Natural Resources Research

, Volume 16, Issue 3, pp 277–292

H2-rich and Hydrocarbon Gas Recovered in a Deep Precambrian Well in Northeastern Kansas

Authors

    • Kansas Geological SurveyUniversity of Kansas
  • John H. Doveton
    • Kansas Geological SurveyUniversity of Kansas
  • Daniel F. Merriam
    • Kansas Geological SurveyUniversity of Kansas
  • Barbara Sherwood Lollar
    • Department of GeologyUniversity of Toronto
  • William M. Waggoner
    • WTW Oil Co., Inc.
  • L. Michael Magnuson
    • Kansas Geological SurveyUniversity of Kansas
Article

DOI: 10.1007/s11053-007-9052-7

Cite this article as:
Newell, K.D., Doveton, J.H., Merriam, D.F. et al. Nat Resour Res (2007) 16: 277. doi:10.1007/s11053-007-9052-7

Abstract

In late 2005 and early 2006, the WTW Operating, LLC (W.T.W. Oil Co., Inc.) #1 Wilson well (T.D. = 5772 ft; 1759.3 m) was drilled for 1826 ft (556.6 m) into Precambrian basement underlying the Forest City Basin in northeastern Kansas. Approximately 4500 of the 380,000 wells drilled in Kansas penetrate Precambrian basement. Except for two previous wells drilled into the arkoses and basalts of the 1.1-Ga Midcontinent Rift and another well drilled in 1929 in basement on the Nemaha Uplift east of the Midcontinent Rift, this well represents the deepest penetration into basement rocks in the state to date. Granite is the typical lithology observed in wells that penetrate the Precambrian in the northern Midcontinent. Although no cores were taken to definitively identify lithologies, well cuttings and petrophysical logs indicate that this well encountered basement metamorphic rocks consisting of schist, gneiss, and amphibolitic gneiss, all cut by aplite dikes.

The well was cased and perforated in the Precambrian, and then acidized. After several days of swabbing operations, the well produced shows of low-Btu gas, dominated by the non-flammable component gases of nitrogen (20%), carbon dioxide (43%), and helium (1%). Combustible components include methane (26%), hydrogen (10%), and higher molecular-weight hydrocarbons (1%). Although Coveney and others [Am. Assoc. Petroleum Geologists Bull., v. 71, no, 1, p. 39–48, 1987] identified H2-rich gas in two wells located close to the Midcontinent Rift in eastern Kansas, this study indicates that high levels of H2 may be a more widespread phenomenon than previously thought. Unlike previous results, the gases in this study have a significant component of hydrocarbon gas, as well as H2, N2, and CO2. Although redox reactions between iron-bearing minerals and groundwater are a possible source of H2 in the Precambrian basement rocks, the hydrocarbon gas does not exhibit the characteristics typically associated with proposed abiogenic hydrocarbon gases from Precambrian Shield sites in Canada, Finland, and South Africa. Compositional and isotopic signatures for gas from the #1 Wilson well are consistent with a predominantly thermogenic origin, with possible mixing with a component of microbial gas. Given the geologic history of uplift and rifting this region, and the major fracture systems present in the basement, this hydrocarbon gas likely migrated from source rocks and reservoirs in the overlying Paleozoic sediments and is not evidence for abiogenic hydrocarbons generated in situ in the Precambrian basement.

Keywords

Forest City Basinlow Btu gasPrecambrian metamorphicsthermogenic vs. abiogenic gas

Introduction

The Precambrian rocks underlying Kansas are primarily igneous and metamorphic rocks and are not generally considered to be potential targets for oil and gas. The Kansas Precambrian lacks a source for petroleum and contains few possible reservoirs with the exception of metasediments in the Midcontinent Rift System (MRS), where source beds and potential traps and reservoirs have been identified (Berendsen and others, 1988; Newell, Burruss, and Palacas, 1993). A small amount of production comes from rocks of this age along the Central Kansas Uplift, parts of the Cambridge Arch, and Nemaha Uplift (Newell and others, 1987), but this production is attributed to migration into these localities from adjacent sedimentary rocks. The production is mainly from the granite wash, a porous weathered residuum of the granitic rocks, and fractures in the crystalline basement (Walters, 1953). Precambrian hydrocarbons as a fraction of Kansas oil and gas production is minimal, accounting for about 0.1% of the total (Merriam and Goebel, 1959).

Almost 4500 of the 380,000 wells drilled in the state are recorded to have reached the Precambrian basement (Cole, 1976; Cole and Watney, 1985). The basement, lacking conventional source rocks and reservoir rocks, usually signals an end to prospective hydrocarbon zones and so is normally penetrated to the depth of a few feet or at most, a couple of hundred feet before cessation of drilling. However, 17 wells are known to have penetrated the basement more than 1000 ft (300 m) and most of these were drilled in the 1920s, 30s, and 40s (Table 1). Direct information on Precambrian geology is thus superficial with the exception of records from these rare, deep, and mostly older wells.
Table 1

Wells in Eastern Kansas Drilled more than 1000 ft (305 m) into the Precambrian

County (Section, Township, Range)

Date Drilled

Precambrian Footage Drilled Lithology Encountered

WA (sec. 35, T5S, R5E)

1985

8450 arkose, basalts*

NE (sec. 19, T3S, R11E)

1929

2552 granite

MS (sec. 4, T4S, R7E)

1986

1840 clastics*

BR (sec. 32, T1S, R17E)

2005

1826 metamorphics**

RL (sec. 8, T7S, R5E)

1986

1784 gabbro*

MS (sec. 4, T4S, R7E)

1926

1725 clastics*

MS (sec. 4, T4S, R7E)

1929

1690 metamorphics*

DK (sec. 13, T13S, R2)

1932

1593 arkose*

MS (sec. 29, T2S, R9E)

1922

1580 granite

WO (sec. 29, T26S, R17E)

1972

1476 granite

GW (sec. 4, T24S, R10E)

1924

1290 schist

BT (sec. 17, T19S, R11W)

1927

1285 schist

AL (sec. 25, T24S, R18E)

1908

1277 metamorphics***

WO (sec. 29, T26S, R17E)

1923

1231 granite

CS (sec. 2, T20S, R7E)

1917

1225 granite

EW (sec. 3, T16S, R8W)

1926

1055 arkose

AL (sec. 34, T23S, R21E)

2003

1000+ metamorphics***

*Located in the Midcontinent Rift System.

**Well discussed in this report.

***Material is a fine-grained sandstone subjacent to the Reagan Fm. (U. Cambrian), but may not be Precambrian.

Kansas county abbreviation: AL, Allen; BR, Brown; BT, Butler; CS, Chase; DK, Dickinson; EW, Ellsworth; GW, Greenwood; MS, Marshall; NE, Nemaha; RL, Riley; WA, Washington; WO, Woodson.

It is difficult to understand why a well would be drilled so deep into the basement unless the basement rocks were softer than expected and drilled exceptionally easy. In early days, some operators possibly concluded that the “granite” was a shallow, thin sheet and thus could be drilled through to sediments beneath. Some, including State Geologist Erasmus Haworth (Haworth, 1896), when basement on the Nemaha Uplift was encountered at less that 600 ft (183 m) depth, simply did not believe that granite had been encountered so shallow, because if it had, he conceived that it would negate the occurrence of petroleum in western Kansas.

The MRS is sparsely drilled, but the deepest well in Kansas, the Texaco #1 Poersch well, was drilled in the MRS in Washington County to a depth of 11,296 ft (3443 m) in 1984–1985. No significant hydrocarbons shows were encountered in the numerous arkoses and basalts flows penetrated by this well, but other wells in the MRS have encountered organic-rich shale with oil-filled fluid inclusions in calcite-filled fractures (Newell, Burruss, and Palacas, 1993).

Data on wells penetrating the Precambrian are sparse in the central parts of the basins and most dense on the backs of the uplifted areas—Nemaha Uplift, Central Kansas Uplift, and Cambridge Arch. Whereas the stratigraphic section is most complete in the basins, it is truncated on the uplifts where widespread and lengthy unconformities are present reflecting the structural development of these features (Merriam, 1963).

Regional Geology

The WTW Operating LLC #1 Wilson well is located in the central part of the Forest City Basin in northeastern Kansas (Lee, 1943; Merriam, 1963). This shallow cratonic basin extends to the northeast into Nebraska and Iowa, and is filled with Paleozoic sediments ranging in age from Cambrian to Permian with a thin veneer of Cretaceous, Tertiary, and Pleistocene. The lower Paleozoic strata are mainly thick bodies of thin-bedded carbonates alternating with thin units of clastics with a fairly uniform thickness across the central part of the basin (Fig. 1). Near the end of the Mississippian and in the early part of the Pennsylvanian, the area was downwarped and a thick section of clastics was deposited punctuated by thin and widespread carbonates (Fig. 2). A long interval of tectonic quiescence and erosion occurred until the Cretaceous Western Interior Sea spilled over into eastern Kansas. An apron of Tertiary deposits, mainly clastics shed from the rising Rocky Mountains to the west, and the thin veneer of Pleistocene, residual material from the Kansan and Nebraskan glaciers, tops the sedimentary sequence.
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig1_HTML.gif
Figure 1

Generalized Paleozoic stratigraphic column and relative thicknesses for northeastern Kansas. Lower Paleozoic reservoirs—Ordovician Simpson Gp. and Viola Limestone, and Silurian-Devonian Hunton Gp.—mainly produce oil from several structural traps along the axis of the Forest City Basin. Pennsylvanian (Desmoinesian) Cherokee and Marmaton Groups, and Pennsylvanian sandstones and Mississippian limestones at the basal Pennsylvanian unconformity produce both oil and gas on the eastern flank of the Forest City Basin in the vicinity of Kansas City

https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig2_HTML.gif
Figure 2

Burial history diagram for the center of the Forest City Basin in northeastern Kansas (from Newell and others, 1989). The deepest burial occurred in the Cretaceous when the Precambrian was buried to a depth of about 6000 ft (1830 m)

The basement in northern Kansas has been characterized as a granitic terrane with patches of metasediments (Fig. 3). Based on sample dating from wells in this area, Van Schmus and others (1993) suggested that this Southern Central Plains terrane (sCP) ranges in age from 1.579 to 1.649 Ga and is intruded by younger granite bodies and crosscut by the major but younger Midcontinent Rift System (∼1.1 Ga). The MRS is a Precambrian failed continental rift filled with lightly metamorphosed sediments intruded by a series of basaltic and gabbroic bodies, with a thick accumulation of arkose derived from the immediately surrounding area. The WTW Operating LLC #1 Wilson is located a short distance from the Nebraska border and its basement lithologies seem to have similarities with the Precambrian of southeastern Nebraska, which is within the Northern Central Plains terrane (nCP). Carlson and Treves (2005) described the Nebraskan basement rocks as composed of metamorphic gneiss and schist of island-arc origin (1.84 to 1.71 Ga in age) penetrated by younger granite plutons (1.78 to 1.35 Ga). A map of the accreted terranes extends into Kansas and the well seems to be located in the Dawes terrane with an estimated age of 1.78 Ga (Carlson and Treves, 2005, Fig. 5). Features of the Precambrian terranes have been cited to favor origin as either accreting arc-crustal elements or as the result of extensional tectonics with the juxtaposition of disparate crustal blocks. Bickford and Hill (2007) proposed an expanded interpretation that suggests that earlier accreted arcs were disrupted by later movement on transcurrent shear zones. The range in ages, rock types, and interpretations of these regional Precambrian studies shows that current understanding of this complex terrane is in flux, which is not surprising considering the limited data available. Consequently, a new and deep well that penetrates the Precambrian in northeast Kansas holds special interest for a variety of perspectives.
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig3_HTML.gif
Figure 3

Sketch map of eastern Kansas and Nebraska showing relation of Precambrian accretionary terranes and boundary sutures (heavy dashed lines) in eastern Kansas and Nebraska (from Bickford and others, 1981; Bickford and Hill, 2007; Carlson and Treves, 2005; Van Schmus and others, 1993; Marvin Carlson, pers. comm., 2006). Age range (in Ga) for each terrane is from Carlson and Treves (2005). Notation: sCP, southern Central Plains (mainly granitic gneiss); nCP, northern Central Plains province (mainly gneiss); SGR, Southern Granite-Rhyolite province; MRS, Midcontinent Rift System (slightly metamorphosed sediments, basaltic volcanics and intrusives); dark gray: metamorphics, mainly schist; light gray: younger granite bodies intruded into the basement. Age dates show the suggested accretionary terranes and boundary sutures (from Carlson and Treves, 2005). The younger intrusive granites (∼1.34 Ga, in light gray) in eastern Kansas may be related to the SGR terrane in southern Kansas of approximately the same age. The MRS, which extends from Minnesota to Oklahoma, cuts across the older terranes (Newell, Burruss, and Palacas, 1993). Accretionary terranes (from oldest to youngest) are: Superior/Penokean, Dawes, Frontier, Hitchcock, and Western Kansas (from Carlson and Treves, 2005). The #1 Wilson well in Brown County, KS sits atop the Dawes terrane (1.78 Ga)

Precambrian Geology of the WTW Operating LLC #1 Wilson Well

In 2005–2006 in extreme northeastern Kansas in the Forest City Basin, the WTW Operating LLC #1 Wilson well was drilled to a total depth of 5772 ft (1759 m), 1826 ft (557 m) into the Precambrian basement on a venture testing the possibility of oil or gas production in these rocks. Elevation of the well is 1000 ft (304.8 m) (KB is 1009 ft [307.5 m]). A bottomhole temperature (BHT) of 135°F (57.2°C) was measured at TD. This well in Brown County (sec. 32, T1S, R17E) penetrated deeper into the Precambrian basement than any other well in the state, except for two wells in the Midcontinent Rift System and an early well drilled on the Nemaha Uplift (Table 1). The data collected from the #1 Wilson well has provided the basis for interpretation on the geologic history of the area supplementing previous work. The Precambrian section of the #1 Wilson well was plugged back to 4000 ft (1219 m) in March 2006 and the casing was shut in pending later testing in several Paleozoic zones. A set of 10-ft (3.05-m) samples were collected from the well, and they have been deposited with the Kansas Well Log Bureau in Wichita; no cores were taken. Only the fines were saved, and there are missing intervals; samples contain abundant cavings or are dirty, all of which makes them of limited value.

The top of the Precambrian is at a depth of 4000 ft (1219 m) under a 100-ft (30-m) section of Upper Cambrian Reagan Sandstone. The uppermost 510 ft (155.5 m) of Precambrian is a biotite schist with fragments of milky quartz interpreted as aplite veins. The biotite is brownish and lustrous. At 4510 ft (1374.7 m) there are fragments of a dark gray to greenish gray crystalline amphibolite in the next 80 ft (24.4 m). A subtle change occurs at 4590 ft (1399 m) to a more gneissic texture and the coloration is darker. This metasediment occurs at the bottom of the borehole again with an interspersion of quartz fragments, which may represent aplite dikes. No granite was encountered in the #1 Wilson well. These metasediments are similar to the limited number of other metasediments described in the Precambrian terrane of eastern Kansas (Van Schmus and others, 1993). However, the metamorphic section is contrasted with the Precambrian in nearby wells, all of which report granite (Cole and Watney, 1985). The closest Precambrian well (Hodgden & Associates #20-1 Mosquito Creek, sec. 20, T5S, R14E) is about 25 miles (40 km) to the southwest. This well-encountered Precambrian quartz monzonite at 3918 ft (1194.2 m) depth (−2673 ft; −814.7 m subsea) (G. Hodgden, pers. comm., 2005).

Existing oil and gas production in the Forest City basin are principally from Paleozoic reservoirs. Sandstones in the Ordovician Simpson Group, dolomites in the overlying Ordovician Viola Limestone, and Silurian-Devonian section (Fig. 1) mainly produce oil from a series of structural traps along the axis of the Forest City Basin. Pennsylvanian (Desmoinesian) Cherokee and Marmaton Groups, and Pennsylvanian sandstones and Mississippian limestones at the basal Pennsylvanian unconformity produce both oil and gas from both structural and stratigraphic traps on the eastern flank of the Forest City Basin (Newell and others, 1987).

In addition to the set of drill-cuttings samples, a suite of wireline logs were run on the #1 Wilson well, and these are on file at the Kansas Geological Survey. The gamma ray, density, and neutron porosity log measurements contain the most diagnostic clues to the rock composition of the Precambrian metamorphic section. Although there is abundant information on the petrophysical properties of sedimentary rocks, sources of wireline petrophysical data on metamorphics are restricted primarily to the recent program of deep continental drilling, and the most useful comparisons come from the KTB borehole in southern Germany. Pechnig, Delius, and Bartetzko (2005) summarized the ranges in log responses associated with acid igneous and metamorphic in continental boreholes, which have aided log interpretation in the #1 Wilson well when coordinated with lithologies observed in drill cuttings. Examination of the gamma ray—density—neutron cross plot (Fig. 4) shows a basic discrimination between two petrophysical facies, which is matched by a sharp break at a depth of 4500 ft (1371.6 m) (Fig. 5). The section above this break is identified as biotite schist, whereas the lower section is considered to represent layered gneisses of differing composition. Amphibolite had been recognized in drill cuttings, but is unlikely to occur in layers of significant thickness, because it has a low gamma ray reading (about 20 API units), whereas the entire Precambrian section is characterized by a high gamma ray response. Instead, the amphibolite is most likely to occur within a gneissic host, and the amphibolitic gneiss seems to form an end member of a gneiss complex ranging through gneiss to aplitic gneiss (Fig. 4). Average log responses of the four-petrophysical lithofacies are summarized in Table 2. Log cut-off values were selected to coincide with apparent breaks in the petrophysical data clouds and applied to the subdivision of the lower section among amphibolitic gneiss, gneiss, and aplitic gneiss (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig4_HTML.gif
Figure 4

Precambrian log zones on cross plot of gamma ray (API units), density (g/cc), and neutron porosity (%), showing differentiation between petrophysical facies matched with biotite schist, amphibolitic gneiss, gneiss, and aplitic gneiss

https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig5_HTML.gif
Figure 5

Gamma ray (GR), density (RHOB), neutron porosity (NPHI%), and gas-chromatograph (GC) logs in the Precambrian section of the WTW Operating LLC #1 Wilson well together with depth profile of interpreted lithofacies (FACIES). Gas shows during drilling occurred deep within the Precambrian section, but are not closely associated to lithology

Table 2

Average Log Petrophysical Properties of Interpreted Precambrian Lithofacies in the WTW Operating LLC #1 Wilson Well

 

Gamma Ray (API units)

Density (g/cc)

Neutron Porosity (%)

Biotite schist

108

2.68

5.3

Amphibolite gneiss

161

2.63

6.7

Gneiss

164

2.57

5.4

Aplitic gneiss

171

2.50

0.7

The mud-gas chromatography shows gas influx in the lower part of the Precambrian section (Fig. 5) which is not associated with a specific lithofacies. However, gas flow into the well is most likely transmitted through communicating fractures, so the petrophysical log suite was evaluated in terms of a variety of “fracture indicators” (Schlumberger, 1987, p. 192). The only indicator that showed a good-depth concordance with gas inflow was provided by the shallow resistivity laterolog with lowered values indicative of water-filled vertical features contrasted with the high-resistivity crystalline rock. However, acceptance of this interpretation implies an additional 300 ft (91 m) of fractured gneiss that immediately underlies the biotite schist. The sonic log shows highly anomalous features that are concentrated over approximately the same depth range as the gas entry. However, these anomalies are high-velocity zones, whose physics are mysterious and differ in both thickness and character from cycle-skipping features usually caused by acoustic attenuation associated with vuggy or highly fractured rocks.

Fluids Recovered from the WTW Operating LLC #1 Wilson Well

When the well was drilling at 5385 ft (1641.3 m), approximately 1400 ft (427 m) into the Precambrian, gas chromatography of the drilling mud suddenly recorded elevated levels of methane, ethane, and propane (Fig. 5). The hot-wire gas detector for the drilling mud registered 80 units of gas over its normal background of 15 units. Elevated gas levels, around 50 units, persisted for the next 7 days until TD was reached, with some spikes to 80 units over 10-to-20-ft (3-to-6-m) intervals. No cut or fluorescence was noted in the drilling samples associated with these elevated levels of mud gas.

During the drilling of the well, only one prospective interval (5710 to 5758 ft; 1740.4 to 1755.0 m) was drillstem tested. The tool was opened 30 minutes with a good surge of 10–20 bubbles and then the well died. At that point, the pipe would not turn so that the pressure bombs on the drillstem test tool could be opened (likely because of the tool being poorly lodged in the hole, which was known, by surveys during drilling, to be deviated by several hundred feet below the top of the Precambrian), thus no shut-in pressures could be taken. The only recovery in the tool chamber was 15 ft (5 m) of drilling mud. An initial flowing pressure of 34 psi (0.23 mPa) and a final flowing pressure of 36 psi (0.25 mPa) were recorded. The initial hydrostatic pressure was 2727 psi (18.8 mPa); the final hydrostatic was 2635 psi (18.2 mPa). The bottom-hole temperature was 123°F (51°C). Further drillstem tests were cancelled, because of the danger of getting the tool stuck in the hole. The well then was drilled on to 5671 ft (1728.5 m), and the wireline log suite was run.

After 5½ in (13.97 cm) production casing was placed in the well, the bore was surveyed vertically by GeoData and noted to be deviated as much as 27° from the vertical below 4000 ft depth (1219 m) (top of the Precambrian). A decision then was made to cement high enough behind the casing to cover the prospective zones of interest. After testing one set of perforations only to determine that the well was channeled on the lower back side of the deviated casing because of the poor cement bonding, the well was then entered again and perforated for several porous intervals. The main purpose with the perforating decision in selecting several intervals was to allow for any possibility of hydrocarbon entry from one or more of the zones of interest, as seen on the mud sample log recording the gas shows. A total of ten zones were perforated (4 shots/ft) in the Precambrian and tested: 5673–5683, 5592–5596, 5540–5546, 5518–5528, 5458–5462, 5419–5425, 5222–5227, 5143–5153, 4916–4924, 4744–4754 ft (see Fig. 5) (1729.1–1732.2, 1704.4–1705.7, 1688.6–1690.4, 1681.9–1684.9, 1663.6–1664.8, 1651.7–1653.5, 1591.7–1593.2, 1567.6–1570.6, 1498.4–1500.8, 1446.0–1449.0 m).

After the attempted drillstem tests, a packer was set at the top of the Precambrian section and the perforated zones were acidized with 250 gallons (946 l) of hydrochloric acid (to eliminate any calcium residue caused by cement or mud filtrates) and then 2000 gallons (7570 l) of hydrofluoric acid (to react with any siliceous rocks), and finally 1680 gallons (6359 l) of KCl water (to wash the acids into the formation). After a few days wait, the well then was swabbed for several days and at least 10,000 gallons (38,000 l) of water were bailed from the well during this process. For those persons uninitiated to well stimulation, a swabbing operation is where a plug-like device with suction cups is lowered down on a cable to near the bottom of the hole, and then quickly pulled to the surface so that formation fluids can be sucked out of the perforated zones. The well thus is cleaned of mud, cement, and other debris that may have entered the borehole prior to cementing the production casing and other unneeded fluids that were later introduced, such as spent acid-treatment water remaining in the formation.

Water sampled from the well after several days of swabbing contained 6.29 wt% dissolved solids (Table 3). The concentration of F anion was only 45 mg/l, down from 3710 mg/l before the swabbing. Acidity was stable at 5.7 pH, as opposed to 3.0 pH before swabbing. This indicates that most of the water injected during acidization was cleaned from the well, but nevertheless, residual effects of these treatments may have significantly impacted the water geochemistry and isotopic composition.
Table 3

Chemical Analysis (by Kansas Geological Survey) of the Formation Water after Several Days of Swabbing

Chemical

mg/l

Ca

7181

Mg

445

Na

14,883

K

235

Sr

196

Fe

1150

Mg

95

Al

45

Ba

6

Pb

<1

Se

4

Zn

1

HCO3

199

SO4

137

Cl

36,963

F

45

Br (SO4 adj.)

99

I

2080

(pH = 5.76)

 

(62,865 mg/l TDS)

 
As swabbing continued, combustible gas burning with a blue flame, would sometimes precede the swabbed water coming to the surface by 5 to 10 sec. Considering the upward velocity of the swab and the volume of the tubing, it is estimated that the volume of this combustible gas was between 5 and 10 cubic ft (0.14 and 0.28 m3) per swab; water released was 4 to 5 barrels (0.64 to 0.79 m3) per swab. 1 to 2½ cubic ft of gas per barrel of formation water (0.03 m3 to 0.07 m3 gas per m3 water) is thus indicated. The gas was captured at the well head and chemically analyzed (Table 4).
Table 4

Chemical Analysis of Swab Gas (by Isotech Labs in Champaign, IL) After Days of Swabbing

Component Gas

%

Hydrogen sulfide (H2S)

0*

Helium (He)

0.508

Hydrogen (H2)

5.81

Argon (Ar)

0.574

Oxygen (O2)

8.59

Carbon dioxide (CO2)

25.3

Nitrogen (N2)

43.68

Methane (C1)

15.21

Ethane (C2)

0.322

Propane (C3)

0.0068

Isobutane (iC4)

0*

Normal-butane (nC4)

0*

Isopentane (iC5)

0*

Normal-pentane (nC5)

0*

Hexane and heavier (C6+)

0.0018

*Below detection limit of gas chromatograph (i.e., 0.001%).

It is unclear if the gas was originally in solution in the formation water or in a free state. In the weeks before any acidizing or swabbing operations began, the well pressured up to 28 psi (0.19 mPa) according to a gage at the wellhead. This build-up of pressure may have been caused by a fill-up of the well bore, but fluid levels in the well before and after the pressure build-up are unknown. During subsequent swabbing operations, the well bore was bailed and water levels typically averaged around 4600 ft (1400 m) depth at the end of the day when swabbing ceased. Overnight rise of the water level in the well bore averaged 700 ft (213 m). Accounting for the diameter of the well bore (5½ in; 13.97 cm), 13.8 barrels (2.19 m3) of water were added to the well bore each day during this typical fluid rise.

Volumetric calculations of gas and water that reach the surface with each pull of the swab (based on upward velocity of the swab (14.4 ft/sec.; 4.4 m/sec.), interior diameter of the tubing (2 7/8 in; 7.30 cm), and number of seconds each type of fluid were bailed at the surface [10 sec. for gas; 45 sec. for water]) indicate a typical pull of a swab brought approximately 27 cubic ft (765 l) of water and 6.5 cubic ft (184 l) of gas to the surface. In other terms, 600 ft (183 m) of water column was lifted with each swab in the tubing. The liters of gas can be divided by 22.4 l/mole to obtain the number moles of gas (i.e., 8.21 moles). A hypothetical molarity (M; moles/l) of the formation water, assuming all the gas came out of solution as it depressurized on its way to the surface, is 0.011 M, or 11 mM. This calculation represents a minimum value, for it assumes that the gas hitting the surface during a pull of the swab was evolved from the total amount of water brought to the surface, and in reality, perhaps only the top few feet of the swab water released this gas, if it was in solution at all.

Although formation waters are complex mixtures of dissolved gas, solute, and water, we can ask if the volume of gas per barrel of water during the swabbing operations can be accounted for by what is known about the solubilities of certain gas species in saline water. A solubility calculator for CO2 in water, available at the Midcontinent Interactive Digital Carbon Atlas and Relational database website (i.e., MIDCARB; http://www.midcarb.org/), indicates that a barrel (159 l) of 70,000 ppm saline water at 130°F (54.4°C) and 285 psi (1.97 mPa) pressure—conditions approximating the depth of a 600 ft (183 m) water column drawn up by a swab pull, assuming a hydrostatic pressure gradient of 0.476 psi/ft (3.28 Pa/m)—can hold 26 cubic ft (736 l) of CO2. Tables for CH4 solubility in water, presented in McCain (1990, p. 451–452), indicate a barrel (159 l) of formation water at such conditions can hold between 2 and 3 cubic ft (57 to 85 l) of dissolved CH4. The net conclusion is that it is feasible that the 1 to 2½ cubic ft of gas released per barrel of formation water during a pull of the swab could have come out of solution as pressure was released on the water column with the removal of water from the preceding swab. Also, as the formation water reached the surface it was observed to effervesce, thereby indicating that the formation water contained dissolved gas.

Assuming the oxygen is contributed by atmosphere that was in the well tubing and sample bottle, atmosphere was subtracted from this analysis (based on oxygen content and atmospheric ratios of atmospheric gases to oxygen) and the components recalculated to 100% (Table 5).
Table 5

Recalculated Chemical Analysis Minus Atmosphere

Component Gas

%

He

0.860

H2

9.84

Ar

0.326

CO2

42.83

N2

19.81

C1

25.76

C2

0.545

C3

0.0115

C6+

0.0030

The gas has a rather low heating value only 283 BTU/scf. Some or all of the concentration of CO2 in the sample may be spurious, because of acidization of the well and dissolution of minerals likely resulting from acidization, even though the well had been swabbed several days to unload its spent acid and its products. If CO2 is subtracted and remaining percentages are recalculated to 100% (Table 6), the BTU of the gas improves to 495 BTU/scf.
Table 6

Adjusted Chemical Analysis Minus Atmosphere and CO2

Component Gas

%

He

1.505

H2

17.22

Ar

0.571

N2

34.66

C1

45.07

C2

0.954

C3

0.0201

C6+

0.0052

Isotopic analyses were performed on chemical constituents of the water and gas (Table 7).
Table 7

Isotopic Analyses of Water and Gas from WTW Oil #1 Wilson Swab Tests

Isotope

Measure

Laboratory

Sample 1

Sample 2

Sample 3

δ13C(methane)

−57.37‰ (PDB)

  

(1)*

  

−56.0‰ (PDB)

−57.5‰ (PDB)

(2)**

δD(methane)

−347.2‰ (SMOW)

  

(1)*

  

−362.0‰ (SMOW)

−376‰ (SMOW)

(2)**

δ13C(ethane)

−52.1‰ (PDB)

  

(1)***

  

−52.7‰ (PDB)

−53.4‰ (PDB)

(2)**

δD(ethane)

 

−275‰ (SMOW)

−262‰ (SMOW)

(2)**

δ13C(propane)

 

−51.9‰ (PDB)

 

(2)**

δ13C(carbon dioxide)

−5.79‰ (PDB)

  

(1)*

  

−5.9‰ (PDB)

−7.1‰ (PDB)

(2)**

δD(hydrogen)

 

−780‰ (SMOW)

−798‰ (SMOW)

(2)**

He

0.036 Ra

  

(1)

δD(water)

−25.9‰ (SMOW)

  

(1)

δ18O(water)

−3.3‰ (SMOW)

  

(1)

δ13C(diss. inorg. C)

2.06‰ (PDB)

  

(1)

(1) Isotech Laboratories, Champaign, IL.

(2) University of Toronto Stable Isotope Laboratory.

*Analysis by chemical analysis based on standards accurate to within 2%.

**Two samples taken 1½ hr apart, on same day as Isotech Lab sample; total uncertainty carbon isotope measurements is ±0.5‰; total uncertainty in hydrogen isotope measurements is ±5‰ (see Sherwood Lollar and others, in press).

***Ethane isotopes obtained via online GC-IRMS at the University of Toronto.

Discussion

The nearest production to the Wilson well is from the Ordovician Viola Limestone and Silurian-Devonian Hunton limestones about 15 miles (24 km) to the northwest and the shallower Mississippian and Cherokee production 45 miles (72 km) to the south and east. By any standard, #1 Wilson was a rank wildcat.

Cross plots of δ13C(methane) vs. δD(methane) (i.e., D is deuterium, or 2H), and δ13C(methane) vs. hydrocarbon wetness (Fig. 6) show the CH4 at the #1 Wilson well is somewhat different compared to other conventional and unconventional (coalbed and shale) gases collected to date from the Forest City Basin, Bourbon Arch, and Cherokee Basin in eastern Kansas. Although δ13C(methane) values and the degree of hydrocarbon wetness, like other wells in the area, are broadly consistent with a thermogenic origin, the δD(methane) values for this gas are more depleted than others in the region and suggest mixing with a component of CH4 produced by microbial methanogenesis.
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig6_HTML.gif
Figure 6

Isotopic cross plots for the methane in the #1 Wilson well, compared to other eastern Kansas gases. Shape of symbol, which corresponds to “biogenic,” “mixed,” and “thermogenic” are defined from Jenden and others (1988), but the “thermogenic” and “biogenic” fields on the δ13C(methane) vs. δD(methane) cross plot are from Schoell (1988). The drastic depletion of the hydrogen isotope deuterium (δD) places the isotopic signature of the #1 Wilson methane in the realm indicated for a biogenic origin, but very near the thermogenic field

The isotopic signature of the Wilson gas is not consistent with abiogenic gases that Sherwood Lollar and others (2006) extracted from Precambrian Shield rocks in deep mines in Canada and South Africa. Normally, there is a progressive 13C and D enrichment that characterizes increasingly heavier alkanes in natural gas that is derived from thermal maturation of organic matter. Precambrian gases published upon by Sherwood Lollar and others (2006) typically exhibit a 13C depletion and D enrichment in some successively heavier alkane gases compared to methane. In contrast, the isotopic values of the #1 Wilson gas conform more to a thermogenic origin rather than an abiogenic origin.

The presence of ethane (0.322%) and propane (0.0068%) with methane (15.21%) produces a C1/(C1 + C2) ratio of 46.3 (Fig. 7). This number (also termed the “Bernard parameter;” see Bernard, Brooks, and Sackett, 1978; Whiticar, 1999) is consistent with thermogenic gas with some component of microbial CH4. CH4 solely generated by methanogenic bacteria in an environment devoid of any thermogenic hydrocarbons would be characterized by a Bernard parameter of 1000 to 10,000 (Whiticar, 1999). Mixing of a biogenic gas with thermogenic gas, or preferential microbial oxidation of CH4 in a hydrocarbon gas that originally contained trace amounts of higher molecular-weight hydrocarbons also can reduce an originally high Bernard ratio down to something on the order of 50. However, the low-sulfate concentration of 137.4 mg/l (1.4 mM) in the formation water (Table 3) indicates the present environment is reducing, so if microbial oxidation of large amounts of CH4 were to occur so as to reduce the Bernard parameter to 50, then this oxidation could not have taken place in the reducing environment indicated for the present-day formation water. Oxidation also would preferentially target the lighter isotopes. The present depleted nature of these species indicates no evidence for such oxidation.
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig7_HTML.gif
Figure 7

Bernard plot of the δ13C(methane) vs. gas wetness (expressed by methane in a ratio to the sum of the ethane and propane) of the #1 Wilson gas. Hydrocarbon gas in the #1 Wilson has an origin that is marginally thermogenic, with a possible addition of microbial methane

As noted, the water recovered at #1 Wilson (Table 7) may not be representative of pristine formation water, for it may have been contaminated by water injected when the well was acidized, and by acid reactions with various minerals near the well bore. Nevertheless, a sample of water recovered with the gas during swabbing operations and stored in a plastic bottle registered a δD of −25.9‰ and a δ18O of −3.3‰ (see Table 7). These isotopic values fall close to the global meteoric water line (see Craig, 1961).

The hydrogen-gas content of the #1 Wilson gas is notable. Even if the CO2 present in the gas is not subtracted, the gas from the #1 Wilson could contain 9.8% H2, the third-highest H2 percentage recorded in approximately 2500 gas analyses compiled for Kansas (unpubl. data by K.D. Newell). The two Kansas gases recording the most hydrogen are discussed in Coveney and others (1987). These gases were obtained from two wells—the CFA Oil #1 Scott and #1 Heins. These wells and several other wells recording H2-rich natural gases overlie the eastern margin of the Midcontinent Rift System in east-central Kansas. The study area in Coveney and others (1987) is 60 miles (95 km) west-southwest of the #1 Wilson well, across the Nemaha Uplift, which is 25 miles (40 km) west of the #1 Wilson well. The major difference between the gases discussed by Coveney and others (1987) and the gas recovered from at #1 Wilson well is the presence of significant hydrocarbon phases in the latter. Similarly, CO2 also is more abundant at #1 Wilson, but a significant fraction of this gas possibly may be a remnant of the acidization of the well or oxidized hydrocarbon gas. Argon (Ar) at #1 Heins and #1 Scott first assayed as 0, and then later in quantities as great as 1.1%. Wilson #1 Ar is 0.326%. N2 is a major constituent gas in both areas, but is more abundant at #1 Scott and #1 Heins.

Helium also is more abundant with gas at #1 Wilson (0.86 vs. <0.1%). The #1 Wilson gas nitrogen-to-helium ratio is somewhat higher (20:1) than most other eastern Kansas gases, which generally are characterized by ratios 10:1 and less (Fig. 8). Two shallow, low-BTU gases from Permian strata in Wabaunsee County 65 miles (100 km) southwest of the #1 Wilson well are the exceptions. Nitrogen-to-helium ratios of gases in the Midcontinent are relatively constant, usually differing only with the age of the reservoir (Jenden and others, 1988).
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig8_HTML.gif
Figure 8

Nitrogen and helium percentages for conventional natural gases in eastern Kansas (dark squares, data from Jenden and others, 1988) compared to that of the Precambrian gas at #1 Wilson

Coveney and others (1987) discussed numerous mechanisms that could be responsible for the H2-rich gases they discovered. They favored either the process of serpentinization of ultramafic minerals in kimberlites known in the region or in hidden ultramafic bodies, or alternatively, the process of redox reactions involving mafic minerals contained in the Precambrian basement. Other types of the H2-generating mechanisms discussed by Coveney and others (1987) include redox reactions with mafic minerals caused by influx of meteoric water along regional Paleozoic aquifers. A battery-type reaction between subsurface fluids with differing Eh was considered not likely by Coveney and others (1987) because the wells from which H2 gas was detected were then only known to be spatially confined to the region of the MRS. They ruled out the possibility of the H2 being derived from the mantle because the isotopic ratios of helium and hydrogen present in their gases had a crustal signature. This last hypothesis also can be ruled out for the #1 Wilson gas, as the isotope ratio of the He recovered at #1 Wilson, reported as a ratio Rsample/Ra (note that Ra is the present-day atmospheric 3He/4He ratio), is 0.035 indicating a predominantly radiogenic origin from continental crustal rocks. The presence of the H2 gas at #1 Wilson, well away from the MRS, may indicate that these and various other processes of H2 generation suggested by Coveney and others (1987) are possible. The dissociation of water molecules when subject to natural radiation (Lin and others, 2005) also cannot be discounted as generating the H2 gas at #1 Wilson. Sherwood Lollar and others (in press) note dissolved and free H2 gas in Precambrian Shield terranes is more abundant than previously thought.

A common origin for the various gases at #1 Wilson is not necessary, but rather some mixing of gases originating from separate processes is possible. Both hydrogen gas (H2) and hydrocarbons (methane, ethane, propane, etc.) have been suggested to form through the serpentinization reaction in which a mafic mineral, usually iron-bearing olivine, is hydrolyzed to produce serpentine, magnetite, and hydrogen gas (see Thayer, 1967). In this reaction, oxidation of Fe2+ chemically reduces the hydrogen in water to produce free hydrogen (H2), possibly even in vapor phase. Generation of abiogenic hydrocarbons from the products of the serpentinization reaction can be achieved with a long-known reaction identified as a Fischer-Tropsch synthesis (see Lancet and Anders, 1970). This reaction combines oxidized carbon compounds (usually CO2 or CO) with the aid of a catalyst (usually iron-bearing compounds or minerals, such as hematite or magnetite) in the presence of water to produce methylene groups (H2C:). In turn, these methylene groups can bond with hydrogen atoms to produce methane, or polymerize with each other and then bond with hydrogen atoms to produce higher molecular-weight hydrocarbons. The predominantly thermogenic nature of the hydrocarbon gases at the #1 Wilson well, however, argues that H2 and the hydrocarbons are produced by separate processes, and even likely in distinct geologic settings.

The most likely geologic events to have caused a major movement of formation waters in the Precambrian rocks may have occurred either in (a) Late Cambrian—Middle Ordovician time, or (b) in Late Mississippian—Early Pennsylvanian time. In the former instance, Precambrian rocks were exposed on a broad uplift called the Southeast Nebraska Arch (Lee, 1943; Merriam, 1963) located approximately 20 miles (32 km) west of the #1 Wilson well locality (Fig. 9). The basement complex was subject to an unknown amount of uplift and erosion before the crest of the Southeast Nebraska Arch was covered by siliciclastics of the Middle Ordovician Simpson Group. In the latter instance, the Precambrian basement west of the #1 Wilson well locality was again uplifted and exposed along the NNE-SSW-trending Nemaha Uplift (Merriam, 1963). This tectonic movement, which occurred in Late Mississippian—Early Pennsylvanian time, accounts for most of the present-day structural features in Kansas. Approximately 2000 ft (600 m) of throw characterizes the faulted eastern side of the Nemaha Uplift and this uplift likely stood in considerable topographic relief, for it only onlapped and covered by relatively flat-lying Lansing Group strata in Late Pennsylvanian time (Merriam, 1963). During both of these times when Precambrian basement was exposed to the west of the #1 Wilson well locality, a markedly different hydrostatic regime could have displaced pre-existing formation waters away from the wellsite. Perhaps during one of these periods of nearby exposure of Precambrian basement, a flux of fresh water derived from the nearby uplift could have coursed through fractures at the wellsite locality. Organic material washed in could have been subject to fermentation reactions to produce the hydrocarbons present, or thermogenic and microbial gases present in the overlying Paleozoic source rocks and reservoirs could have migrated, in phase or dissolved, into the Precambrian basement at the wellsite locality along fractures. Hydrogen, generated at relatively shallow depths, also could have been produced and trapped at this time, but it is a difficult gas to trap, for it is chemically active and has a small molecular diameter that makes it easy for it to diffuse though other materials.
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig9_HTML.gif
Figure 9

A, Present-day structure of the top of the Precambrian in the vicinity of the WTW Oil #1 Wilson well (after Burchett and others, 1983). Locations of the two wells discussed in Coveney and others (1987) and the #1 Poersch well (the deepest well in Kansas to date, which tested the MRS) are noted. B, Areas where Precambrian was exposed on the Southeast Nebraska Arch in Late Cambrian to Middle Ordovician time, and on the Nemaha Uplift in Late Mississippian to Late Pennsylvanian time (from Bunker and Witzke, 1988)

The relationship among the δD values determined for the water, H2, and CH4 from the #1 Wilson well also supports a decoupled origin for the gases and water in this well. Isotopically, the H2 gas at #1 Wilson (δD = −780 to −798‰, relative to SMOW [standard mean ocean water]) is depleted, and thus similar to the H2 gas recovered at the #1 Scott and #1 Heins wells (−740 to −836‰; see Coveney and others, 1987). Based on relationships outlined in Horibe and Craig (1995), the calculated value for the αH2O–CH4 value is 420. Because this value is substantially higher than the maximum theoretical value of 276 (see Horibe and Craig, 1995), it indicates that the methane is not in isotopic equilibrium with the water sampled from the well. Calculated values for αH2O–H2 and αCH4–H2 correspond to estimated geothermometer temperatures of <0 and 25°C, both well below the actual water temperature of approximately 50°C. Again, this indicates that the water, H2 and CH4 are not in isotopic equilibrium and supports the interpretation that the H2 and hydrocarbon phases may have been produced in different geologic settings and later mixed because of migration. Similarly, these results suggest that the sampled water is likely not the water in which either the hydrocarbons or hydrogen formed.

For comparison purposes, the five gases in Kansas with the greatest percentage of hydrogen are shown as pie diagrams with the analysis for the #1 Wilson gas (Fig. 10). The composition of the #1 Wilson gas, with atmospheric contribution subtracted (see Table 5), is shown in Figure 10A. Inasmuch as the CO2 in the sample may be extraneous because of acidization or partial oxidation of hydrocarbons, subtracting it (see Table 6), produces the gas characterized in diagram Figure 10B. For reasons discussed, the hydrocarbon component also may have a separate origin from the H2 component in the sample. If the hydrocarbon component is subtracted and the residual composition is recalculated to 100%, (C) is the resultant composition. This recalculated composition for the gas at #1 Wilson shows it to be similar to the two H2-rich gases discussed by Coveney and others (1987). In addition to H2 being somehow generated in the basement, these apparently uniform compositions (approximately a 2:1 ratio of N2 to H2, with minor percentages of other gases) also imply that some N2 also may be originating in the basement with the H2, but an identical origin is not necessarily implied.
https://static-content.springer.com/image/art%3A10.1007%2Fs11053-007-9052-7/MediaObjects/11053_2007_9052_Fig10_HTML.gif
Figure 10

Pie diagrams showing major component gases of the #1 Wilson natural gas: A, (see Table 5) and other gases in Kansas that have significant hydrogen. Despite cleaning out the geological formation by several days of swabbing operations, it is possible that the CO2 in the #1 Wilson gas may be the result of acidizing of the well. Recalculated component gas percentages ignoring the CO2 (see Table 6) are shown in B, Recalculated component gas percentages ignoring the CO2 and hydrocarbon gases are shown in C. The resultant hypothetical gas, which is dominated by N2 and H2, is similar to the gases recovered at the #1 Heins and #1 Scott wells (lower left) discussed by Coveney and others (1987). See text for discussion

The abundance of H2 in the #1 Wilson gas and its isotopic similarity to that recovered from the wells discussed by Coveney and others (1987) argues for a common mechanism of origin, but the geologic settings of the localities are different, for they are in two different basins separated by the Nemaha Uplift (Fig. 9). The presence of a substantial show of H2-rich gas away from known kimberlites in Kansas and mafic volcanics that dominate the fill in the MRS (see Berendsen and others, 1988), and the acidic nature of the formation water at the #1 Wilson locality militates against a serpentinization reaction generating the H2 gas at #1 Wilson, and perhaps even in the region discussed by Coveney and others (1987). Nevertheless, earlier formation water, long since gone, may have been more amenable to the serpentinization reaction.

A combustible gas partly composed of hydrocarbon components and free hydrogen, present in crystalline and metamorphic Precambrian basement along the axis of a cratonic basin in the U.S. Midcontinent, is certainly a novelty. Had the gas been expected, maybe more strictly controlled testing could have been planned and performed. Ironically though, had this been a scientific core hole rather than an unusually located deep wildcat well, the gas may never have been discovered and recovered, for a scientific test does not usually resort to swabbing operations that so typify oil and gas tests. Nevertheless, a summation of our knowledge of the chemistry of the well fluids and the local geology is perhaps the best that can be done at present. Ultimately, a question needing answering is: can this gas be commercially produced, or is it strictly of scientific value? The gas certainly needs upgrading to pipeline standards if it can be produced, but many conventional natural gases do too. Rates of deliverability to the well bore of both the formation water and the gas need to be ascertained, and a better idea of the reservoir geometry, trapping mechanism, and extent of the gas in the subsurface need to be determined. More drilling and testing may be needed to better answer these questions.

Acknowledgments

We would like to thank Frank Benson of WTW Operating Company, LLC for sharing information with us on the well. Although his contributions also were made as an author of this manuscript, the co-authors of Bill Waggoner (geologist at W.T.W. Oil Company) thank him for access to the wellsite and to other data before, during, and after the drilling of the well. Randy Van Schmus of the University of Kansas identified and confirmed the Precambrian rock identifications and provided information on the Precambrian of Kansas. We thank Kansas Geological Survey members Lynn Watney for information on the post-1983 Precambrian wells and Dana Adkins-Heljeson for sorting and printing the list from the KGS database. Pieter Berendsen and Tim Carr read a preliminary version of the manuscript and offered helpful comments and suggestions. We would like to thank P. Acker for producing the graphics.

Copyright information

© International Association for Mathematical Geology 2007