Comparison of the Direct-Inject Method and the Pangaea Gas-Sieve Method
of Determining Soil Vapor Hydrocarbon Concentrations
Above the Cunningham Field, Kingman County, Kansas

OBJECTIVE:  The Cunningham Project was performed at the request of Phillips Petroleum.  Phillips designed the survey and it was conducted by Pangaea.  The survey goals were to compare results of the commonly used "direct-inject" method of soil vapor analysis and Pangaea's new method that allows the concentration of soil vapor hydrocarbons in the Pangaea Gas-Sieve to determine which method best delineated the Cunningham Field.

1. INTRODUCTION

In 1931, the Cunningham Field was discovered, utilizing conventional geology, with the drilling of the Skelly #1 Miles into the Pennsylvanian Age Lansing/Kansas City Formation at 3350' BGL.  Both oil and gas were produced from this zone. Additionally, commercial quantities of "dry" gas at 2005-2020' BGL and at 2133-2142' BGL were produced from two shallow Permian zones within this well.  Four wells, below the Lansing Kansas City Limestone, each producing large quantities of "wet" gas, were drilled the following year. The "wet " gas was produced from the Viola, Simpson and the Arbuckle Dolomite (Map 1).

Well data suggests that the Cunningham Field is an anticlinal fold, faulted on three sides by high angle normal faults, which throw the Mississippian Osage Limestone or Mississippian Kinderhook Shale against the Orodvician Age Viola Formation (Figure 1).  The major oil and gas reservoir of this field is located in the up thrown block of this structure. The field was expanded in subsequent years and produced from 1931-1978. In 1978 the field was converted to a natural gas storage unit with gas piped in from various sources and injected into the unit for later withdrawal and transmission. The "heart” of the storage unit occurs in Sec. 25 -T27S-R11W and Sec. 30-T27S-R10W.

A gas sample taken from the final Viola gas well producing in the Cunningham Field was analyzed in 1977.  The results of the analysis occur in Table 1.  Gases of two or more carbon length comprises about 22% of this “wet” gas including N-butane at 2.63%.   

 Map 1:  Cunningham Field mapped on the top of the Lansing Formation below:

Cross-Section of the Cunningham Field:

Table 1.  Results of the analysis of the "wet viola" natural gas from the A.J. Conally #2 on 7/21/77 (provided by Storage Unit Operator).

Since 1978, stripped pipeline gas has been repeatedly injected for storage and withdrawn from the storage unit as a result of consumer demand. The pipeline gas was gathered from various sources and transmitted to the storage unit. The composition of the mixed gas stream withdrawn from the storage unit for transmission on 1/15/97 was provided by the storage unit operator (Table 2).  Note that 7.78% of the gas stream is composed of two or more carbon units, but N-butane represents only 0.07% of the gas.  Currently, the gas in the storage unit is a "dry" gas.  The composition of the shallow Permian gas, occurring both on and off the field is provided in Table 3.

Table 2.  Analysis of the "dry" storage gas from the unit on 1/14/97 (Unit Operator).

2.   Methods of Investigation

Two soil vapor samples were collected at each sample location. One was collected using gas-sieve and the other was taken using standard "direct inject" sampling procedures.

The gas-sieve sampling method employs the Pangaea Gas-SieveÔ and a Geoprobeâ soil gas sampling system.  The Geoprobe's hydraulic hammer is used to insert 1" hollow steel rods to 8' BGL or, if in the Ninnescah River Valley, just above the capillary fringe.  A section of polyethylene tubing was lowered into the Geoprobeâ rods after total depth was reached. The tubing was thereafter screwed into the advancing front of the sampling apparatus, at 2" from the advancing front, using a stainless steel adapter and o-ring to facilitate a tight seal.  At the surface, the polyethylene tubing was attached to the vacuum-volume pump, using a section of Tygon tubing.  The vac/vol pump is used to remove atmospheric contamination by purging 1/4 liter from the tubing. The line pressure is allowed to return to zero, broken apart at the Tygon connection, and the gas-sieve was attached in-line, using two Tygon tubing sections.  One liter of soil vapor from each location was drawn across the Gas-Sieve by the vacuum-volume pump.  After sampling, the gas-sieve was disconnected from the line and delivered to the lab for analysis.

The "direct inject” sampling was accomplished using the Geoprobe system to pull soil vapor to the surface in the tubing after the 1/4 liter purge. A syringe was inserted into the tubing and 10 ml of soil vapor was taken directly from the line, transported to the filed lab and 1ml was analyzed within 2 minutes of collection. Concentrations of the hydrocarbons was determined by the same analytical method as were the Gas-Sieves.

The mobile gas chromatography lab was equipped with a Hewlitt-Packard 5890 GC using an FID detector.  The analytes were methane (C1), ethane (C2), propane (C3), butane (C4).  Each component of the analytical equipment was computer linked to insure a standard operating procedure, free of operator error.

3.  Findings and Conclusions

The raw butane data, not transformed by statistical manipulation, obtained from both the analysis of the "direct inject " samples and the gas-sieve samples, was placed on a map of the study area (Map 2).  Butane was mapped because the concentrations were high in the original Cunningham Field gas (2.63%), low in the stored pipeline gas (0.07%) and low in the shallow Permian gas (0.62%).

Using the "direct inject" method, butane concentrations in and around the storage unit ranged from 0-3.1 ug/L, suggesting that the storage unit "does not exist”.  The "direct inject" method fails to define the storage unit because the volume of soil vapor sample that can be introduced into the GC is very small and thus the hydrocarbon mass in the sample is small.  The small injected sample volume is an unavoidable problem with a GC and the concentrations detected soil vapor hydrocarbons are often in the "noise" range of the GC.  Thus "data" and "noise" look the same.

The small volume problem becomes particularly important in geological settings such as at the Cunningham Field because there is a 400' thick NaCl bed between the reservoir and the surface.  This thick salt is thought to "dampen" or reduce the hydrocarbon signal.  Low "signal to noise" ratio has been the major source of failure for soil vapor geochemical prospecting.  Other instances of signal depression are depth of reservoir as very deep reservoirs have dampened light hydrocarbon signatures, as well.

Utilizing the gas-sieve methodology, butane in and around the storage unit ranged from 10-735 ppb. Anomalously high butane concentrations were found to occur above the storage unit, while low background levels were found to occur outside the storage unit. The storage unit is well defined by the gas-sieve method. The gas-sieve overcomes the "loss of hydrocarbon signal" which generally is associated with the dampening effect of thick subsurface salt by sieving, concentrating and preserving the light hydrocarbons stripped from a liter of soil vapor.  After sampling, the gas-sieve transported to the lab and thermally desorbed into the GC. The hydrocarbons from a liter of soil vapor are carried into the GC by a stream of N2 carrier gas.  The geochemical signature is enhanced approximately 1000 times by this method bringing the data well above the noise of the GC.  Clearly, the gas-sieve method is superior to the "direct-inject" method.

Information obtained from the storage unit operator suggests that variations in the butane values inside the boundaries of the storage unit are explained by known permeability variations across the anticline. Observation well pressure data at the storage unit suggests that the south 3/4 of the storage unit has less permeability and stores less pipeline gas.  The majority of the pipeline gas is stored in the north portion of the unit.  Anomalous butane values outside of the storage unit boundaries are explained by fault locations, occurrence of reservoir outliers charged by the faults, and other oil and gas production, near but unrelated to, the storage unit.