The present study continues a series of publications on the analysis of the composition of crude oils of major oil-and-gas basins (OGBs) in Russia and neighboring countries based on 1H and 13C NMR examination [1–7]. The ultimate goal of this series is to correctly identify the generalized composition of crude oils in, at least, North Eurasia. Moreover, given the diversity of the OGBs to be analyzed, these data could further be extrapolated, with an acceptable approximation, to all crude oils on earth. It should be noted, in this context, that a misconception has prevailed in the modern petroleum geochemistry as to the appropriate methods for describing the composition of a set of crude oils. Researchers typically rely on the mean, highest, and lowest values of parameters, as well as (sometimes) on their root-mean-square deviations (e.g., [8]). However, the mathematical statistics principles with regard to data processing [9, 10] dictate that, for a fairly large set of oils, a composition parameter cannot be considered identified properly unless either a distribution function or distribution density function has been determined for this parameter. To date, such data have only been published for the sulfur content in crude oils [11]. Thus, within the scope of our works, the distribution density functions were first obtained for all the OGBs under study [1, 3, 6, 7].

This study focused on crude oils from the North Caucasus OGB. This basin differs substantially [12] from those previously explored [1–7]. Its commercial oil and gas occurrence has been proven in a section from Lower Triassic to Neogene [12]. The reservoir occurrence depth range is very wide: from several hundred to six thousand meters. The range of reservoir temperatures is also broad: 20–180°C. In particular, the highest oil reservoir temperatures in Russia have been recorded in this OGB [12]. The region’s source rocks have been studied, at least, since the early 1960s. A comprehensive study on this subject by Bazhenova et al. [13], which includes a brief review of prior relevant publications, showed that these rocks range from the Lower Triassic to Oligocene–Miocene age. It was demonstrated that most of these formations consist of mixed type (II + III) kerogene and some are type II and type III. The paper further notes some presence of terrigenous components (in various concentrations) in the precursor organic matter (OM) of almost all the source rocks [13].

We have found only one prior study dedicated to the analysis of the composition of Ciscaucasian crude oils by advanced methods [14]. The referenced paper reports data on the types of the source rocks and precursor OM that were derived from assessing the compositions of a number of cyclic biomarkers in 50 crude oil samples from 33 oil fields. The assessment covered all major stratigraphic plays of the region from Triassic to Neogene. These crude oils were shown to be of marine origin, with significant terrigenous admixtures in some deposits.

EXPERIMENTAL

We tested crude oil samples from the TIPS RAS collection gathered in the 1960s, 70s, and 80s. According to the adopted sample preparation procedure [1], total C8+ or C9+ fractions were selected as objects of structural-group analysis for this study. The reasons for this decision are detailed in reference [1].

1H NMR spectra of 250–350 mg/mL oil solutions in CDCl3 (400 MHz) were recorded at 313 K on a DRX-400 spectrometer (Bruker, Germany) equipped with a 5-mm probe without sample spinning. The spectrum record modes were as follows: data acquisition time 4 s, relaxation delay 3 s, pulse width 55°, 128–256 scans. The correction factors for all analytical groups of signals, taking into account their saturation, were measured with an error of ±1% by comparing the integral signal intensities in the spectra obtained both without saturation (relaxation delay 20 s, pulse width 70°) and with saturation (12 samples, three spectra in each mode for each sample). The correction factors were determined separately for gas condensates and light oils, for “conventional” oils, and for oils with significant broadening of 1H NMR signals. Chemical shifts against a tetramethyl silane standard are known to strongly depend on the concentration of the fraction in the solution, the composition of the fraction, and the recording temperature. Therefore, the most intense signal, which corresponded to the CH2 resonance in the middle of alkyl chains, was taken as a reference point for measuring the chemical shifts. For this signal, we assumed: δ = 1.280 ppm.

RESULTS AND DISCUSSION

We examined 86 samples of crude oils from 54 fields and exploration areas (Table 1) that cover, in total, the entire basin area. The samples represent all the productive plays, from Lower Triassic to Neogene, and major tectonic structures (Table 1) [1]. The occurrence depth and the reservoir temperatures (Tres) ranged from 300 to 5800 m and from 20 to 180°C, respectively. The distributions of depths and temperatures for the tested oil samples (Fig. 1) are reasonably consistent with their frequency of occurrence in the OGB [12].

Table 1. Oil fields represented in tested samples
Full size table
Fig. 1.
figure 1

Distribution density of occurrence depth and reservoir temperature for crude oil samples under study.

Full size image

The following structural-group parameters were measured for all the samples:

Hγ—hydrogen of CH3 groups separated by at least three C–C bonds from aromatic rings, carbonyl groups, or heteroatoms;

Hβ—hydrogen of –CH2– and –CH< groups, in β- and more distant positions to the same structural units;

Hα—hydrogen of CH3–, –CH2–, and –CH< groups in α-position to the same structural units;

Har—hydrogen in aromatic rings, further divided into two ranges:

Har,1—hydrogen predominantly in monoaromatic structures; and

Har,2+—hydrogen predominantly in bi- and polycyclic aromatic structures; and

Hdbl—hydrogen in isolated double bonds.

In addition, the values of Har,2+/Har,1 and Hα/Har [1] were calculated.

For all parameters, the average characteristics of distribution were calculated, namely: mean; median; root-mean-square deviation (RMSD); and median confidence interval (Table 2). We also evaluated the simplest criteria to distinguish a distribution from normal: the ratio of the average absolute deviation (AAD) around the mean to the RMSD; eccentricity; and excess kurtosis [10, 15].

Table 2. Distribution characteristics of composition parameters for North Caucasus crude oilsa
Full size table

In contrast to the previously investigated basins, the content of unsaturated compounds in the majority (46 out of 86) of the North Caucasus crude oil samples proved to be below the detection limit (0.002% Hdbl with respect to the total hydrogen in the sample). This finding corroborates our previous data obtained by high pressure liquid chromatography in combination with ozonolysis [16]. This explains the incompleteness of the Hdbl data in Table 2.

The data in Table 2 are suitable for estimating a deviation of the distribution of the composition parameters from normal. In this regard, the table shows that none of the measured parameters obeys the normal distribution law. The distribution density plots presented in Figs. 2 and 3 provide a detailed picture of the respective distributions. These plots, combined with Table 2, clearly indicate that many distributions are bimodal and almost all are asymmetric. Consequently, these oil composition data can only be processed by non-parametric statistical methods. In particular, medians and median confidence intervals determined by signed-rank test [9, 10], rather than means and RMSDs, are informative (see Table 2).

Fig. 2.
figure 2

Distribution density of parameters describing aromatic components in crude oils.

Full size image
Fig. 3.
figure 3

Distribution density of Hβ and Hγ.

Full size image

To compare the composition of the North Caucasus oil samples with the previously explored crude oils (from the Volga–Urals, Western Siberia, and Eastern Siberia OGBs) [1, 3, 6], a non-parametric Mann–Whitney test (MWT) was used. This test is designed to indicate the equality probability of two distributions (and to show which of them is shifted upwards) [9, 10]. The results are presented in Table 3. To visualize shifts between the distributions at different significance levels (SL) of the respective MWT values, Fig. 4 shows the distribution density plots for three pairs of parameters in which the differences (MWT values) had SL on the order of 10–6, 10–12, and 10–21.

Fig. 4.
figure 4

Comparison of distribution density for three pairs of parameters between which the significance level of MWT values was on the order of 10–6, 10–12, and 10–21 (see Table 3).

Full size image
Table 3. Comparison of composition parameters of crude oils from North Caucasus vs. previously explored OGBs [1, 3, 6] using MWT [10]a
Full size table

Table 3 indicates that the difference of the North Caucasus crude oils from the Volga-Urals and Western Siberia is primarily in the content of aromatics. The North Caucasus oils exhibit markedly lower concentrations both of monocyclic and bi-/polycyclic aromatics. Moreover, in terms of Har,2+/Har,1 ratio, the North Caucasus oils stand out from the other three OGBs with a reliably higher proportion of bi-/polycyclic aromatics in the total aromatics. Based on the Hα/Har shift––a value primarily indicative of the degree of substitution of aromatic cycles––the amount of the substituents of these cycles in the North Caucasus oils is lower. There is no substantial difference in the total aromatics between the Eastern Siberia and North Caucasus crude oils. On the other hand, the content of bi-/polycyclic aromatics is reliably lower in the Eastern Siberia oils. This is likely explained by the fairly high presence of gas condensates in the Eastern Siberia fluids tested, whereas the North Caucasus samples were essentially free of gas condensate.

When comparing the parameters that describe saturated moieties (Hβ and Hγ), the North Caucasus oils also exhibit a significant difference from the other OGBs. The only exception is the Hγ distribution similarity to the Western Siberia oils. A higher Hβ with a lower Hγ is typical of crude oils that have a large proportion of normal and monomethylated alkanes.

Three groups of crude oils were previously identified in the North Caucasus [14]. The first group was revealed in the Triassic, Jurassic, and Lower Cretaceous deposits. These oils are characterized by high catagenetic maturity; they were generated in clays from shallow marine OM with a potential presence (up to 20%) of terrigenous OM. The second group consists of Upper Cretaceous oils and crude oils from the Cenozoic sediments of the Tersk–Sunzhensk zone. They are catagenetically mature, and “typically marine, generated in clays” [14]. The third group is the Cenozoic oils from the West Kuban Trough. They are also marine oils generated in clays, although catagenically less mature. Thus, the principal difference between the oil groups identified in the referenced study concerned their catagenetic maturity degrees. On the other hand, this paper provides no solid evidence that the crude oils under consideration can be distinctly classified into these groups.

Our data assessment produced a different classification of the crude oils according to their content of aromatics and alkanes (both normal and monomethylated). We were able to expressly distinguish three groups: the oils from Triassic and Jurassic reservoirs; the oils from Cretaceous (both Upper and Lower) deposits; and the Cenozoic oils. This classification considered the medians of the measured composition parameters, with due regard to the median confidence intervals determined by signed-rank test (Table 4).

Table 4. Comparative composition of crude oils in reservoirs belonging to three different age rangesa
Full size table

Table 4 shows that the three groups are only similar in the Har,2+/Har,1 value; in addition, the Cretaceous and Cenozoic oils are similar in Hγ, and the Cretaceous and Triassic/Jurassic oils are similar in Hα/Har. The aromatic content grows monotonically as the reservoir age decreases, such that the difference between the medians of total aromatics (nearly proportional to the difference in Har) in the groups at the age extremes is about three-fold. There is an important similarity in the Pg–N to T–J ratios for the Har, Har,2+, and Har,1 medians (2.8, 2.7, and 3.0, respectively). The table clearly indicates a higher content of normal and monomethylated alkanes (Hβ and Hγ) in the Triassic/Jurassic oils. To the best of our knowledge, such a pattern of differences in the oil compositions has never been reported in prior studies.

The low content of normal and monomethylated alkanes in the Cenozoic oils is due to biodegradation, as noted in [14] and confirmed by our chromatography/mass spectrometry data. However, the high content of these alkanes in the Triassic/Jurassic oils remains unexplained (especially taking into account that, after discarding one crude oil sample with a unique composition, the highest Hβ in the Western Siberia and Volga–Urals oils is below the lower confidence interval of the median for the Triassic/Jurassic oils). We suggest that in actuality the presence of terrigenous precursor OM may be appreciably higher than “up to 20%” as specified in [14].

Regarding the potential causes for the difference in the aromatic content (both among the OGBs being compared and among the stratigraphic plays within the North Caucasus), we can only make generalized hypotheses. One possible explanation is the difference in the precursor OM, with the specific magnitude of this difference remaining unknown. Alternatively, we may hypothesize a difference in the oil generation conditions (for example, a difference in the hydrogen available in the rocks to ensure hydrogenation or preclude dehydrogenation). Presently, given the paucity of research on the potential pathways for aromatics formation in crude oils, a more definite rationale cannot be proposed. Particularly, there are serious doubts about the conventional concept of the dehydrogenation of saturated analogs as a major pathway for aromatics formation because no consistency between, for instance, naphthalenes and decalines in crude oils has ever been identified.

CONCLUSIONS

Using the 1H NMR examination of a representative set of crude oils (86 samples, 54 oil fields and exploration areas) from the North Caucasus oil-and-gas basin, we obtained, for the first time, the structural-group composition of crude oils over the entire basin. Distribution density plots were presented for all the composition parameters measured. It was demonstrated that, because none of the parameters obeys normal distribution, non-parametric statistical analysis is needed for the data processing. In particular, medians and median confidence intervals (the latter being determined by signed-rank test [9, 10]), rather than means or root-mean-square deviations, should be relied upon. The North Caucasus oils were found to have a markedly lower content of aromatics (both monocyclic and bi-/polycyclic) than the crude oils from Western Siberia and Volga–Urals. Compared with the Eastern Siberia basin, the North Caucasus oils are distinguished by a slightly higher presence of bicyclic and polycyclic aromatics. At the same time, the North Caucasus oils exhibit the highest content of normal and monomethylated alkanes among the four basins under consideration [1, 3, 7]. The North Caucasus oils were classified into three stratigraphic groups according to their oil composition. The first group consists of Triassic and Jurassic reservoirs. These crude oils have the lowest content of aromatics and the highest content of normal and monomethylated alkanes. The second group is the Cretaceous sediments, and the third group includes Paleogenic and Neogenic deposits. The content of aromatics increases consistently from the bottom to the top of the section.