The depletion of light oil reserves, as well as the growing consumption of petroleum distillates, have prompted increasing involvement of low-quality heavy petroleum feedstocks (HPFs) in oil refining processes [1]. HPFs, such as high-viscosity oils, natural bitumens, and oil residues, are distinguished by high contents of aromatic hydrocarbons, resins, and asphaltenes. Among the groups of compounds present in oil dispersed systems (ODSs), asphaltenes are the most complex as they consist of heteroatoms, coke residue, and metals, mostly V and Ni. This complexity accounts for the variability of their properties [14].

Currently, two general models have been suggested to describe the molecular structure of asphaltenes, specifically a continental model and an archipelago model [2, 58]. The most sophisticated molecular architecture model, referred to as the modified Yen model, has been suggested by Mullins [7]. This model is based on the continental-type molecular architecture as the most stable and common in the asphaltene composition. According to this model, an asphaltene molecule contains a central polycyclic aromatic fragment (PAF) surrounded by peripheral alkyl substituents. The continental-type molecule consists of a large PAF (7–8 rings) and short peripheral substituents (5– 6 carbon atoms long). In contrast, the archipelago model has a small PAF (5–7 rings) and long alkyl substituents (more than 6 carbon atoms). The composition of asphaltenes is such that the PAF serves as their primary intermolecular and intramolecular interaction site, responsible for the association ability of asphaltenes as their main property [6, 8, 9]. Due to this property, HPF asphaltenes form nano-aggregates and clusters that influence the ODS properties. Thus, depending on their size, asphaltene associations can stabilize oil-water emulsions [3, 4], destabilize ODSs [4, 5], etc.

The rheological and physicochemical properties of ODSs are known to be affected not only by the concentration, but also by the molecular/colloidal structure of asphaltenes [5, 1012]. The exploration of this structure has become increasingly important due to the need to improve existing process technologies for HPF production and refining, and the need to develop new ones. Of particular relevance is the identification of the structural features of HPF asphaltenes and of their structure–property correlation.

This study identified the composition and molecular structure of asphaltenes isolated from heavy oil and two vacuum residue samples. Using a combination of physicochemical analytical methods, we determined the average molecular structural-group and geometrical parameters of the resultant asphaltene stacks. The structural features of asphaltenes from the selected HPF samples were identified, and the effects of their composition and structure on the properties of the petroleum feedstocks were evaluated.

EXPERIMENTAL

Materials

The following petroleum feedstock samples were investigated: a sample of heavy oil from the Ashalchinskoye field (HO); and two samples of vacuum residues (VR) produced from the crude distillation units (CDUs) in the oil refineries of TANECO (VR1) and LUKOIL Ukhta Refinery (VR2). Despite its relatively low asphaltene concentration, the heavy oil from the Ashalchinskoye field has a high heavy metal and sulfur content. This was of research interest from the viewpoint of exploring the composition and structure of HO asphaltenes when compared to those in refinery VRs. The VR1 and VR2 samples were selected based on the criterion of significant difference in their trace-elemental and elemental compositions and physical properties, while having comparable contents of asphaltenes and group chemical compositions.

The recovered asphaltene samples were labeled HO-A, VR1-A, and VR2-A, respectively.

Analysis of Petroleum Samples

Analytical methods for HPF samples. Asphaltene content. The precipitation of asphaltenes and measurement of their weight content in the HPF samples were performed as per ASTM D 6560 (IP 143) [13].

Density. The HO density was measured according to ASTM D 5002 using a DA-500 automatic density meter. The density of VR1 and VR2 was determined by the GOST 32183-2013 pycnometric method.

Softening point. The softening points of VR1 and VR2 were measured by the ring and ball method (GOST 33142-2014). The standard method is inapplicable to crude oils.

Dynamic viscosity. The dynamic viscosity of the HPF samples was measured on a Brookfield DV2TRV rotary viscometer in accordance with its operating instructions. The viscosity of the crude oil and vacuum residues was measured at 20°C and 100°C, respectively. 100°C was selected as the optimum temperature for the vacuum residues because this point allowed both samples to reach a fluid state. Under these conditions, the spindle immersion in the container with a VR sample enabled its free stirring, and no volatile compounds were emitted during heating.

Content of trace elements (V, Ni). The contents of the most common HPF metals, specifically V and Ni, were measured by inductively coupled plasma mass spectrometry (ICP-MS) (ASTM D 5708) on an Agilent 7900 mass spectrometer. Sample preparation was carried out according to ASTM D 7876.

Sulfur content. The sulfur concentration in the samples was detected on a Spectroscan SL energy dispersive X-ray fluorescence (ED-XRF) analyzer (GOST 32139-2013).

Coking capacity (expressed as Conradson carbon residue, or concarbon). The HPF concarbon weight content was measured by the ASTM D 189 Conradson method on an ACR-6 carbon residue automatic analyzer [14].

Group composition. The group chemical composition was examined by thin-layer chromatography with flame-ionization detection (IP 469 TLC-FID) on an Iatroscan MK-6S analyzer [15]. The standard method makes it possible to identify four hydrocarbon groups: saturates, aromatics, polars I, and polars II. According to IP 469, polars I are lower molecular weight polar compounds containing nitrogen, sulfur, and oxygen (e.g., benzoquinolines, carboxylic acids, phenols, and metalloporphyrins); and polars II are higher molecular weight polar compounds (similar, though not identical, to n-heptane-insoluble asphaltenes determined by IP 143/ASTM D 6560) [13].

Analytical Methods for Asphaltene Samples

Elemental analysis. The concentrations of individual chemical elements in the recovered asphaltenes were measured by the Pregla–Dumas method, where the samples were dynamically combusted in a pure oxygen environment and the resulting gases (CO2, N2, H2O, and SO2) were separated using a chromatographic column [16]. A Thermo Scientific Flash EA-2000 elemental analyzer was used for this analysis.

Content of trace elements (V, Ni). The contents of V and Ni in the asphaltenes were measured similarly to the analysis of the same metals in the initial HPF samples.

Average molecular weight. The average molecular weights (MW) of the asphaltenes were measured by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry using a Bruker Autoflex Speed mass spectrometer equipped with a solid-state UV laser (λ = 355 nm) in a linear mode and with detection of single-charged positive ions (detectable MW range 500–5000 Da). A 2,5-dihydroxybenzoic acid solution in dichloromethane (30 mg/mL) was used as a matrix. The matrix solution and analyte solutions (2 mg/mL in toluene) were dried, then deposited on a support according to the matrix-analyte-matrix sequence. The average MWs were derived from the extremum location in the mass spectra.

NMR spectroscopy. 1H and 13C NMR spectra of the asphaltenes were recorded using a Bruker Avance III HD 500 spectrometer equipped with a Bruker Prodigy CPP BBO H&F high-sensitivity cryo-probe. The NMR spectrometer operating frequencies were 500.13 MHz for 1H and 125.76 MHz for 13C. Asphaltene sample solutions (3.5 wt %) were prepared in deuterobenzene (99.9%), with tetramethylsilane (99.6%) as an internal standard. The solutions were held for 12 h. The recording of spectra was followed by integration within the chemical shifts inherent to asphaltenes. Then the structural-group parameters of the asphaltene molecules were calculated using a combination of NMR spectroscopy, elemental analysis, and mass spectrometry methods [5, 1719].

X-ray diffraction (XRD). XRD patterns of the asphaltenes were obtained on a Rigaku Rotaflex RU-200 X-ray diffractometer equipped with a rotating copper anode (wavelength λ = 1.542 Å), a Rigaku D/MAX-b horizontal goniometer, a focusing graphite monochromator crystal, and a scintillation detector. The XRD patterns were recorded in Bragg-Brentano geometry in the continuous θ–2θ scanning mode in the 2θ range of 5°–90°, with a step of 0.04° and a scanning rate of 2°/min. The XRD patterns were smoothed and deconvoluted to identify the diffraction maxima typical of asphaltene structures, and the geometry of asphaltene nano-aggregates was calculated [1722].

RESULTS AND DISCUSSION

Characterization of HPF

The composition and properties of the HPF samples are presented in Table 1.

Table 1. Physicochemical properties of HO and VR samples

As was expected, the oil residue samples exhibit higher contents of asphaltenes and polar resinous-asphaltenic compounds (polars I/II) than the crude sample. Asphaltenes are the heaviest and highest-molecular-weight components of ODSs. They also have the highest aromaticity, which causes significant differences in the density and concarbon between crude oil and vacuum residues (Table 1). An even more drastic change is observed in viscosity, which is the most sensitive to the presence of asphaltenes because they form dispersed phase particles due to their aggregation and self-association abilities. Tukhvatullina et al. [2] have demonstrated a viscosity growth in model solutions caused by an increase in the concentration of A1 asphaltenes (continental-type), which are propagators of aggregation [6]. Furthermore, the enlargement and the increased association ability of asphaltene particles resulting from a combination of intramolecular and intermolecular interactions are among the central factors affecting the rheological and structural-mechanical properties of the ODSs [23]. On the whole, an increase in the asphaltene content in the petroleum samples, regardless of their type and origin, enhances the density, viscosity, and coking capacity as a general trend. Attention should be paid to significant variations in the softening point and viscosity when passing from VR1 to VR2, despite only a minor difference in the asphaltene content between these samples. Specifically, an asphaltene concentration rise of 1.7 wt % boosted the softening point by 16°C and viscosity by 1100 mPa s.

The highest sulfur content among the samples subjected to the test (3.9 wt %) was observed in the heavy oil, despite a lower total polar concentration and a lower asphaltene content in this sample than in the vacuum residues. It is also worth noting that the V content in HO is comparable to VR1 and higher than in VR2. We observe a known correlation between the concentrations of vanadium and sulfur in the residue samples [24]. However, the vanadium variation from VR1 to VR2 as a function of asphaltene content follows the opposite trend: the vanadium content declines despite an increase in asphaltenes (Table 1).

Characterization of HO/VR1/VR2 Asphaltenes

Elemental composition. The elemental analysis data for the asphaltene samples are summarized in Table 2.

Table 2. Elemental composition of asphaltenes from heavy oil (HO-A) and vacuum residues (VR1-A and VR2-A)

The simultaneous presence of heteroatomic functional groups and metals (V, Ni) that form metal complexes when interacting with these groups, governs the polarity of asphaltenes [9]. In the first approximation, the polarity of samples can be estimated by the total amount of metals and heteroatoms without identification of the types of functional groups and metal compounds [25].

The highest content of heteroatoms and metals in HO-A explains the highest sulfur content and the high concentration of vanadium in the HO sample, despite the markedly lower asphaltene content compared to the VR samples (Table 1). At the same time, the contents of metals, sulfur, and nitrogen in the asphaltene samples vary symbatically, which confirms the presence of vanadium and nickel in petroleum in the form of various complex compounds with these heteroatoms, including metalloporphyrins. VR2-A has the lowest content of heteroatoms and vanadium, which most likely accounts for the lower vanadium and sulfur contents in VR-2, while the latter has the highest asphaltene content (Table 1). Despite its presumably high polarity, HO-A exhibits a higher H/C atomic ratio than the VR asphaltenes, which indicates its lower aromaticity (Table 2).

Thus, the polarity of asphaltenes, along with their content, belongs to the key parameters responsible for the contents of heteroatoms and heavy metals in HPF. As a consequence, despite the lower asphaltene concentration, heavy oil may have comparable or higher metal and sulfur content than vacuum residues.

Structural-group analysis of asphaltenes. Based on the elemental analysis data and average MW values, we calculated a number of generalizing parameters to evaluate the PAF and peripheral structures of asphaltene molecules. Table 3 presents the absolute numbers of atoms [17, 26] and the average molecular structural-group parameters of the asphaltene samples. Also, the average parameters of the supramolecular structures of asphaltenes were calculated based on the XRD data on the samples (Table 4).

Table 3. Average molecular structural-group parameters of asphaltenes (A)
Table 4. Average geometry of asphaltenes (A)

HO-A displays lower values of molecular weight, of total amount of aromatic hydrogen and carbon, and of aromaticity factor than the VR asphaltene samples (Table 3), which is consistent with the higher H/C ratio. When comparing VR1-A to VR2-A, the aromaticity and the aromatic carbon amount in molecules are slightly higher in VR2-A. The molecular structure of asphaltenes is known to directly affect the aggregation of molecules. In particular, a higher aromaticity and a larger number of aromatic rings increase the molecular association energy and the resultant aggregate size [9]. In this regard, the NMR data obtained are in good agreement with the XRD data and with the average structural parameters of the supramolecular stacks. Table 4 clearly shows that HO-A, of all the samples, has the lowest values of aromatic sheet diameter (La), number of aromatic sheets per stack (M), and stack height (Lc). The term “aromatic sheets” here refers to the aromatic planes of asphaltene molecules piled into stacks, and “stack height” is the average size (average height) of aromatic clusters oriented perpendicular to the sheet plane [1722, 26]. It is worth noting that the structural variations in asphaltenes observed in our study, in particular an increase in the number of molecules and in the height of respective stacks when passing from the HO sample to the VRs, are consistent with the findings of reference [27]. That study showed, based on HR–TEM data, that the asphaltenes of two VR samples from refineries in China and Venezuela had a greater average number of layers in the stacks when compared to the asphaltenes of a natural bitumen sample from oil sands in Pulau Buton, Indonesia. In addition, the bitumen exhibited a higher average interlayer distance in asphaltene stacks than the VR [27]. In our case, the distance between aromatic sheets (dm) was almost the same in the HO-A and VR-A samples. This may be associated with the high polarity of the crude asphaltenes, which intensifies the interaction and attraction on the PAF due to the large number of heteroatoms and metals, and compensates for their smaller size [9]. The highest dγ value in HO-A (5.90 Å) is also of particular importance. We can assume that the change in the distance between aliphatic layers is partly or wholly responsible for the similarity of distances between aromatic sheets in the asphaltene stacks of the heavy oil and vacuum residue samples. Alkyl substituents in HO-A probably exercise the strongest repulsion to maintain a balance with the highly polar PAF.

For the VR asphaltenes, the geometry of newly-formed stacks is also affected by the molecular structure. For instance, the higher aromaticity of VR2-A leads to an increase in the number of aromatic rings per sheet (NOAr) and in the La value, and decreases dm to some extent. This may serve as evidence of the formation of a more ordered structure in this case. Reference [11] noted that, along with the asphaltene concentration, a more regular and ordered layered structure of asphaltenes (higher ordering of aromatic sheets) increases the asphalt softening point. Viscosity is known to be a function of temperature and the property most sensitive to asphaltene content. As a result, larger concentrations of highly ordered asphaltenes will increase both the softening point and the viscosity of the feedstock. The main precursor of coke is asphaltene’s PAF, which forms a layered supramolecular structure of associates. At the same time, the alkyl substituents in asphaltene molecules cause steric hindrances for association and structural ordering [9]. Enlargement of the PAF is supposed to increase the amount of coke residue resulting from thermal decomposition. Aromaticity, along with aromatic sheet ordering, also affects viscosity [11, 12]. Thus, the VR2-A asphaltene structure is assumed to contribute to the change in the VR sample properties by elevating the softening point and viscosity and increasing the concarbon content in VR2 compared to VR1 at similar asphaltene concentrations (see Table 1).

In addition, the high CAr and fa values, the small HAr amount, and the comparable NOAr numbers suggest the presence of large PAFs in the asphaltene molecules. Assessment of HAli distribution revealed that the length of lateral alkyl chains (n) does not exceed five (5) carbon atoms (short chains) [58, 1719] in any of the asphaltene samples. The HAli distribution was identified as a function of hydrogen position to aromatic ring. All the asphaltene samples exhibit the predominant presence of β-hydrogen and comparable amounts of α-hydrogen and γ-hydrogen (Table 3). Such a distribution is associated with the number and orientation of CH2 and CH3 groups. Based on the amount of β-hydrogen and n ≈ 5, CH2 groups were demonstrated to be predominant in alkyl chains and naphthenic rings. Taking into account the presence of short lateral substituents and the size of the polyaromatic core consisting of an average of seven (7) aromatic rings (NOAr), it is reasonable to assume a significant presence of island-type molecules in the samples under study [6].

CONCLUSIONS

A combination of physicochemical analytical methods was used to determine the properties, elemental composition, and group chemical composition of the selected HPF samples, as well as the elemental compositions, molecular weights, and structural properties of asphaltenes isolated from these samples.

It was demonstrated that heavy oil (HO) asphaltenes from the Ashalchinskoye field, when compared to asphaltenes of the vacuum residue (VR) samples examined in the study, exhibit a lower average molecular weight (1500 vs. ~1700), a smaller number of aromatic carbon atoms (71 vs. ~90), a lower aromaticity factor (0.70 vs. 0.74–0.75), and smaller molecular stacks in the supramolecular structure. The HO asphaltene aggregates consisted, on average, of fewer molecules (6 vs. 7) and had a smaller height of aromatic sheet stacks (~18 vs. ~21–22 Å) and a smaller aromatic sheet diameter (19.0 Å vs. 19.3–19.4 Å).

The metal content in asphaltenes is largely determined by molecule polarity and, in particular, by the content of sulfur and nitrogen atoms capable of forming coordination bonds with vanadium and nickel. Thus, the polarity of asphaltenes, along with their concentration, is one of the main factors responsible for the content of heteroatoms and heavy metals in HPF.

An increase in the content and aromaticity of asphaltenes, size and ordering of their aggregates leads to an increase in the density, viscosity, softening point (for vacuum residues), and coking capacity (concarbon) of the HPF. For example, an increase in the asphaltene concentration in the VR samples of 1.7 wt % and an increase in their aromaticity—from 0.74 to 0.75—with a simultaneous increase in the average aromatic sheet size by 0.1 Å and a decrease in the aggregate stack’s sheet-to-sheet distance by 0.01 Å, was accompanied by an increase in the VR density and viscosity from 1007 kg/m3 and 1400 mPa s to 1014 kg/m3 and 2500 mPa s, respectively. This was also accompanied by an increase in softening point and concarbon by 16°C and 1.1 wt %, respectively. Thus, it can be concluded that the properties of ODSs are affected not only by the asphaltene content, but also by the structural properties of asphaltene molecules and their aggregates.

Further research is needed with large-sampling of heavy oils and oil residues to confirm and/or clarify the patterns identified in this study. The authors hope that such research will extend the knowledge on the structural organization of asphaltenes in various HPF types and on their effects on the ODS properties. Furthermore, additional research in this direction should be aimed at a better understanding and optimization of HPF upgrading and refining processes. In particular, we intend to use the analytical data obtained in this study for further investigation of the effects of the asphaltene composition and structural properties on the efficiency of solvent deasphalting of the HPF samples examined in this study with the addition of nanosized adsorbents [28].