The objects of the investigation are: highly viscous oil from Ashalchinskoye accumulation (Republic of Tatarstan, Russian Federation) and products of its refining: deasphaltizate, bitumen; asphalt. The deasphaltizate has been gained from the heavy oil residue relating to refining process of highly viscous oil with the use of thermal-steam method [23] and propane/butane deasphaltizing [24] .

X-ray fluorescence analysis has been applied to determine the elemental composition. The measurements have been carried out using energy-dispersive X-ray fluorescence spectrometer EDX-800HS2 produced by “Shimadzu” (Japan) with the implementation of semi-quantitative method. Measurement conditions: Rh-anode tube (50 watts); voltage: 50 kV, 15 kV; current: auto; medium-Helium; diameter measured: 10 mm; measuring time: 100 seconds. Measurements of the samples have been carried out in the range of Ti-U (0.00 - 40.00 keV), Na-Sc (0.00 - 4.40 keV).

The investigation on thermodynamic properties (density ρ, viscosity ν) of the heavy oil has been held in the temperature range of 298 - 363 K at pressure 0.098 MPa, whereas thermal expansion coefficient α_t at pressures up to 49 MPa. For the measurement of thermal expansion coefficient α_t the experimental apparatus has been used that implements the method of heat-conducting calorimeter [54] . Calculation formulas of the method and the measurement procedure are described in details in [55] -[57] . Reliability of the experimental data on α_p is confirmed by control (reference) measurements of chemically-pure substances such as hexane (= 1.3749; = 659.1 kg∙m⁻³) [54] , and butyl alcohol (= 1.3995; = 809.5 kg∙m⁻³) [57] [58] . For the measurement of isobaric heat capacity at pressures up to 50 MPa in the temperature range of 273 - 673 K the upgraded scanning calorimeter ITS-400 has been used.

Density ρ at atmospheric pressure has been measured by glass pycnometer with the volume of 10 ml in accordance with GOST 7465-67 [59] [60] , whereas at high-pressures the metal pycnometer has been used with the volume of 2.2 ml [57] . For the viscosity measurement there have been used capillary viscometers VPG-1 [61] . The refractive index has been determined by IRF-22 refractometer [62] [63] . Temperature control of pyknometers and viscometers has been carried out by ultra-thermostat U-10 and electronic thermostat by Huber with an accuracy of better than ±0.02 K. Weighting was performed by scales brand VLA-200 and VLTE-150 with an accuracy of ±0.0001 g and ±0.001 g, respectively.

On the basis of structural analysis (Table 1) and the measurements of density, refractive index and kinematic viscosity of the oil (Figure 1, Table 2), such molecular characteristics as molecular weight and similarity criterion α have been calculated (Table 3).

The calculation of the molecular weight has been carried out using approximation equations [64] :

where and―values of kinematic viscosity and density at temperature t = 20˚C. According to the

characteristics, the oil samples correspond to a fraction of bituminous oil with average molecular weight of 465 g∙mol⁻¹. Viscosity index VI = 150 [68] . During the calculation of the similarity criterion the following formula has been applied [65] -[67] :

where M_i and ω_i―molecular weight (g∙mol⁻¹) and weight percentage of a chemical element in the compound; n―number of elements in the compound.

The obtained data on molecular and structural characteristics have been implemented for the analysis of experimental data on heat capacity.

Control measurements of the apparatus have been carried out on two samples: stearic acid (chemically pure grade) at atmospheric pressure and butyl alcohol at pressures up to 30 MPa [69] [70] . Deviations from [58] [71] [72] for these substances do not exceed 2% - 3%. Confidence limits of the total measurement error of C_p at P = 0.95 are less than 2%. Heat capacity behavior of the oil-products are presented in Figures 1-6.

For the oil-product samples (Figure 1, Figure 4 and Figure 5) some common patterns can be noted, typical for phase transitions. These changes relate primarily to the composition of oil-products [43] -[47] . Thus, according to [45] -[48] data on hydrocarbon, structural-group and component analysis for a fraction of 350˚C - 450˚C, which corresponds to the characteristics of the oil investigated, paraffin-naphthene compounds are represented mainly by solid alkanes and naphthenic hydrocarbons with various numbers of rings, while the aromatic hydrocarbons consist primarily of polycyclic naphthene-aromatic structures with short side chains. This is also confirmed by chromatograms from the results of [45] . Thus 350˚C - 450˚C oil fraction contains some n-alkanes C26-C31, also a large number of high-molecular pentacyclic structures-hopanes C27-C35 is noted, where adianthan C29 and hopane C30 have the highest concentrations. Thus, the presence of solid phase in the composition of the oil and its refinery-products causes a particular behavior of the heat capacity (Figure 1, Figure 4

and Figure 5). The temperature dependences include phase transition regions. For example there are two such regions in Figure 1: the first (more evident) in the range of 350 - 420 K, the second (narrower) in the range of ~430 - 460 K. As for the asphalt (Figure 4) and bitumen (Figure 5) there are more than two such regions. Similar patterns have been observed previously for the samples of crude oil and bitumen [37] -[42] . Such behavior according to [39] is explained by the presence of two phases of asphaltenes, one of which melts or vitrifies (endothermic effect), and the other one dissolves in the first phase (exothermic effect). For example, in [39] the discovery of 4 phases is noted.

Different methods of oil-products’ heat capacity calculation have been analyzed: group component method, thermodynamic similarity method and empirical equations. All these methods have a common feature-they all require data on composition, structure and properties to calculate the heat capacity. In many cases, especially for multi-component compounds, such information is not available. According to the obtained data on elemental and structural composition of oil-products the computational methods have been analyzed, which require a minimum of information on the example of oil sample No.1. The results of calculations according to [52] [65] -[67] [73] and their comparison are presented in Figure 1. This figure shows that the smallest deviations of specific heat are calculated by methods of Shaw [65] -[67] and Filippov [73] . For the calculations by the method of Shaw [65] -[67] the elemental composition data (Table 1) have been used, while for the calculations by the method of Filippov [73] ―the data on the content of saturated and aromatic hydrocarbons [43] [44] and molecular weight (Table 2). The calculation method [52] gives much lower results (Figure 1), since it is not intended for heavy oils and bitumen. Thus, the method of calculation in [65] -[67] allows estimating the heat capacity values with the smallest error on the basis of minimal information. To analyze the phase behavior of oil there has been used an approach based on the calculation of excess heat capacity, which represents the difference between the measured heat capacity and heat capacity of the solid phase [38] [65] -[67] . Elemental composition data is used for the calculations. Based on the elemental analysis (Table 1), thermal capacities of solid phases and the excess heat capacities have been calculated for the oil-products (Figure 7). This figure shows both endothermic and exothermic effects. Similar dependences have been built for the oil (sample No.2), measured under pressure (Figure 2 and Figure 8). As can be seen from these figures the pressure has no significant effect on phase behavior.

As a result, it is found that the phase behavior is greatly influenced by temperature and composition of the oil. Pressure has no significant effect on behavior of the specific heat of the oil.

Experimental apparatus [54] has been applied for the measurement of α_р thermal expansion coefficient of the oil (sample No.2) at temperatures T = 298 - 363 K and pressures up to 50 MPa with an accuracy of ±2%. Results of the measurements are presented in Figure 9.

Experimental data on, obtained in a wide range of state parameters, allow estimation of the change patterns of α_p depending on temperature and pressure.

At temperatures 298 K and 303 K with the increase of pressure up to 25 MPa α_p first increases and then decreases. As for the other isotherms throughout the whole range of pressures a gradual decrease in thermal expansion coefficient can be observed. It should be noted that at a pressure of 25 MPa an intersection of α_p isotherms occurs. Such behavior has been already noted [54] for n-hexane, where the intersecting point is near 70 MPa.

The results of the thermal expansion coefficient measurements have been applied for the calculation of density by using the following equation:

where and―density of oil at pressure Р and temperatures T and Т_R, kg∙m⁻³; T_R―reference temperature for the experiments, equal to 323 K. Density values and are presented in Table 4.