Gerry G. Calhoun
Independent Geologist
Midland, Tex.
Richard Burrows
Enseco Rocky Mountain Analytical Laboratory
Arvada, Colo.
This study points out a direct relationship between the fluorescence signature between the oil produced from a well and that found in the soils overlying that oil accumulation.
Under the right conditions, an operator can know before drilling which of the potential pays in its prospect is trapped and microseeping to the surface.
Some have suggested that half the oil that has been produced was found using visible oil seeps as indications of traps. With those obvious opportunities on production, only more subtle microseeps remain.
Instruments sensitive enough to measure these miniscule amounts of aromatic hydrocarbons have only recently become available.
Barringer Laboratories, Golden, Colo., has the instruments and provided invaluable support for the authors' research.
OIL SIGNATURES
All crude oils contain a varying proportion of aromatic compounds.
These compounds are hydrocarbons with at least a portion of their structures formed by combinations of one or more six membered benzene rings.
The distribution of these compounds is shown in relation to their fluorescent response (Fig. 1).
The unsubstituted aromatics that occur within oils include benzene (one ring), naphthalene (two rings), and phenanthrene and anthracene (each three rings with different structures). However, they more commonly occur with substitutions to the basic aromatic skeleton.
FLUORESCENCE ANALYSIS
Fluorescence analysis provides two compatible reports: 1) the relative amounts of naphthalene, phenanthrene, and anthracene, and 2) selected synchronous scans of the samples comparing them to scans of pertinent oils.
These aromatic hydrocarbons are soluble in formation water to more than 1,000 ppm and thus can be carried vertically in water solution through fractures too small to permit the movement of in-phase oil.
The size of these molecules is remarkably small, considering their complexity.
The real advantage that the aromatics have over the alkanes (ethane, propane, etc.), in geochemical exploration is that they are far less prone to be metabolized by microbes than are alkanes. There are more than 100 species of bacteria, yeasts, and molds that feed on the alkanes and only a few that metabolize aromatics.
Thus the aromatics tend to accumulate on the surface over time. In addition, they accumulate at the point of maximum vertical migration rather than in a halo position.
SITE SAMPLING
The sampling protocol in the Geochemical Evaluation Research Team (GERT) project conducted in the Permian Basin was to take seven samples at the wildcat site and three at a nearby deep dry hole.
Averaged separately, a 50-100% increase in concentration at the wildcat compared to the dry hole was indicative of commercial production. Parity in the wildcat to dry hole ratio predicts a dry hole (see table).
Ratios between wildcat and dry hole of the three compounds studied-naphthalene, phenanthrene, and anthracene -in the Hunt Oil Co. Ordovician Ellenburger discovery in Garza County, Tex., range from 6:1 to 8:1.
Background levels associated with barren conditions vary from one locality to another depending on the factors affecting vertical migration.
A sample density of 70 stations/1,500 acres is recommended. This is about the maximum number of samples that can be collected in a day and is about the minimum sized area to study.
The original fluorescence work in the GERT project used a generic combination of West Texas oils for comparison with surface samples. This report involved taking an oil sample from each of the wildcats that produced oil and comparing it to the surface samples collected earlier.
Of 18 wells drilled on 12 prospects in the GERT project, this fluorescence article will refer only to those that produced oil. They are Hunt in Garza County, Marshall R. Young in Fisher County (both commercial successes), a Robert Landreth well in Dawson County, Tom Hillin and Meridian Oil Inc. wells in Borden County, and a Primary Fuels Inc. well in Glasscock County.
SCAN FORMAT
Here is the format of the synchronous scan:
The abcissa is scaled in wave length from 270-530 manometers. The ordinate is scaled in increasing units of illumination.
The simple compounds like benzene appear as a spike on the left, with the two ring naphthalene group next. This is followed by three ring compounds like phenanthrene and anthracene.
Finally, the larger molecules appear beyond 400 nm. In essence this is like a precise measurement of the fluorescence that can be seen in an ultraviolet box.
The excitation laser is hundreds of times stronger than the light source in a UV box, and the photon counter is hundreds of times more sensitive than the human eye, but the principle is the same. This study examines surface soil samples rather than drill cuttings.
GARZA RESULTS
A sync-scan of Hunt's Ellenburger oil discovery in southern Garza County shows the trace of the intensity or concentration of the various aromatics in the oil superimposed with the trace of the same compounds in the soil sample (Fig. 2). Of course, the oil sample has been diluted to compensate for scale.
Many oils have a characteristic signature on this sync-scan display, and surface samples over an accumulation of that oil will have a similar signature.
The correlation coefficient, or similarity, of the Hunt curves, is .89. Thus the soil sample is related to the Ellenburger oil sample to a significant extent. A discovery was predicted.
BORDEN, FISHER COUNTIES
A sample from an Enserch dry hole in central Borden County revealed the contrast between soil samples from productive and barren areas.
A spike in the naphthalene group is probably due to a naphthenic base used to spray mesquite brush. Reported aromatic concentrations in the Enserch experiment are in the teens, while the Hunt experiment recorded 94 intensity units of anthracene and 204 units of phenanthrene.
Sampling at the Fisher County Palo Pinto discovery showed 306 units of phenanthrene and a strong correlation coefficient between oil and soil sync-scans (Fig. 3).
This well should ultimately recover more than 100,000 bbl of oil and has been offset successfully. A discovery was predicted.
DAWSON EXPERIMENT
The Landreth experiment in Dawson County resulted in a 9 b/d oil discovery from Permian San Andres that depleted quickly.
The sync-scan comparing the San Andres oil to the soil sample shows a correlation coefficient of .75. The average phenanthrene concentration is 274 units. A discovery was predicted.
This points up one limitation of using fluorescence analysis. Predicting volumes is beyond the scope of this method at present.
The intensity of the surface signature is controlled by the degree of fracturing in the overburden, the composition of the overburden, pressures in the reservoir, the percentage of the oil that is aromatic, and other factors.
The volume of oil trapped in the reservoir is only one of many variables that influence fluorescence response.
Comparing the Landreth soil samples to a syn-scan from a nearby Spraberry oil well revealed that the Spraberry is not trapped in this locale. The "fingerprint" of the soil samples showed little relationship to the Spraberry oil.
BORDEN COUNTY TESTS
The Hillin wildcat in Borden County was potentiated for 16 b/d of oil and 125 b/d of water from Mississippi limestone and is a marginal producer (Fig. 4).
The actual concentrations of the aromatics are lower than the controlling dry hole, and a dry hole was predicted.
Meridian's Strawn oil discovery in Borden County is 1 mile south of the Hillin well. The Meridian soil sample displays many of the characteristics of the Hillin soil sample rather than the marginal Pennsylvanian oil discovered here (Fig. 5).
Based on the intensity of the aromatic values, the well was predicted to be dry. Meridian potentiated it for 14 b/d of oil and 12 b/d of water.
GLASSCOCK FUSSELMAN
Primary Fuels drilled a seismic controlled Siluro-Devonian Fusselman trap in Glasscock County and made a 6,000 ft plugback completion in Permian Abo for 99 b/d of oil and 2 b/d of water.
The average Abo well makes about 10 000 bbl of oil in this area.
Aromatic concentrations were low, and a dry hole was predicted.
Comparing a Fusselman oil sync-scan to a soil "fingerprint" in the Primary Fuels experiment shows no relationship (Fig. 6). However, the Abo oil sync scan shows a very high correlation coefficient (Fig. 7).
This indicates that the operator could anticipate completion in Abo rather than Fusselman.
These experiments show conclusively that the trapped oil is most influential in determining the soil signature. Now for the first time, the fluorescence method can predict not only producer or dry hole with confidence but also anticipate which formation contains trapped oil.
GERT SUMMARY
Fluorescence experiments in GERT accurately predicted both commercial successes with pays as deep as 8,500 ft.
They also picked the Landreth, Sandia, and Encino Petroleum Co., San Antonio, experiments as producers for a 40% success rate.
The wildcats predicted to be dry were all dry or marginal producers for an accuracy of 100%.
The overall predictive accuracy, considering marginal wells as dry holes, is 11 correct from a total of 14 experiments, or 78.57%. This compares favorably with the industry average of one discovery of any amount of oil in seven wildcats, or 14% success.
Since there are only 14 iterations of the experiment, Fisher's Exact Test is the only statistical analysis that can be conducted to assess the significance of the data field generated by the study.
Applying the test resulted in a .01 evaluation. This means that the probability that the table could be derived from random data is 100:1 and that there is a strong relationship between prediction and result.
MORE RESULTS
Eleven of the wildcats in GERT were controlled with modern seismic data.
Of that group, two resulted in commercial discoveries, or a predictive accuracy of 18%. At one tenth of the cost and four times the accuracy, fluorescence analysis should be used as a screening device before using more expensive methods and before lease acquisition.
Correlation coefficients may be used as an exploration tool (Fig. 8). A stratigraphic trap produces at the location indicated, and well control is too sparse to provide help in locating the sand body.
When the produced oil is compared to soil samples with sync-scans, a band of high correlation coefficients develops. The trend of these high correlations parallels the trends of other fields producing from this interval. This discovery will be offset in the near future.
These data can help today in predicting potential pays in a multipay prospect.
For example, there is no point in drilling to the Ellenburger when the signature observed on the surface is San Andres in origin. Likewise there's no point in drilling at all unless a signature of oil is present on the surface.
Data collected over a multipay producing area are generally quite strong but are not diagnostic of any individual pay in the field.
BIBLIOGRAPHY
Brennan, M.C., Smith, P.Y., The Chemical Relationships Between Crude Oils and Their Source Rocks, Habitat of Oil, AAPG, 1958, pp. 818-49.
Brooks, James M., Kennicutt, Mahlon C., Offshore Surface Geochemical Explorations, OGJ, Oct. 20, 1986, pp. 66-72.
Calhoun, G.G., How 12 geochem methods fared in GERT project, OGJ, May 13, 1991, pp. 62-68.
Eastwood, DeLyle, Use of Luminescence Spectroscopy in Oil Identification, Chemistry Branch, U.S. Coast Guard R&D Center, Avery Point, Conn. 06340
Jones, T.S., Smith, H.M., Relationship of Oil Composition and Stratigraphy in the Permian Basin, in Fluids in Subsurface Environments, AAPG Memoir 4, 1964, pp. 101-244.
Riecker, Robert E., Hydrocarbon Fluorescence and Migration of Petroleum, AAPG Bull., Vol. 46, No. 1, 1962, pp. 60-75.
Hebert, Carroll F., Geochemical prospecting for oil and gas using hydrocarbon fluorescence techniques, in Unconventional Methods Symposium III, Institute for the Study of Earth & Man, Southern Methodist University, 1984, pp. 40-58.
Copyright 1992 Oil & Gas Journal. All Rights Reserved.
