INTRODUCTION

Except for a few papers about Prudhoe Bay field, little has been published pertaining to oil-source rock relationships on the Alaskan North Slope. Sources for the Prudhoe Bay accumulation were implied in early publications by Morgridge and Smith (1972), Jones and Speers (1976), and Young et al (1977). However, no direct geochemical oil-rock correlations were reported until Seifert et al (1979) concluded from biological marker comparisons that Prudhoe Bay oil was co-sourced by the Triassic Shublik, Jurassic Kingak, and Cretaceous post-unconformity shales. Essentially no geochemical data were presented on oils other than Prudhoe Bay until Magoon and Claypool (1981) reported on results from 40 North Slope oils. They concluded that two major oil types, the Barrow-Prudhoe and the Simpson-U iat, are present in the basin with both types recognized over wide areas. They did not attempt to identify the source for either oil type.

To learn more about oil-source relationships across the North Slope, Magoon and Claypool initiated a cooperative study among a group of geochemistry laboratories, all of which have some expertise in oil-rock correlations. Each laboratory was provided with 9 oil samples from various North Slope locations and 15 core samples from 8 wells (Fig. 1). Sample descriptions are provided in Table 1. The primary objectives of this study were to answer the following questions:

1. How many genetically unrelated oil types are represented by the 9 oil samples?

2. What are the most likely source horizons for the oil types found?

METHODS

The methods listed in Table 2 were used for characterizing oil types and making oil-rock correlations. All of these methods have been reported in the literature (Tissot and Welte, 1978; Hunt, 1979), but some are not commonly employed by many geochemical laboratories. As has been

End_Page 165------------------------

the case in other regional oil correlation attempts, we found that not all methods can be equally applied. When undertaking such a study, it is invariably true that some techniques do not make sufficient distinction or show too much variability to be useful even though each method has been successfully applied in at least one other area. Also, inclusion of a few severely biodegraded oils in the suite made some methods less useful than they normally might be. The methods found to be most definitive for North Slope samples were GC-MS, whole oil GC, isoprenoid distribution including pristane/phytane, carbon isotope ratio, optical rotation, and vanadium-nickel analysis. Light ends analysis, infrared spectroscopy, aromatic compound distribution, sulfur compound distribution, and sulfur iso ope ratios did not provide information useful in distinguishing among 9 oil samples provided. Rock extracts used for oil-rock correlations were obtained by Soxhlet extraction of pulverized rock samples with methylene chloride. Solvent was removed by roto-evaporation. Vitrinite reflectance and elemental analysis measurements were made on kerogens isolated by HCl-HF demineralization.

Some core samples with low maturity were subjected to hydrous pyrolysis to determine the characteristics of oil derived from more mature equivalent sections. Procedures used were generally similar to those described by Lewan et al (1979).

OIL-OIL CORRELATION

Carbon Isotope and Optical Rotation Data

Williams (1974) showed that carbon isotope ratios and optical rotation values, when cross-plotted, helped to distinguish principal oil types found in the Williston Basin. The same technique was employed by Momper and Williams (1979) to characterize various secondary processes affecting oil composition in the Powder River Basin. Figure 2 is a similar display for the 9 oils in this study. For comparison, the range of values determined for 15 oils from Prudhoe Bay field is outlined. These oils had been analyzed previously at Amoco Production Company Research Center. Saturate fraction carbon isotope ratios are shown on the

Fig. 1. Oil and rock sample locations, Alaskan North Slope.

End_Page 166------------------------

linear vertical scale, while specific rotations of the same fractions are plotted on the logarithmic horizontal scale. The distribution of points suggests that the Fish Creek, Dalton, and Barrow oils cannot be distinguished from Prudhoe Bay oil. The two Simpson oils and one from Umiat field appear to be related to one another but distinct from the Prudhoe Bay type oil. The Seabee oil is somewhat anomalous. Its position on the plot suggests that it could be either a thermally matured Umiat oil or an unrelated third type.

Gas Chromatography of Oils

Whole oil GC is commonly used for oil characterization. In addition to correlation by pattern comparison, it also provides useful information about secondary alteration processes such as biodegradation, thermal maturation, and fractionation. Also, the presence of volatile contaminants is usually revealed by GC. Chromatograms of the 9 North Slope oils are shown on Figure 3.

The lack of n-paraffins and isoprenoids on chromatograms for the Fish Creek and Simpson Seismic Line samples

Table 1. Sample descriptions, Alaskan North Slope.

Table 2. Oil characterization methods used on North Slope oils.

End_Page 167------------------------

Fig. 2. Plot of carbon isotope ratio (¹³C) vs. optical rotation (specific rotation) for aliphatic fractions of North Slope oils. O.R. is plotted on logarithmic scale and ¹³C on linear scale. O.R. was measured using Hg 5460 A source. ¹³C is relative to PDB standard.

Fig. 3. Capillary gas chromatograms for North Slope oils. Numbered peaks indicate n-paraffin carbon numbers. PR = pristane; PH = phytane.

End_Page 168------------------------

Fig. 3. Continued.

End_Page 169------------------------

indicates that these oils have been severely biodegraded (Winters and Williams, 1969). Low API gravities for these oils are also consistent with biodegradation. Although the Dalton sample contains some n-paraffins, it also appears biodegraded. This oil apparently contains no normal paraffins above n-C₂₀ yet shows a large envelope of unresolved material in the n-C₁₀ to n-C₃₀ range. These observations suggest biodegradation; however, all recognized n-paraffins are in the C₈ to C₂₀ range, which is unusual for a biodegraded crude. One possible interpretation is that the reservoir was charged twice, first with a normal crude oil that was severely degraded and later by a light crude oil that has not been degraded.

The South Barrow No. 20 oil contains an atypical hydrocarbon distribution with an anomalously large concentration of material in the C₁₁ to C₁₆ range. Such distributions usually indicate the presence of contamination by hydrocarbons in the diesel oil or kerosene boiling range. Another possible explanation for the composition of South Barrow No. 20 oil is that a second migration of very light oil moved into an accumulation of normal range oil followed by a loss of light components from the mixture. There are not sufficient data to determine which, if either, of these two possibilities actually occurred but there was no record of any diesel being used during recovery procedures in this well. It is possible that the sample was contaminated after collection, but the ikelihood of such contamination is not great. If this sample is contaminated, it makes questionable any correlation based on parameters that include the diesel oil hydrocarbon range.

Oil from the Seabee well, as suggested by its high API gravity, is composed of relatively low-boiling-point compounds. However, its relatively high pristane/n-C₁₇ ratio (0.63) does not support the conclusion that it is a thermally degraded oil. The high pristane/n-C₁₇ ratio is more consistent with an interpretation that the Seabee oil is composed of fractionated light components from a heavier (lower API gravity) oil. This suggests that the higher (less negative) isotope ratio for this oil is not attributable to thermal degradation and that the Seabee oil (produced from the Torok Formation) may be genetically unrelated to Umiat type oils (produced from the Nanushuk Formation).

Isoprenoid Distribution

A technique that has proved useful for identifying related crude oils is that of plotting normalized relative abundances of C₁₃-C₁₆ plus C₁₈-C₂₀ isoprenoids versus carbon number (Clayton and Swetland, 1977; Ross, 1980). The resulting distribution patterns often show remarkable similarity between related oils over very large geographic areas. Isoprenoid distributions, determined for six of the oils in this study, are shown in Figure 4. The Simpson Seismic Line and Fish Creek oils are too severely degraded for this type of analysis to be useful. The Seabee oil shows such a rapid decrease in components in this range that its isoprenoid distribution is so distorted as to be misleading.

The isoprenoid patterns (Fig. 4) confirm the similarity between Umiat and Simpson oils. The distribution for South Barrow No. 20 oil may not be meaningful because of the suspected contamination or mixing as mentioned above. The South Barrow No. 19 and Dalton oils appear significantly different from Prudhoe Bay oil. Pristane/phytane ratios (Fig. 4) are more consistent, with Umiat and Simpson oils almost identical and Barrow-Prudhoe oils in a close grouping. Although pristane/phytane ratios determined for Umiat and Simpson oils are not particularly large, their higher values relative to those for the Barrow-Prudhoe oils suggest that Simpson-Umiat oils were derived from a source or sources that contain greater amounts of terrestrially derived organic matter.

Vanadium and Nickel

Abnormally high concentrations of vanadium and nickel have been noted in some organic-rich shales (Vine and Tourtelot, 1970) and crude oils (Yen, 1975). Investigations into the chemistry of these metallo-organic complexes suggest that they have high thermal stability and are resistant to microbial degradation, water washing, and weathering. These properties make them ideally suited for use as a parameter in oil-oil and oil-source rock correlation studies (Lewan, 1980; Lewan and Maynard, 1983). V/(V + Ni) ratios determined for 7 oils in this study are summarized in Table 3. The Seabee oil was too light to contain detectable amounts of vanadium or nickel while the Umiat sample has not yet been analyzed.

Examination of the data in Table 3 suggests that the 7 oils can be divided into two groups. Prudhoe Bay, South Barrow No. 19, Fish Creek, and Dalton oils have similar ratios of V/(V + Ni), while the Simpson Seismic Line, Simpson Core Test, and South Barrow No. 20 oils make up a group with lower values.

Gas Chromatography-Mass Spectrometry

GC-MS analysis is rapidly becoming the primary method for oil-oil and oil-rock correlation studies because of the vast amount of data that can be generated from one analysis. Unfortunately, it is not possible to compare visually all of the mass chromatograms the instrument is capable of generating. Therefore, based primarily on experience gained in other areas, we selected and monitored 20 ions that have provided useful correlation information. Of these, m/z 169 (tri- and tetrasubstituted naphthalenes) and m/z 123 (triterpanes) appeared to be the most definitive for oil-oil correlation on the North Slope.

A generally high relative abundance and insensitivity to both biodegradation and thermal degradation are primary factors allowing m/z 169 to be particularly useful in oil correlation. Figure 5 shows m/z 169 mass chromatograms for 8 of the 9 oils analyzed in this study. Five peaks representing unidentified compounds used to characterize these oils are shaded. The relative heights of peak 5 compared to peaks 3 and 4 seem to be significant for both

End_Page 170------------------------

oil-oil and oil-rock correlation. Mass chromatograms m/z 169 for the Prudhoe Bay, Fish Creek, and Dalton oils appear quite similar as do those of the Umiat and Simpson oils. South Barrow Nos. 19 and 20 oils are noticeably different from both Prudhoe Bay and Umiat oils. There is also considerable difference between the two South Barrow oils, particularly in the lower range (left side) of the m/z 169 mass chromatograms. This could be due to contamination or mixing. The mass chromatogram of the Seabee oil reflects a higher concentration of tri- relative to tetrasubstituted naphthalenes, consistent with the low-boiling range of this oil. However, taking this fact into account, the Seabee oil does not seem to match any of the other oils. The relatively high peak 1 suggests similarity to Um at oil, while a low peak 5 relative to peaks 3 and 4 is more comparable to Prudhoe Bay type oil.

The m/z 123 mass chromatograms (Fig. 6) show close similarity among the Prudhoe Bay, Fish Creek, Dalton, and South Barrow No. 19 oils, but the South Barrow No. 20 oil is significantly different. Again, this could be due to contamination or mixing. The Umiat and two Simpson oils are similar to one another, but the Seabee oil is again somewhat different. The latter has a relatively large peak 1 in common with the Simpson and Umiat oils. The Prudhoe Bay and related oils have a much smaller peak 1 and larger peak 3 than the Simpson-Umiat oils.

Conclusions on Oil-Oil Correlation

Conclusions from oil characterization data are summarized on Figure 7. Two genetically unrelated types

Fig. 4. Isoprenoid distributions for North Slope oils. Vertical scale shows normalized percentages of C₁₃, C₁₄, C₁₅, C₁₆, C₁₈, C₁₉ (pristane), and C₂₀ (phytane) isoprenoids. PR/PH = pristane/phytane ratio determined from peak heights on Figure 3.

Table 3. Vanadium-nickel analysis, Alaskan North Slope oils.

End_Page 171------------------------

Fig. 5. Mass chromatograms of m/z 169 (diagnostic for tri- and tetrasubstituted naphthalenes) for North Slope oils. Five shaded peaks (compounds not identified) are the most diagnostic for oil correlation.

End_Page 172------------------------

can be recognized, but some oils show modifications sufficient to have them classified as subtypes. The Prudhoe Bay and Fish Creek oils are very similar and seem to represent the basic Barrow-Prudhoe (B-P) type reported by Magoon and Claypool (1981). The Dalton oil is generally similar to Prudhoe Bay oil but shows minor differences in isoprenoid distribution. It is therefore designated as a Barrow-Prudhoe subtype (subtype 1). The Barrow oils show similarities to the Barrow-Prudhoe type but have some noticeable differences in GC-MS (ion 169) and isoprenoid distribution. They are concluded to be a second subtype of Barrow-Prudhoe oil (subtype 2). There are some differences between the two Barrow oils, but most of these differences can be attributed to the suspected added components in S uth Barrow No. 20 oil. However, the very low V/(V + Ni) determined for the South Barrow No. 20 sample cannot be attributed to contaminant and raises the possibility that this oil represents still another B-P subtype.

Fig. 6. Mass chromatograms of m/z 123 (diagnostic for triterpanes) in 13 to 16 carbon number range for North Slope oils. Four numbered shaded peaks (compounds not identified) are the most diagnostic for oil correlation.

End_Page 173------------------------

The Umiat and two Simpson oils are quite similar and are designated as the Simpson-Umiat type reported by Magoon and Claypool (1981). The Seabee oil is significantly different in composition from the Simpson-Umiat type oils. Some of these differences can be attributed to the predominance of low molecular weight compounds in the Seabee oil believed to have resulted from fractionation of a full-range oil. It is difficult to establish whether the Seabee oil should be considered a separate genetic type or a subtype of Umiat type. Based on geological considerations, source rock characteristics, and oil-rock correlation data (discussed below), the Seabee oil is assigned as a Simpson-Umiat subtype (subtype 1), but this assignment is tentative.

OIL-ROCK CORRELATION

Obtaining suitable oil-rock correlations in any area is inherently more difficult than making oil-oil correlations. The difficulty is due to several factors such as: (1) more natural variability in composition of rock extracts than of oils; (2) the need to have source rock samples in a relatively narrow range of thermal maturity to be useful for correlation, relation, and (3) the small quantity of extractable bitumen being insufficient for some measurements to be made with reasonable accuracy. As a result of these limitations, the conclusions drawn from oil-rock correlation data are less definitive than those drawn from the oil-oil correlation data.

Source Rock Evaluation

Before attempting to correlate the various oil types with available core samples it was important to establish hydrocarbon generating capability and level of thermal maturity of each rock sample. Organic carbon content (Fig. 8) was high in most samples, suggesting that the Torok, Neocomian pebble shale, Kingak, and Shublik formations are all possible hydrocarbon sources. Vitrinite reflectance analyses indicate that five of the samples (011, 013, 014, 015, 021) are matured beyond the oil window. Extracts from

Fig. 7. Locations and oil type designations for the 9 North Slope oils analyzed. Numbers in parentheses represent subtypes of the two major oil types identified.

End_Page 174------------------------

these samples were not suitable for oil-rock correlation. All other samples are in or below the oil window and should provide useful data for oil-rock correlation.

Extract quantities from the entire suite of samples were quite low (Table 4), especially relative to organic carbon content. For the five overmature samples, such small amounts of extract were expected, because most of the bitumen originally generated would have been expelled or thermally destroyed. However, the small extract quantities from those samples in the oil window are surprising, especially considering that the kerogens consisted of mostly amorphous or a mixture of amorphous and structured material. Only Kingak sample 023 contained a significant amount of extractable bitumen, and it is still relatively lean for an oil-prone source rock near the stage of maximum oil generation. In addition, generating capability determined by Rock-Eval and elemental H/C ratios (Table 4) are bo h unusually low for oil-prone kerogens.

These results immediately raise the question of whether any of the 15 cores analyzed actually represent effective oil source rocks, even though they contain sufficient organic matter. It appears that the kerogens in these sampled intervals have undergone some form of oxidative degradation. It is not possible to determine whether such degradation would have occurred contemporaneously with deposition, or at a later time by encroachment of oxygen-containing surface water. Kerogen oxidation could have been significant during a period of uplift and erosion such as that which produced the Neocomian unconformity.

Despite the apparent deficiencies in this group of samples, we thought they might give some clues about oil-source rock relationships on the North Slope. Some of the same intervals sampled for this study apparently are effective sources for the oil in Prudhoe Bay Field (Seifert et

Fig. 8. Organic carbon (wt%) and vitrinite reflectance values for 15 core samples from NPRA wells shown on generalized northwest-southeast cross section of 8 sampled wells. Oil window occurs approximately in 0.6 to 1.5 range of vitrinite reflectance.

End_Page 175------------------------

al, 1979). The small amounts of bitumen in these samples may still be representative of extracts from richer lateral equivalents that have been oil sources.

Gas chromatograms of the rock extracts are shown in Figure 9. As would be predicted, GC traces of the five very mature samples (011, 013, 014, 015, 021) do not resemble traces for crude oils. Also, bitumen from the Torok sample, Seabee well (012), was unlike crude oil, even though this sample was indicated by vitrinite reflectance to be within the oil window. Torok samples 020 and 022 and Neocomian pebble shale sample 010 yielded immature-appearing extracts consistent with their low R₀ values. The large peak appearing between n-C₂₇ and n-C₂₈ on many of the extract chromatograms is believed to be a contaminant.

The six extracts that are at a maturity level suitable for GC correlation (samples 009, 016, 017, 018, 019, 023) do not provide much basis for identifying a specific source for either the B-P or S-U oil types. Kingak sample 023 most closely resembles Prudhoe Bay oil (sample 001) in overall hydrocarbon distribution but has a distinctly higher pristane/phytane ratio (2.57 vs. 1.41). Shublik sample 018 is the only "oil-window" extract with a pristane/phytane ratio close to B-P type oil (1.46 vs. 1.33-1.45). It is significant that all Cretaceous samples (Torok and Neocomian pebble shale) have pristane/phytane ratios greater than 2. Higher ratios indicate an increased input of terrestrially derived organic matter, and oils sourced from such horizons should reflect that characteristic. Umia oils have pristane/phytane ratios (2.07 and 2.25, Fig. 4) more consistent with ratios determined for Cretaceous rock extracts.

Two Torok and two Neocomian pebble shale samples (016, 020, 010, 019) were subjected to hydrous pyrolysis at 350°C for 3 days to mature them to the stage of maximum oil generation. Gas chromatograms of the pyrolysates and original extracts are shown on Figure 10. These data suggest that only one Torok sample (016) yields a pyrolysate with a GC pattern similar to that for Prudhoe Bay type oil; however, it has an appreciably higher pristane/phytane ratio (3.27 vs. 1.46) than the oil. The other three pyrolysates, in addition to having high pristane/phytane ratios, contain greater proportions of long-chain paraffins (>C₂₀) than observed in any of the oils.

Gas Chromatography-Mass Spectrometry

Of the correlation parameters applied, m/z 169 mass chromatograms provide the most definitive oil-rock relationships (Fig. 11). Especially significant is peak 5, which also was diagnostic in characterizing the oils. All Torok extracts have a large peak 5 relative to the other extracts. Also, all Neocomian pebble shale samples except number 17 have a peak 5 appreciably higher than do any of the oils. Kingak and Shublik samples 023 and 018 have distributions resembling some of the oils, especially with regard to the relative proportions of peaks 1, 3, and 5. Barrow-Prudhoe oils (left side of Fig. 5) seem to be intermediate between the Kingak and Shublik samples while Simpson-Umiat oils (except for Seabee oil) are more intermediate between Neocomian pebble shale and Kingak extracts. By t is measurement, the Seabee oil appears similar to Shublik sample 010 extract. However, this similarity must be considered coincidental because other considerations rule out any correlation.

There does not seem to be any relationship between maturity and size of peak 5 on m/z 169 mass chromatograms. A comparison was made between m/z 169 mass chromatograms before and after hydrous pyrolysis of Torok and Neocomian pebble shale samples (Fig. 12). Peak 5, although slightly reduced, remains quite high. Only pebble shale sample 009, among all the Cretaceous samples analyzed, seems to show any relationship to any of the oils.

Table 4. Source rock evaluation data on North Slope cores.

End_Page 176------------------------

Fig. 9. Capillary gas chromatograms of methylene chloride extracts from 15 NPRA core samples. Numbered peaks represent n-paraffin carbon numbers. PR = pristane, PH = phytane. PR/PH ratio was determined from peak heights on chromatogram.

End_Page 177------------------------

Fig. 9. Continued.

Fig. 10. Capillary gas chromatograms of methylene chloride extracts from aliquots of four immature NPRA core samples compared to extracts obtained from second aliquots of the same cores after being subjected to hydrous pyrolysis at 350°C for 3 days. Pyrolysates from samples 016, 022, and 019 suffered light end loss during removal of extracting solvent. Sample 010 pyrolysate was recovered without extraction.

End_Page 178------------------------

Fig. 11. Mass chromatograms of m/z 169 for methylene chloride extracts from 9 NPRA cores. Six core samples did not yield sufficient extract to get meaningful data. Numbered shaded peaks correspond to those on Figure 5.

End_Page 179------------------------

Fig. 12. Mass chromatograms of ion 169 for methylene chloride extracts from four immature NPRA core samples compared to extracts after hydrous pyrolysis of cores at 350°C for 3 days. Numbered shaded peaks correspond to those on Figure 5. Peak 5 is not significantly reduced by the increased maturity effected by hydrous pyrolysis.

End_Page 180------------------------

The m/z 123 mass chromatograms (Fig. 13) do not show any consistent relationship between oils and rock extracts. None of the samples give the high peak 3 observed in Barrow-Prudhoe oils (Fig. 6). The Torok and Neocomain pebble shale samples generally give a sizable peak (a) not observed in any of the oils. The high peak 1 observed in Simpson-Umiat oils and especially in the Seabee oil occurs only in the Kingak and Shublik samples. Actually, the closest similarity is between the Kingak extract and the Simpson-Umiat oils. Hydrous pyrolysis does not help in this case because the compounds giving the 123 ion are thermally sensitive and the abundances are too greatly reduced to give usable mass chromatograms.

Some of the parameters used successfully in characterizing the 9 oils could not be used on the rock extracts. There was not enough extracted material to obtain optical rotation or to get accurate vanadium-nickel analysis. Carbon isotope ratios of the extracts were not diagnostic, as was the case with C₁₃ to C₂₀ isoprenoid distributions, although pristane/phytane ratios were meaningful.

Conclusions on Oil-Rock Correlation

For reasons already mentioned, it is not possible to draw firm conclusions from oil-rock comparisons with this suite of samples. However, the data suggest that Barrow-Prudhoe type oils are co-sourced for the most part by the Kingak and Shublik shales. A contribution from Neocomian pebble shale in some areas cannot be ruled out, but it does not seem to be a major contributor. Such a conclusion is consistent with the observation that Barrow-Prudhoe oils are generally found in older, mostly Paleozoic, reservoirs. Although it is geologically feasible for oil from Cretaceous source beds to enter the Paleozoic Prudhoe Bay reservoirs, it is less likely to occur in areas to the west where contact between Neocomain pebble shale and Paleozoic horizon is unusual.

The source (or sources) for Simpson-Umiat oils is even less clear. Similarities to both Neocomian pebble shale and Kingak were observed; possibly both served as the source. Again, however, the stratigraphic distribution of the oils must be considered. The Simpson-Umiat type oils are found in Cretaceous reservoirs, which suggests a Cretaceous source. From the data obtained, the Neocomian pebble shale is the most likely Cretaceous source. The possibility of a separate source for the Seabee subtype must also be considered. There are not enough data available to determine whether the Seabee oil is due to a local unsampled anomaly in one of the major source intervals or if it was derived from a separate stratigraphic interval not sampled for this study.

Fig. 13. Mass chromatogram of m/z 123 for methylene chloride extracts from 9 NPRA cores. Six core samples did not yield sufficient extract to get meaningful data. Numbered shaded peaks correspond to those on Figure 6. Peak A was not detected in oils.