to coarse-grained sandstone, which is locally pebbly. Clasts of the underlying Precambrian red clastic sandstone are present just above the nonconformity.
MATERIALS AND METHODS
To understand pore-scale characteristics that control heterogeneity in flow properties of the Mount Simon Sandstone and Eau Claire Formation, we used a combination of gas porosimetry and permeametry, MIP, XRD, and standard petrographic techniques. Samples were taken from the Keith No. 1 well core at depths of the major lithologic units of the Mount Simon Sandstone and several lithofacies of the Eau Claire Formation (Fig. 2; also see Heath et al., 2013; Dewers et al., 2014). Core plugs included those of vertical and horizontal orientation for directional gas permeametry and MIP. Gas porosimetry and permeametry analyses were performed on dried samples by methods detailed in Heath et al. (2013). A set of samples of mudstones or low-permeability lithologies were analyzed by TerraTek, a Schlumberger company, using their tight rock analysis method to determine pressure- and/or pulse-decay permeability and porosity. Poro-Technology in Sugar Land, Texas (now Micromeritics), performed MIP on a Micromeritics AutoPore IV 9500 V1.07 instrument. MIP was done on 28 cylindrical core plugs, ∼1.91 cm long × 1.91 cm diameter, that were jacketed with epoxy for directional mercury intrusion (Table 1). The epoxy jacket coated the outside of the cylindrical samples, but not the tops and bottoms. Thus, mercury intruded from the top and bottom of the plugs during the measurements. Omnidirectional measurements were performed on two samples that were too thin for epoxy jacketing (Table 1). Closure pressure was performed using a compressibility method (see Heath et al., 2013, for full details). Conversion of intrusion pressure to pore-throat diameter used the Washburn equation (version of the Young-Laplace equation; Washburn, 1921; Dullien, 1992). Permeability calculated from MIP data was based on Swanson’s (1981) equation. Samples were analyzed for drainage, but not imbibition.
Mercury breakthrough pressure, which represents the pressure when a continuous connected path of mercury has been attained across the sample, was estimated for each MIP sample (using methods after Dewhurst et al., 2002). Conversion from the mercury-air-rock system to the groundwater-air-rock, groundwater-CO2-rock, and gas column heights held by capillary forces (the sealing capacity) followed procedures of Ingram et al. (1997). Density of pure nitrogen as a proxy for injected air for CAES was determined from Span et al. (2000). We used the value of interfacial tension (IFT) for mercury-air of 484 mN/m and 140° for contact angle (Shafer and Neasham, 2000; Dewhurst et al., 2002). We used an IFT value of 67 mN/m (Wiegand and Franck, 1994) for water-air and a contact angle of 0° for the groundwater-air-rock system. Carbon dioxide and groundwater densities were estimated using TOUGH2 software (Pruess et al., 2012), with geothermal gradient and groundwater density gradients for the sample depths of 0.025 C°/m and 0.009792 MPa/m, respectively. Interfacial tension values for CO2-groundwater were estimated following the algorithm of Heath et al. (2012). The groundwater-CO2-rock contact angle range of 17°–70° was based on the general contact angle ranges of mineral phases found in the samples for which contact angle data are available; the phases included quartz, calcite, feldspar, and mica (Table 2), and the contact angle data was from Iglauer et al. (2015).
The X-ray diffraction samples were taken at 28 depths, near the locations of MIP and thin section samples; X-ray diffraction was performed with a Siemens model D500 θ-2θ powder diffractometer (see Heath et al., 2013). Thin section samples were selected after hand-sample observations of the core for major lithofacies. Thin sections were impregnated with epoxy and cut and ground to ∼30 μm thickness (see Heath et al., 2013). We performed optical petrography with a Leitz Wetzlar Orthoplan-Pol polarizing microscope and a Lecia DFC 425 digital camera. Intergranular volumes were determined for selected regions of thin sections using ImageJ (Ferreira and Rasband, 2012) with two-dimensional binarized images of selected regions. Binarized images were created by thresholding and manual area selection.
Here we describe the mineral composition, porosity, permeability heterogeneity, and diagenetic history of the Mount Simon Sandstone, with an emphasis on events that had the greatest impact on porosity and permeability. We then present MIP data, which provide a quantitative assessment of the pore-throat size distribution in the Mount Simon Sandstone. We show that the current range in pore-throat distributions is largely the result of complex diagenetic modification of the original primary porosity of the unit. We also present XRD, porosity, permeability, and MIP results of the Eau Claire Formation to allow for comparison; however, detailed paragenetic observation of the Eau Claire Formation is beyond the scope of this paper.
Variability in Mineral Composition, Porosity, and Permeability
Mineral phases of the Eau Claire Formation determined by XRD (Heath et al., 2013) include quartz, feldspar (microcline and orthoclase), dolomite, calcite, illite, kaolinite, and minor hematite, hydroxyapatite, and anatase (Table 2; Fig. 2). Dolomite can be predominant, reaching values as high as 89 wt%. The Mount Simon Sandstone mineralogy is dominated by quartz, with lesser amounts of feldspar (microcline and orthoclase) and illite, and typically minor hematite, gypsum, hydroxyapatite and anatase; however, hematite and gypsum can attain values as high as 16 wt% and 17 wt%, respectively, at certain depths (Table 2; Fig. 2). Permeability within the Mount Simon Sandstone of the sampled lithofacies of the Keith No. 1 core varies over five orders of magnitude, testifying to dramatic heterogeneity in pore size and connectivity (Figs. 2 and 3). A pronounced transition to more variable permeability occurs at depths >∼899.16 m in the Mount Simon Sandstone, and corresponds to a lithofacies change from a quartzarenite to arkose and subarkose sandstones with depth (Fig. 2). Unlike the Mount Simon Sandstone in some portions of the Illinois Basin (Medina et al., 2011), porosity and permeability display no systematic trends with depth and are not correlated with one another (Figs. 2 and 3).
Paragenesis of the Mount Simon Sandstone
We detail porosity-altering diagenetic changes that have affected the Mount Simon Sandstone, as well as evidence supporting our inferred paragenetic sequence (Fig. 4). This work is based upon analysis of 16 thin sections, representing samples from the upper, middle, and lower units (Fig. 2). This is a general paragenetic sequence for the entire unit; individual samples do not contain evidence for all the alterations. Prior work on the paragenesis of these phases in the Illinois Basin is also noted, in part to establish whether the Mount Simon Sandstone at the Dallas Center structure has an alteration history similar to elsewhere.
Quartz overgrowths occur in all samples from the Mount Simon Sandstone portion of the core. The presence of quartz overgrowths within barite and gypsum indicates that the overgrowths predate these phases (Fig. 5A, 5B). Some overgrowths (and gypsum) fill pores in which the adjacent grains have sutured contacts (Fig. 5C). This suggests that some quartz and gypsum may have formed after pressure solution of quartz. In the Illinois Basin, the origin of quartz overgrowths has received considerable attention (Duffin et al., 1989; Fishman, 1997; Chen et al., 2001; Pollington et al., 2011; Bowen et al., 2011). Most previous workers concluded that the quartz overgrowths formed at elevated temperatures (Fishman, 1997; 100–130 °C; Pollington et al., 2011; 40–124 °C). Bowen et al. (2011) noted that in the Illinois Basin overgrowths are present in both relatively shallow and deep samples, with no depth-related pattern to their occurrence; they speculated that either the conditions suitable for quartz cementation existed in shallow as well as deep settings (e.g., hydrothermal fluids at shallow depths) or that overgrowths formed at greater depths and subsequently underwent uplift.
Feldspar overgrowths are very common at all stratigraphic levels in the Mount Simon Sandstone. Because quartz overgrowths terminate against feldspar overgrowths, the feldspar overgrowths most likely formed first (Fig. 5B). The presence of feldspar overgrowths in gypsum-cemented regions indicates that some overgrowths formed prior to gypsum (Fig. 5D). The paragenesis, geochemistry, and environment of formation of the feldspar overgrowths have received considerable attention in studies of the Mount Simon Sandstone in the Illinois Basin (Fishman, 1997; Duffin et al., 1989; Chen et al., 2001). Radiometric dating of the feldspar indicates that it may have begun to precipitate in the Late Ordovician and continued into the Devonian (Duffin et al. 1989; Liu et al., 2003).
Natural fractures are relatively common in the Mount Simon Sandstone at Dallas Center: they are present at numerous depths in the core and occur in 25% of our thin sections. Most of the fractures are filled by gypsum, anhydrite, and barite cements. In some cases the cements that fill the fractures are confined to the fractures and do not penetrate the surrounding sandstone (Fig. 6A), despite the porous nature of the sandstone. This suggests that a soluble cement, such as halite or another evaporite mineral, was present filling pores in the sandstone at the time of fracture filling and was subsequently dissolved. Fractures were also described in the Mount Simon Sandstone of the Illinois Basin by Chentnik (2012). In addition to throughgoing fractures (i.e., fractures are not confined to single grains), intragranular fractures are also common in the Mount Simon Sandstone (Makowitz and Milliken, 2003).
Compaction and Pressure Solution
The Mount Simon Sandstone has undergone variable amounts of mechanical and chemical compaction. The intergranular volume of sandstones provides a measure of the degree of compaction that the rock has undergone (e.g., Paxton et al., 2002). Freshly deposited sand typically has intergranular volumes (IGVs) of >40% (Atkins and McBride, 1992). The highest IGVs that we have noted in the Mount Simon Sandstone are ∼30%, indicating that significant compaction has occurred. Additional evidence for compaction includes common long, concavo-convex, and sutured grain contacts as well as stylolites (Figs. 5C, 6B, and 6C). Makowitz and Milliken (2003) studied the compaction of the Mount Simon Sandstone in the Illinois Basin and concluded that it accounts for the majority of porosity loss in the unit.
The degree of compaction is exceptionally heterogeneous, both among and within samples (e.g., Figs. 7–9). Heterogeneity in IGVs can also been seen in the data for the Mount Simon Sandstone from the Illinois Basin provided by Makowitz and Milliken (2003). Such heterogeneity is relatively uncommon in sandstones (McBride, 2012). Possible explanations for the highly variable IGVs include the following.
1. Differences in the degree of sorting can cause variation in primary IGV, with more poorly sorted regions having lower IGVs. Although some of the heterogeneity may be explained to some degree by this mechanism (Fig. 7), significant variations in IGV occur in regions that are texturally similar (Figs. 8 and 9). Thus, this explanation cannot explain most of the observed heterogeneity.
2. Grain size control on chemical compaction. Grain size is thought to strongly influence the amount of pressure solution and resultant chemical compaction undergone by sandstones, with finer grained regions undergoing greater amounts of solution (Houseknecht, 1984, 1988). However, this explanation does not apply to the Mount Simon Sandstone, as regions of similar grain size have very different IGVs (Figs. 8 and 9), and where there is a relationship between IGV and grain size, lower IGVs are observed in coarser regions (Fig. 7).
3. Preferential pressure solution in association with phyllosilicates. Because phyllosilicates facilitate pressure dissolution (e.g., Heald, 1955; Dewers and Ortoleva, 1991; Bjørkum, 1996; Oelkers et al., 1996; Kristiansen et al., 2011), more phyllosilicate-rich regions can become more compacted due to greater pressure solution. Some of the IGV heterogeneity in the Mount Simon Sandstone likely results from such a process (Figs. 6C and 8); however, many of the compacted areas do not contain evidence for pressure solution and do not have greater amounts of phyllosilicates than nearby less-compacted regions, so this cannot be the only explanation.
4. Variable cementation can produce heterogeneous IGVs (e.g., McBride, 2012) if cementation is spatially heterogeneous and occurs prior to significant compaction. In the case of the Mount Simon Sandstone, patches of intergranular evaporite cements are common in areas with high IGV, making this the most likely cause of most of the heterogeneity.
The compaction textures exhibited by portions of the samples seem incompatible with their present, relatively shallow burial depths. In particular, features such as sutured grain contacts and stylolites are often associated with considerably greater burial depths. For example, Renard et al. (1999), based upon their integrated model, noted that significant pressure solution should not normally occur at depths <2 km because quartz kinetics are too slow at low temperatures. Alternatively, some workers concluded that such features can form at shallow depths if phyllosilicates are present at the grain contacts to facilitate pressure solution (e.g., Bjørkum, 1996; Kristiansen et al., 2011). Although phyllosilicates appear to have facilitated some of the pressure solution in the Mount Simon Sandstone, many sutured contacts are free of phyllosilicates, so this mechanism cannot explain all of the pressure solution features. This suggests that the samples were either (1) buried to considerably greater depths in the past, or (2) subjected to anomalously high temperature fluids (which would facilitate chemical compaction).
The Mount Simon Sandstone in the Dallas Center area appears to have undergone both greater burial depths than the present day and anomalously high temperatures. Hegarty et al. (2007) examined the thermal history of Pennsylvanian through Precambrian units in a nearby well (∼64 km west of Dallas Center) using apatite fission track and vitrinite reflectance data and concluded that three thermal events affected the area. The first occurred from 300 to 200 Ma, resulting from an elevated geothermal gradient (∼35 °C/km versus 16 °C/km for the present gradient). The other events, which occurred at 70–50 Ma and 35–10 Ma under the present-day geothermal gradient, resulted from higher temperatures related to 3100 m and 2100 m of paleoburial, respectively. R. Anderson (2012, personal commun.) and others also concluded that the region was subjected to greater burial depths in the past (∼760 m greater depths), though by considerably less than that calculated by Hegarty et al. (2007). In addition, the area may have been covered by as much as 3 km of glacial ice between ∼2.5 Ma and 13 ka (R. Anderson, 2012, personal commun.).
Data from the Illinois Basin indicate that the Mount Simon Sandstone also underwent temperatures considerably greater than present due to greater than present-day burial depths and flow of deep basinal fluids. As noted here, Bowen et al. (2011) indicated that some of the quartz overgrowths occur at unexpectedly shallow depths and suggested that this might be the result of greater burial in the past; they noted that coal maturation studies (Damberger, 1971) suggest that as much as 1.5 km of additional sediment filled the Illinois Basin at its maximum burial. In addition, Hoholick et al. (1984) estimated that the Illinois Basin underwent 914–1524 m of additional burial based upon work by Altschaeffl and Harrison (1959), and Grathoff et al. (2001) concluded that authigenic illite in Ordovician and Cambrian shale partings formed ca. 300 Ma from hot deep basinal brines brought to shallower depth by gravity-driven flow. Rowan et al. (2002) used vitrinite reflectance data and fluid inclusions to conclude that that the rocks underwent a combination of additional burial of as much as 1.2 km and elevated temperatures due to Permian magmatic activity and advective heat transport due to fluid flow. Thus, although there may be disagreement concerning the absolute amount of burial, it is clear that the Dallas Center area underwent both considerably greater burial depths than the present day and a prolonged period during which the geothermal gradient was almost double the present gradient. The combination of these factors explains the anomalously compacted nature of the Mount Simon Sandstone at this locality.
Gypsum and Anhydrite
Gypsum and to a lesser extent anhydrite are present as intergranular pore and fracture filling cements. The gypsum commonly has microporosity, giving it a pinkish color in plane-polarized light (Figs. 5 and 6A). While gypsum and anhydrite are common in the Mount Simon Sandstone at Dallas Center, Fishman (1997) noted that these minerals are rather rare in the Illinois Basin. However, where present, the paragenesis of the gypsum and anhydrite in the two areas is apparently similar (Fishman, 1997).
As discussed here, multiple lines of evidence suggest that cement dissolution occurred, including (1) unusually large heterogeneity in IGVs, (2) the presence of microporosity in gypsum and irregular patches of gypsum and anhydrite in high IGV regions, (3) solution pits in some feldspar overgrowths, and (4) the observation that gypsum cements are often confined to fractures, suggesting that a more soluble cement may have been present, filling intergranular porosity. Carbonate cements are often reported to have undergone extensive subsurface dissolution in sandstones; however, there is no evidence for former carbonate cements in the Mount Simon Sandstone at Dallas Center. Given the presence of remnants of gypsum cement and direct evidence of gypsum dissolution (i.e., microporosity in gypsum), the most likely cement phases to have undergone dissolution are gypsum and anhydrite. However, other more soluble phases such as halite may also have been affected, as is suggested by the gypsum-filled fractures in porous sandstone.
Cement dissolution has also been reported in the Illinois Basin. Hoholick et al. (1984) concluded that all of the porosity in the unit is secondary (with the exception of small amounts of primary porosity near outcrops), resulting mainly from subequal amounts of cement and grain dissolution. Fishman (1997) concluded that most of the current porosity in the Mount Simon Sandstone resulted from the dissolution of cements, perhaps ferroan dolomite.
Skeletal feldspar grains indicate that dissolution of feldspar has occurred (Fig. 6D). The timing of feldspar dissolution is somewhat difficult to interpret. Because dissolution voids in feldspar overgrowths are not filled by adjacent gypsum and quartz overgrowths (Fig. 5B), the dissolution must postdate these phases, as well as the overgrowths. However, possible authigenic feldspar inside skeletal feldspar grains (Fig. 6D) suggests that some feldspar precipitation may have occurred after feldspar dissolution.
Oversized pores are common in the Mount Simon Sandstone and are most likely the result of grain dissolution. In some cases the grains may have been unstable heavy minerals, as discussed here. However, because dissolution of such grains typically results in nearby precipitation of Fe-bearing and sometimes Ti-bearing authigenic minerals, most of these oversized pores that contain no such authigenic phases probably resulted from feldspar dissolution. Dissolution of feldspar grains has been reported for the Mount Simon Sandstone of the Illinois Basin by most previous workers (Hoholick et al., 1984; Fishman, 1997; Bowen et al., 2011).
Pore System Characterization
The size distribution of pore throats resulting from the complex paragenetic history of the Mount Simon Sandstone at this locality is highly variable. Using the pore-throat distributions as derived from the MIP analysis (summarized in Fig. 10), most samples fit into one of three main groups based upon the dominant pore-throat size. These three groups are: mesopore-throat to macropore-throat–dominated (4–40 µm diameter), micropore-throat dominated (<0.07 µm), and an intermediate group (0.15–1.2 µm). Although the intermediate group can be classified as micropore-throat–dominated (cf. Nelson, 2009; Figs. 10–13), for this investigation it is referred to as intermediate. A fourth group is characterized by no dominant pore-throat size, but rather a relatively uniform distribution of sizes. One sample is bimodal, with weak modes in both micropore and macropore ranges. All of the micropore-throat–dominated samples are from the Eau Claire Formation, whereas all of the macropore- and mesopore-dominated samples are from the Mount Simon Sandstone. The intermediate micropore throat and evenly distributed types are present in both units.
The micropore-throat–dominated samples can be further subdivided into three broad categories: those characterized by a single dominant pore-throat size; those with a bimodal pore-throat size; and those characterized by a broad range of pore-throat sizes (Fig. 11). Although many of these samples contain abundant macroporosity, it does not appear in the mercury porosimetry data; which indicates that it is not effective porosity (i.e., poorly interconnected). This can be clearly seen in thin section, where macroporosity occurs in poorly interconnected subdomains (Fig. 14). The variability in pore-throat size distribution reflects the considerable textural heterogeneity of the samples, including heterogeneities imparted by depositional texture, as well as variable distribution of pore-filling cements.
The mesopore-throat to macropore-throat samples fall into two main distributions: those that are characterized by a single dominant pore-throat size, and those that have a broader distribution of sizes with several modes (Fig. 12). There is a clear primary textural influence on the two populations; the former type occurs in lithologically and texturally homogeneous samples, and the latter type occurs in texturally and lithologically heterogeneous samples. For a subset of four of the mesopore- to macropore-dominated samples, mercury intrusion data were collected for both horizontally and vertically oriented jacketed core plugs (Fig. 15). For all but one of these horizontal-vertical sample pairs, the MIP measured porosity (Table 3) is slightly greater for the horizontally oriented samples than the vertically oriented samples. Because textural and lithologic heterogeneity is greatest in the vertical direction in most of the samples, this difference in accessible pore volume may result from three-dimensional effects in which zones of porosity are enclosed by very small-pore-throat domains. This interpretation is strengthened if we consider the sample exhibiting the greatest difference in horizontal and vertical pore-throat size distribution. For three of the four paired samples the pore-throat size distribution did not differ significantly between horizontal and vertical analyses; however, for sample 2967.20 (904 m), the horizontally-oriented plug shows a distinct 9 µm pore-throat diameter peak, whereas the vertically oriented plug shows a relatively evenly distributed pore-size distribution with no clear peaks. Sample 2967.20 (904.4 m) is the most lithologically heterogeneous of the four samples, containing conglomerate, sandstone, and mudstone. In addition, zones of relatively large macropores are bounded above and below by thin mudstone laminae. Such zones would have been accessible to mercury injected horizontally, but are less likely to have been accessed by vertical injection due to the bounding mudstone. One of the four samples, 2931.10 (893.4 m), exhibits no difference in porosity between the horizontal and vertical analyses. This sample has the least textural heterogeneity of the four samples. However, it is cut by a gypsum-filled fracture that apparently did not influence the measurements. This is likely due to the fact that the gypsum fracture fill only traverses ∼60% of the sample, allowing ready access of mercury to both sides of the fracture. In conclusion, the results for paired directional samples indicate that the MIP-measured porosity for many, perhaps most, of the vertically-oriented measurements underestimate the total sample porosity. In addition, given the clear textural control on the pore-throat results, and the high degree of lithologic and textural heterogeneity of many of the samples, a portion of the vertically measured samples that exhibit no clear pore-throat peak probably would if measured horizontally.
Samples exhibiting intermediate, evenly distributed, and bimodal distributions in pore-throat diameters have one thing in common: they are all texturally and lithologically heterogeneous, with the greatest heterogeneity in the vertical direction (Fig. 13). All of these samples were measured in the vertical direction; given the apparent influence of sample orientation discussed here, it is likely that the measured values were influenced by this vertical heterogeneity, and that if measured horizontally, the results may have been considerably different. Data for one of the samples, 2967.20 (904.4 m), support this contention. This sample was measured both vertically and horizontally. The horizontal measurement produced a pore-throat size distribution in the mesopore- to macropore-throat category, whereas the vertical measurement placed it in the evenly distributed category. The sample exhibiting a bimodal pore-throat distribution (2983.42; 909.3 m) appears to have been dramatically influenced by this vertical heterogeneity. It consists of interlayered sandstone and mudstone (Fig. 13). Two populations of pore sizes are evident, with a major peak in the micropore throat range and a weaker peak in the mesopore throat range. The sample is heterolithic, mainly consisting of sandstone (∼90%) with a layer of mudstone at its base. Although sandstone makes up the majority of the sample, it appears that the thin mudstone layer formed a continuous barrier in the jacketed sample, controlling mercury intrusion and providing a skewed analysis of overall pore-throat size distribution.
Sealing capacity is the column (vertical) height of a non-wetting phase (e.g., supercritical CO2 or compressed air) that can be held by the capillarity of a water-saturated (water wetting) rock. Variability in sealing capacity as estimated from MIP data (see Materials and Methods section) for a reservoir rock or caprock, respectively, can indicate a potential for local (intraformational) trapping and/or change in local multiphase flow conditions, or a range in quality of sealing behavior of a caprock. Sealing capacity can also be compared to the hydrostatic and lithostatic pressure gradients at a given site in order to estimate pressures or non-wetting phase column heights that would lead to capillary breakthrough of a caprock, or pressures that may cause fracturing of the caprock (Heath et al., 2012). Our goal is to use sealing capacity to investigate heterogeneity of the reservoir and caprock lithofacies; placing the data in terms of hydrostatic pressure gradients, lithostatic gradients, and a measure of the pressure that would lead to failure and/or fracturing of the caprock is beyond the scope of this study.
The Eau Claire Formation has the highest breakthrough pressures and sealing capacity (in terms of the columns heights of air or CO2 that can be held by capillary forces) near its contact with the subjacent Mount Simon Sandstone (see Figs. 2 and 16; Table 3). The Eau Claire Formation in general shows high variability in sealing capacity (and capillary pressure behavior), reflecting a range in pore size distribution, structure, and connectivity. This is consistent with the findings of Lahann et al. (2014) in the Illinois Basin. Estimated pore-throat size diameters at breakthrough range from 0.014 to 0.936 μm, equivalent to a range in mercury-air breakthrough pressure over 109.2–1.6 MPa. The Mount Simon Sandstone exhibits generally more uniform capillary pressure behavior as expressed as air and CO2 column heights; however, in the middle unit much higher values can occur that reflect interbedded mudstones (Fig. 16). Mount Simon Sandstone pore-throat diameters range from 0.024 to 37.327 μm, with a corresponding mercury-air breakthrough pressure of 60.98–0.04 MPa. This large change in breakthrough pressure and hence sealing capacity indicates that the interbedded mudstones may act as an upward barrier to the injected air or CO2. Compressed air energy storage requires formation of an initial large bubble or zone of injected air that efficiently displaces ambient groundwater to irreducible saturations (Heath et al., 2013). Similarly, CO2 storage requires sweep efficiencies that allow for suitable storage volumes. At this site, the range in sealing capacity reflects that sweep efficiencies may be poor, so that air or CO2 focused flow may occur underneath the interbedded mudstones (the zones of high column heights within the Mount Simon Sandstone; Fig. 16; Heath et al., 2013).
The range in groundwater-CO2-rock contact angles of 13°–70° used in this analysis (see Materials and Methods section) can result in a large difference in sealing capacity (e.g., ∼3054 m difference in CO2 column height at the depth of 8890 m). This range in groundwater-CO2-rock contact angles was used due to the multiple mineral phases in rock and the contact angles associated with those phases. Estimates of CO2 sealing capacity can thus be highly uncertain due to uncertainty in contact angle values and what minerals line pore throats. However, even with the high uncertainty, the estimated CO2 column heights are much greater than the thickness of the Mount Simon Sandstone at the Dallas Center structure for the interbedded mudstone at a depth of 909 m (Fig. 16). The difference in sealing capacity between the air and CO2 systems is predominately due to the density difference. At the depths mentioned above, CO2 has an average density for the samples of ∼820 kg/m3, whereas nitrogen has an estimated density of 90 kg/m3. The buoyancy force is thus much greater for air for these depths. CO2 is a dense state due to its presence as a supercritical fluid at these depths. Therefore, uncertainty in interfacial tension and contact angle has less influence on the sealing capacity differences between the air and CO2 systems as compared to the density differences between the air and CO2 for the estimated in situ depths and temperatures.
Regional Mount Simon Sandstone Paragenesis and Anomalous Evaporite Cements
As noted herein for individual alterations, the major diagenetic events and their timing are remarkably similar in both this study and in the previous studies that examined samples from throughout the Midwest. The similarity is striking given the large regional extent of the unit, and the widely variable burial depths of the samples examined; this similarity was also noted by Fishman (1997). Although the overall paragenesis of the Mount Simon Sandstone is similar throughout the region, there are a few significant differences. The most striking difference is the relatively large amount of gypsum and anhydrite found in the Dallas Center area versus elsewhere. There are two possible explanations for this difference: (1) significant gypsum and anhydrite precipitation was restricted to certain areas, such as Dallas Center, due to a heterogeneous distribution of saline brines, perhaps resulting from locally focused flow of deeper basinal fluids; and (2) gypsum and anhydrite were more abundant throughout the region in the past and were removed in most places by dissolution. The strong evidence for significant cement dissolution in the Mount Simon Sandstone favors the latter explanation. If this interpretation is correct, then something about the geologic setting (e.g., structural position) of the Dallas Center area must have protected the Mount Simon Sandstone from the circulation of the more dilute (evaporite dissolving) fluids, allowing more of the evaporite cements to be preserved in this region.
Evolution of Porosity in the Mount Simon Sandstone: Timing of Porosity Development
Primary porosity in the Mount Simon in the Dallas Center area was greatly reduced by mechanical and chemical compaction, as well as precipitation of authigenic minerals, most notably gypsum, anhydrite, feldspar, and quartz. We agree with Fishman (1997), who concluded that most of the present porosity originated though cement dissolution. Our data suggest that the dissolved cement consisted of evaporites, such as gypsum, anhydrite, and perhaps halite.
The timing of evaporite dissolution, and the probable creation of most of the current porosity, can be constrained relative to other diagenetic events, some of which have been dated either directly or indirectly. Based upon thin section observations, evaporite dissolution is a late-stage event that postdates all diagenetic alterations in the Mount Simon Sandstone, with the possible exception of feldspar dissolution (Fig. 4). Determining the absolute timing of the dissolution event is not possible with the current data set; however, the relative timing of dissolution can be constrained. There is clear petrographic evidence that the evaporite cements postdate feldspar overgrowths, which, as discussed above, precipitated prior to ca. 400 Ma (Early Devonian). Thus, the dissolution event must have occurred subsequent to 400 Ma. Because the dissolution event must postdate significant compaction, based upon the discussion of overcompaction and thermal history, we can further conservatively constrain the timing to younger than 300 Ma (Pennsylvanian).
The possibility of widespread evaporite dissolution in the Mount Simon Sandstone and other units in the Illinois Basin was discussed by Bethke (1986). In considering the origin of Mississippi Valley Type (MVT) ore bodies in the Upper Mississippi Valley mineral district, Bethke (1986) noted, based upon mass balance considerations, that meteoric waters that entered aquifers on the Pascola arch might have become saline by dissolving evaporate as the water moved northward to the Wisconsin arch. Thus, it is possible that the mineralizing fluids that formed the district in part gained salinity from evaporite dissolution in the Mount Simon Sandstone. If this was the case, based upon radiometric dating of ore mineralization (Pannalal et al., 2004), the dissolution event may have occurred ca. 270 Ma (Permian). Rowan and Goldhaber (1995) concluded that the mineralizing fluids flowed through the Mount Simon Sandstone over a period of ∼200 k.y., providing a minimum duration of any dissolution event.
An additional possibility for widespread dissolution of evaporite cements in the Mount Simon Sandstone is that it occurred at least in part in response to enhanced recharge of meteoric water related to the presence of Pleistocene ice sheets. Person et al. (2007) and McIntosh et al. (2002) provided evidence indicating that groundwater flow in the Illinois and other midcontinent sedimentary basins was greatly affected by changes in hydraulic head associated with ice sheet topography. McIntosh et al. (2002) concluded that the south to north tectonically controlled flow system of the Illinois Basin was reversed to a north to south flow system during the Pleistocene; they documented recharge of meteoric waters associated with this mechanism to depths of at least 1 km. Thus, in addition to dissolution that may have occurred in association with tectonically driven groundwater flow and MVT ore mineralization, dissolution likely also occurred more recently, driven by Pleistocene ice sheet hydrology. Consequently, the ice sheets, in addition to possibly contributing to the locally overcompacted nature of the Mount Simon Sandstone, may have provided a source of dilute water and elevated hydraulic head necessary to drive further cement dissolution and porosity creation at depth. However, this mechanism cannot account for dissolution at depths >∼1 km (McIntosh et al., 2012), so it cannot be the sole mechanism.
Implications of Pore-Throat Size Heterogeneity for Fluid Flow
The pore and pore-throat size distribution of the Mount Simon Sandstone samples examined at this locality are often highly heterogeneous, a reflection of (1) variable IGV due to cement dissolution and (2) primary textural heterogeneity at the thin section scale. The net result of this heterogeneity is that there is a greater than normal opportunity for capillary trapping (see Holtz, 2002; Saadatpoor et al., 2010). The variation in capillary behavior (expressed through breakthrough pressures and sealing capacity column heights; see Table 3; Figs. 2 and 16) of the middle Mount Simon Sandstone may lead to local regions of higher than average (relative to the rest of the formation) capillary trapping of air or CO2 (or the non-wetting phase). The amount of a non-wetting fluid (e.g., air or CO2) in a water-wet system that is held in place by buoyancy due to capillary forces is affected by local variations in capillary breakthrough pressure (Saadatpoor et al., 2010). When the capillary breakthrough pressure is locally larger than the average of the underlying reservoir rocks (of the same formation), additional capillary trapping can occur. Saadatpoor et al. (2010) noted that the volume stored by this trapping mechanism can be much larger than the average amount residually trapped for a particular rock type. For the site in question, capillary trapping heterogeneity could possibly lead to the nonuniform formation of a bubble for CAES and poorer sweep efficiency during cycling of air into and out of the formation (see Heath et al., 2013). For CO2 storage, local capillary trapping may be desirable as a mechanism that improves the overall security of confinement by promoting residual-saturation trapping within the storage reservoir for CO storage (see Saadatpoor et al., 2010). However, variations in capillary breakthrough pressure may also affect CO2 injectivity and lead to poor sweep in storage reservoirs that have interbedded lithofacies with highly variable capillary entry and/or breakthrough pressures, such as mudstones interbedded with sandstones. Future modeling of the site with the measured variation in capillary pressure behavior could quantify the predicted site performance of CAES and CO2 storage. Our results quantitatively demonstrate the significant impact millimeter-scale to micron-scale porosity heterogeneity can have on the migration of multiphase fluids in reservoir sandstone. Such heterogeneity should be taken into account when attempting to model multiphase flow in such reservoirs.
1. The mineralogy, texture, and paragenesis of the Mount Simon Sandstone at Dallas Center are very similar to those of the Mount Simon Sandstone elsewhere in the Midwestern United States, with one exception, relatively large amounts of evaporite cements (gypsum, anhydrite, and barite).
2. Most of the porosity in the Mount Simon Sandstone at this locality is the result of cement and grain dissolution. The majority of the porosity formed through the dissolution of highly soluble cements, such as gypsum and anhydrite, as well as some framework grains. This is probably the case for the Mount Simon Sandstone in other areas as well, including the Illinois Basin. This dissolution event most likely occurred either in the Permian, in association with MVT ore body development, or possibly in the Pleistocene as a result of deep circulation of meteoric water associated with variable ice-sheet geometry.
3. The pore-size distribution in the Mount Simon Sandstone at Dallas Center, Iowa, is unusually heterogeneous at small scales due to the combined effects of primary textural heterogeneity, compaction, and cement dissolution. This heterogeneity is particularly evident when comparing vertical and horizontal mercury porosimetry data sets for the same samples, which are in some cases dramatically different.
4. The unusually heterogeneous character of the pore system within in the Mount Simon Sandstone should result in an unusually large amount of local capillary trapping of supercritical CO2 or injected compressed air. Neglecting to characterize such millimeter-scale to micron-scale heterogeneity in similar reservoir sandstones could result in erroneous modeling efforts.
5. Although the storage capacity and caprock of the Dallas Center structure are adequate for a CAES facility, the generally low permeabilities and heterogeneous nature of the permeability and multiphase flow (capillary pressure behavior) make it an undesirable location for development as a CAES reservoir.
The U.S. Department of Energy (DOE) Storage System Program, the Iowa Stored Energy Plant Agency, and the DOE National Energy Technology Laboratory (NETL; grant DEFE0004844) funded this work. Dewers and Heath were supported in part by the Center for Frontiers of Subsurface Energy Security, an Energy Frontier Research Center funded by the DOE, Office of Science, Office of Basic Energy Sciences, award DE-SC0001114. Mozley was funded by the NETL portion of the project, which was managed and administered by the New Mexico Institute of Mining and Technology and funded by DOE-NETL and cost-sharing partners.
We thank Raymond Anderson and Robert McKay of the Iowa Geological and Water Survey for discussions about regional lithofacies variation of the Mount Simon Sandstone and the Eau Claire Formation. Mark Rodriguez of Sandia National Laboratories performed the X-ray diffraction. John Neasham of Poro-Technology performed mercury porosimetry (Poro-Technology has subsequently been acquired by Micromeritics). Sandia National Laboratories is a multiprogram laboratory managed and operated by the Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the DOE’s National Nuclear Security Administration under contract DE-AC04–94AL85000. The manuscript benefited greatly from the comments and suggestions of Shanaka de Silva (editor) and reviewers Brenda Bowen and John A. Rupp.