Influence of Conversion Conditions on Heavy-Oil Coking During in

Mar 27, 2018 - Coking of heavy oil plays a key role in in situ combustion, which is affected by conversion conditions. In this study, a fixed-bed reac...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/EF

Cite This: Energy Fuels XXXX, XXX, XXX−XXX

Influence of Conversion Conditions on Heavy-Oil Coking During in Situ Combustion Process Dong Liu,† Lijuan Chen,‡ Long Chen,‡ Ruonan Zheng,† Qiang Song,*,† and Gang Cai‡ †

Key Laboratory of Thermal Science and Power Engineering of Ministry of Education, Department of Energy and Power Engineering, Tsinghua University, 100084 Beijing, China ‡ Engineering and Technology Research Institute of Xinjiang Oilfield Company, 834000 Karamay, China ABSTRACT: Coking of heavy oil plays a key role in in situ combustion, which is affected by conversion conditions. In this study, a fixed-bed reactor and a pressurized reactor were used to thermally transform a heavy oil sample from China into coke under different atmospheric conditions, heating rates, and pressures. Elemental composition and surface functional groups were studied by an elemental analyzer and diffuse reflectance Fourier transform infrared spectrometer, whereas oxidative activity and coke yield were characterized by a thermal gravimetric analyzer. Results showed that yields, characteristics, and oxidative activities differed between the cokes produced in inert and oxidizing atmospheres. Compared with an inert atmosphere, an oxidizing atmosphere presented lower coke formation temperature and higher yield. In comparison with inert atmosphere, the coke produced in oxidizing atmosphere contained more oxygen, thus increasing the amount of surface functional group, but its oxidative activity was poorer. In oxidizing atmosphere, conversion conditions influenced yield and characteristics of coke but exerted minimal influence on oxidative activity. Oxygen content increased, whereas coke yield decreased with increasing O2 concentration. Oxygen content decreased, and yield increased initially and then decreased with increasing heating rate. Temperature of coke formation and oxidation decreased, while oxygen content and yield increased with increasing pressure. Through analysis of elements and functional groups of residues produced at different holding temperatures, our study confirmed that the following processes occurred during low-temperature oxidation: evaporation, oxygen-adding reaction, dehydrogenation and dealkylation, polymerization, decarbonylation, and mild oxidation of coke. Conversion conditions, including heating rate, oxygen concentration, and pressure, affected conversion rate of these processes, thus influencing coke characteristics.

1. INTRODUCTION Petroleum is one of the most important fossil fuels. With the recovery of light oil, heavy oil accounts for a high proportion of oil reserves. Currently, thermal recovery is widely used for highviscosity heavy-oil exploitation, which mainly includes cyclic steam stimulation and steam flooding.1 Given that oil production by these two methods decreases with increasing stimulation cycles and flooding time, a large proportion of heavy oil reserves remains underground.2,3 In situ combustion (ISC) shows potential as a thermal recovery method that enhances oil recovery. ISC reduces heavy-oil viscosity by burning petroleum coke to generate heat with combustion front propagation.4,5 Therefore, stability of combustion front propagation is the key to ISC.6 The coke from heavy oil acts as fuel, and its formation and characteristics serve as an important research subject. Given that reservoir conditions and operation methods vary, conversion conditions of heavy-oil coking differ in different ISC projects. Such differences may influence yield, characteristics, and oxidative activity of coke. Reaction atmosphere is the most important conversion condition. In an ISC project, coking atmosphere may be either pyrolytic without oxygen or oxidizing with low-concentration O2. Reaction atmosphere has been proven to significantly affect heavy-oil coking.7,8 Studies have shown that coke formation temperature and yield differ between the coke produced in oxidizing and inert atmospheres: Xu et al. observed that compared with that under inert atmosphere, coke formation temperature reduced by 200 °C © XXXX American Chemical Society

and yield increased by four times in oxidizing atmosphere under 5 MPa.9 On the other hand, several oxygen-containing functional groups exist in the coke formed from oxidizing atmosphere, indicating differences in coke characteristics.10,11 These differences are related to coke formation mechanism during pyrolysis and oxidation. In inert atmosphere, with increasing temperature, heavy-oil coking can be divided into low-temperature evaporation and high-temperature pyrolysis.12 During high-temperature pyrolysis, hydrocarbon molecules crack and polymerize, forming molecules with high molecular weights and numerous aromatic rings.13 These large molecules precipitate from the oil phase and form the mesophase. The coke produced by pyrolysis is obtained with increasing mesophase.14−19 In oxidizing atmosphere, low-temperature oxidation (LTO) occurs and changes the conversion process of heavy oil, thus influencing the coking process.20 Studies have shown that O2 can be consumed by heavy oil even at reservoir temperature.21,22 During LTO, oxygen-containing functional groups appear in the gas, liquid, and solid products.21−26 Appearance of these functional groups influences formation temperature, molecular structure, and yield of coke precursors. Molecules with these functional groups are easily polymerized and generate coke precursors which transform into coke by further cracking.27 Increasing the amount of precursors also Received: January 9, 2018 Revised: March 27, 2018 Published: March 27, 2018 A

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

saturates, aromatics, resins, and asphaltenes (SARA) analysis results of the sample oil.8 Oil was mixed with quartz sands according to the SY/ T 6953-2013 standard. Chemically pure quartz sand was produced by Sinopharm Chemical Reagent Co., Ltd. in China. Sand particle diameter ranged from 10 to 20 μm. Oil-sand mass ratio approximated 1:9, which was measured by a TGA oxidation experiment for precise oil mass ratio for each patch of oil-sand sample. 2.2. Experimental System. In this study, a fixed-bed reaction system and a pressurized reactor were used. Figure 1a shows the fixedbed reaction system. The quartz tube reactor consisted of an outer tube and an inner tube. The outer and inner tubes measured 30 and 15 mm in diameter, respectively. A quartz crucible measuring 20 mm in diameter was used as sample container. A hole plate was welded in the middle of the crucible. Quartz cotton was spread on the hole plate, and the sample was laid on the quartz cotton. The crucible was inserted into the constant-temperature region of the electric furnace. Relative deviation of the fixed-bed reaction system was less than 3% (measured by coke yield). Figure 1b illustrates the pressurized reactor. Maximum temperature and pressure of the reactor measured 500 °C and 20 MPa, respectively. Volume measured 500 mL. The oil-sand sample was placed in a crucible with a quartz glass holder to isolate the samples from the heating wall. The diameter of quartz tube reached 15 mm, and sample layer thickness approximated 10 mm. The thermocouple, which controls the heating power, was buried in the reactor wall, and another thermocouple was inserted into the reactor beyond the sample through the reactor top cover to directly measure sample temperature. 2.3. Methods. Coke was prepared under 0.1 MPa in the fixed-bed reaction system. The basic experimental conditions were summarized in Table 2. For the oxidizing atmosphere, 5% O2 in Ar was used as gas reactant, and Ar was used for the pyrolytic atmosphere. Gas flow rate totaled 500 mL/min, and sample weight approximated 1.3 g. Experiments were performed using a linear temperature program with a constant heating rate of 5 K/min. Terminal temperatures reached 400 and 600 °C during oxidation and pyrolysis, respectively. In oxidizing atmosphere, coke is well formed at 400 °C, and with further increase in temperature, the oxidation of coke begins, so the terminal temperature was set as 400 °C. In an inert atmosphere, coke formation starts at around 450 °C and the chemical characteristics of coke produced from 500 to 600 °C appears to be similar.11 As the fire front could possibly reach 600 °C, this temperature was set as the terminal temperature of pyrolysis. To study the influences of conversion conditions on heavy-oil coking, several heating rates (3, 5, 10, and 15 K/min) and oxygen concentrations (2%, 5%, 10%, and 20%) were selected under atmospheric pressure. Coke was prepared under high pressure in the pressurize reactor with 5% O2−Ar mixed gas as a gas reactant. The experimental method was as follows. Sample was heated to holding temperature with high heating rate and then held for 1 h. Holding temperatures totaled 350 and 400 °C, and pressure values measured 0.1 and 3 MPa. Gas flow rate was 500 mL/min. All solid samples from preparation experiments were soaked in 50 °C toluene and then filtered. These steps were continued until toluene became clear. The undissolved substance was regarded as coke.17 The solid samples obtained under the above coke preparation conditions were completely coked with minimal oil absorbed. HTO performance of coke is vital for the ISC technology, which is determined by the characteristics of coke. Therefore, the chemical characteristics of coke are analyzed. An elemental analyzer (Eurovector EA3000) was used to study the carbon−oxygen−hydrogen composition of coke. A diffuse reflectance Fourier transform infrared spectrometer (DRIFTS, Thermo Scientific Nicolet 6700) was used to analyze the functional groups on coke surface. A thermogravimetric analyzer (TGA, Mettler-Toledo TGA/DSC 3+) was used to investigate oxidative activity and to analyze coke yield. Air was used as reactant gas in all TGA tests in this study. Coke yield of heavy oil is defined as the mass of coke produced divided by the mass of original oil. As sample recovery of fixed-bed reactor experiments may be incomplete, using mass changes in the sample before and after experiments may introduce additional errors. Therefore, coke samples from the fixed-bed reactor experiments were

influences the stability of oil colloidal dispersion, thus reducing coking temperature.22 As differences exist between the coking mechanisms in inert and oxidizing atmospheres, coke yield and formation temperature also differ. Therefore, differences in characteristics and oxidative activities require further studies. In ISC, air injection is often applied excessively to fully use the coke produced. Therefore, heavy-oil coking is usually performed in oxidizing atmosphere. As discussed above, LTO is the key to coking in an oxidizing atmosphere. Oxygen inserts into the carbon chain and then polymerizes to form a coke precursor, which cracks into coke. However, LTO is a complicated process. Coke formation could not be theoretically predicted accurately until now. Conversion conditions affect LTO, as reflected from coke characteristics. Several experimental studies have shown that conversion conditions influence coking. According to studies of heavy-oil oxidation in differential scanning calorimeter, activation energy of hightemperature oxidation (HTO, mostly coke oxidation reaction) increased with heating rate, whereas total heat release decreased.28,29 For the influence of pressure, by using a pressurized thermal gravimetric analyzer, Kok et al. observed that weight loss in LTO increased with pressure (0.69−2.07 MPa). Activation energy of HTO slightly changed when pressure varied.30 However, according to the results of ramped temperature oxidation (RTO) experiments, Anto-Darkwoh and Cinar discovered that activation energy of HTO increased from 76 to 120 kJ/mol with pressures from 4.76 to 11.89 MPa.31 The conflict between these results may be related to differences in reaction systems and kinetic calculation methods. Limited research can be found for the direct influence of oxygen concentration on coke characteristics. Some studies showed that total O2 consumption of bitumen, as well as the proportion of asphaltenes and coke, increased with O2 partial pressure when bitumen was treated at a specific LTO temperature range and high pressure.32 The above studies indicate the importance of conversion conditions on heavy-oil coking. Petroleum coke is the main fuel for combustion in ISC. Thus, yield of petroleum coke and its characteristics play vital roles in the stability of combustion front propagation.33 With regard to heavy-oil coking, conversion conditions, including heating rate, oxygen concentration, and pressure, vary in ISC projects. Therefore, this study investigated the influence of conversion conditions on coking. A fixed-bed reaction system was used to prepare coke with different atmospheres and heating rates, whereas a pressurized reactor was used to prepare coke under varying pressures. After coke preparation, the yield, elemental composition, functional group, and oxidative activity of coke were analyzed to evaluate the influence of conversion conditions on coking of heavy oil.

2. EXPERIMENTAL SECTION 2.1. Samples. The heavy oil from an oilfield in China was used in this study. Viscosity measured 1878 mPa·s (50 °C). Water and impurity were removed from the oil sample according to Chinese national standards SY/T 6316 and SY/T 6520. Water content measured less than 0.5% after processing. Table 1 summarizes the

Table 1. SARA Composition of Heavy Oil Sample (wt %) saturates aromatics resins asphaltenes

41.6 19.1 37.2 2.1 B

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 1. Diagrams of reaction system. (a) Fixed-bed reaction system. (b) Pressurized reactor.

Table 2. Experimental Conditions of Coke Formation and Oxidation section

reaction system

V (mL/min)

O2 (%)

terminal temperature (°C)

heating rate (K/min)

pressure (MPa)

3.2.1 3.2.2 3.2.3 coke oxidation

fixed-bed reaction system fixed-bed reaction system fixed-bed reaction system fixed-bed reaction system pressurized reactor TGA

500 500 500 500 500 70

0 5 5 2, 5, 10, 20 5 21

600 400 400 400 350, 400 holding for 1 h 700

5 5 3, 5, 10, 15 5 ∼25 30

0.1 0.1 0.1 0.1 0.1, 3 0.1

3.1

ground uniformly, and a small portion of each coke sample was burned in TGA. Mass loss is regarded as the coke content in the coke-sand mixture. Together with the mass ratio of oil in the oil-sand mixture evaluated by TGA, coke yield can be calculated using the following formula:

γ=

⎛ m − mr ⎞ ⎛ m0′ − mr′ ⎞ mcoke moil =⎜ 0 / ⎟/⎜ ⎟ ′ msand msand ⎝ mr ⎠ ⎝ mr′ ⎠

yield and elemental composition presented large differences. Coke yield reached 5% in pyrolysis and 29% in oxidation. Coke from pyrolysis mainly consisted of C and H, with an oxygen concentration lower than 2%. By contrast, the coke produced in oxidizing atmosphere contained higher amounts of oxygen, the proportion of which approximated 21%, whereas that of carbon was 74%. As shown in Figure 2b, two weak typical hydrocarbon absorption peaks existed in coke from pyrolysis; the wavenumber ranges measured 3000−3100 (aromatic C−C bond) and 1450−1650 cm−1 (aromatic C−H bond).34 Comparing with pyrolysis atmosphere, surface functional groups were more abundant due to the presence of high amounts of oxygen in the coke produced in oxidizing atmosphere. Infrared absorption peaks were detected within the ranges of 3350−3420, 3000− 3030, 1650−1750, and 1450−1600 cm−1. Among these peaks, the absorption at 3350−3420 cm−1 refers to the hydroxyl group; absorptions at 3000−3030 and 1450−1600 cm−1 refer to the aromatic ring; and that at 1650−1750 cm−1 refers to the carbonyl group.34 On the other hand, no infrared absorption of alkyl groups (2920−2970 cm−1) was observed in the cokes produced in both atmospheres, indicating that the alkyl groups either cracked into small hydrocarbons which were released into the gas phase or polymerized during heavy-oil thermal conversion. Given that coke elemental composition and surface functional groups differ between the cokes produced in different atmospheres, their oxidative activities may also differ. Figure 2c presents the temperature-programmed oxidation DTG curve of coke produced by pyrolysis and oxidation. Light-off temperature of coke produced by oxidation (393 °C) was 25 °C lower, whereas burnout temperature (560 °C) was 30 °C higher than those obtained in inert atmosphere. For a temperature-

(1)

where m0 and mr refer to the initial mass and residual mass after burnout from the coke-sand mixture oxidation TGA result, respectively. m′0 and m′r represent the initial mass and residual mass from the oil-sand mixture oxidation TGA result, respectively. The units of the variables are mg. The experimental errors are analyzed in Table 3.

3. RESULTS AND DISCUSSION 3.1. Characteristics of Coke Produced from Heavy Oil in Inert and Oxidizing Atmospheres. Figure 2 presents the yield, elemental composition, surface functional group, and oxidative activity of coke that was converted from heavy oil in inert and oxidizing atmospheres. As shown in Figure 2a, coke Table 3. Analysis of Experimental Errors experimental error (standard deviation) coke yield elemental composition

DTG peak temperature

pyrolysis: ±0.5% oxidation: ±0.3% C: ± 1.5% (oxidation), ±1.3% (pyrolysis) H: ± 0.1% (oxidation), ±0.2% (pyrolysis) O: ± 0.3% (oxidation), ±0.1% (pyrolysis) N: ± 0.07% (oxidation), ±0.08% (pyrolysis) ±2.5 °C C

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 2. Yield, element, surface functional group, and oxidative activity analysis of coke formed in oxidizing or pyrolysis atmosphere. (a) Coke yield and elemental analysis, (b) DRIFTS, and (c) derivative thermogravimetry (DTG) curve of coke oxidation.

programmed oxidation test, light-off temperature here is defined as the temperature at which the weight loss is 5% of the total weight loss, while the burnout temperature is the temperature at which the weight loss is 95%. Peak temperature of the DTG curve of coke produced by oxidation approximated 545 °C, which was 20 °C higher than that produced in inert atmosphere. This temperature was directly related to oxidative activity.35 Maximum relative oxidation rate of coke produced by oxidation was 50% lower than that of coke by pyrolysis. Therefore, the coke produced in inert atmosphere featured better oxidative activity than that in the oxidizing atmosphere. 3.2. Influence of Conversion Conditions on Heavy-Oil Coking in Oxidizing Atmosphere. As air injection is often excessively applied in ISC practice, heavy-oil conversion mainly occurs in oxidizing atmosphere. This section investigates the influence of conversion conditions on coke characteristics in oxidizing atmosphere. Conversion conditions, including heating rate, oxygen concentration, and pressure, were investigated. 3.2.1. Influence of Heating Rate on Coke Characteristics. Figure 3 shows the yield, chemical characteristics, and oxidative activities of cokes produced in 5% O2 atmosphere with varied heating rates. With increasing heating rate, coke yield increased from 26% (3 K/min) to 29% (5 K/min) then decreased to 25% (15 K/min). Oxygen proportion decreased with increasing heating rate, that is, from 23% at 3 K/min to 18% at 15 K/min. Surface functional groups were extremely similar among the cokes produced with various heating rates. Oxygen-containing

functional groups, such as hydroxyl and carbonyl groups, and aromatic groups can be found in these cokes. Absorption of carbonyl groups was strong at low heating rate owing to the large extent of oxygen-adding reactions and oxidation of hydroxyl. From the coke temperature-programmed oxidation DTG curve in Figure 3c, light-off temperature remained nearly constant (around 394 °C). Except for the sample produced with a heating rate of 3 K/min, peak temperature of the DTG curve of other samples was approximately 545 °C. Maximum reaction rate was also constant among samples. As for the coke produced with a heating rate of 3 K/min, although the DTG peak temperature reached 6 °C lower than that of other samples, the burnout temperature measured 565 °C, which was 5 °C higher than others. Therefore, oxidative activities of coke produced remained nearly the same under various heating rates. 3.2.2. Influence of O2 Concentration on Coke Characteristics. Figure 4 shows the results of analysis of coke produced with varying oxygen concentrations. As shown in Figure 4a, coke yield decreased with increasing O2 concentration, that is, from 29% in 2% O2 atmosphere to 24% in 20% O2 atmosphere. This phenomenon may be related to the enhancement of mild oxidation at 400 °C when O2 concentration increased. The oxygen elemental proportion of coke increased from 18% to 22%, whereas carbon proportion decreased with increasing O2 concentration from 2% to 15%. The oxygen proportion of cokes produced in 15% and 20% O2 atmosphere were similar. D

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 3. Yield, element, surface functional groups, and oxidative activity of coke formed in oxidizing atmosphere with various heating rates. (a) Coke yield and elemental analysis, (b) DRIFTS, and (c) DTG curve of coke oxidation.

3.2.3. Influence of Pressure on Coke Characteristics. Figure 5a compares the results between coke yield and elemental composition for the cokes obtained under 0.1 and 3 MPa. At a holding temperature of 350 °C, coke yield increased by 11% under 3 MPa but decreased by 1.7% at 400 °C under 3 MPa compared with that at 0.1 MPa. On the other hand, oxygen content of coke obtained under 3 MPa was 5% higher than that under 0.1 MPa at 350 °C. However, the content was nearly the same at 400 °C. In accordance with DRIFTS, all samples presented the same type of functional groups. Coking at 350 °C under 0.1 MPa was incomplete with a few remaining alkyl groups. However, when pressure increased to 3 MPa, these alkyl groups disappeared, indicating that coke formation temperature can be reduced with the increase in pressure. At the same time, DRIFTS showed the lower relative intensity of the carbonyl peak under 3 MPa and 400 °C in comparison with that at 350 °C. This condition agrees with elemental results, revealing the elimination of a part of the oxygen-containing functional groups. We compared the DTG curves of coke samples obtained under various pressures. Results showed that higher light-off temperature (around 420 °C) for coke obtained under high pressure owing to the increase in partial pressure of O2. This condition promoted the removal of active residual groups and increased coke formation conversion in LTO. By contrast, lightoff temperature (360 °C) was relatively low as a number of

Figure 4b demonstrates that surface functional groups, such as hydroxyl, carbonyl, and aromatic groups, were similar among the samples except for the coke produced with 2% O2. For the coke produced with 2% O2, a weak absorption peak of alkyl groups appeared, and absorption of carbonyl groups was relatively weaker than that of other samples. This condition resulted from the remarkably low O2 concentration. In turn, the amount of oxygen atom inserted into the carbon chain reduced, thus influencing polymerization and coking performance and reducing the amount of carbonyl groups. From results of TGA analysis of coke samples obtained at different O2 concentrations, as shown in Figure 4c, with increasing oxygen concentration, light-off temperature increased gradually, from 350 to 405 °C. For these occurrences, when O2 concentration was low, coke formation was incomplete, and numerous residual active groups remained on the coke surface. These active sites were easily oxidized, thus reducing light-off temperature. The peak temperature of the DTG curve of the samples was approximately 544 °C, whereas burnout temperatures were nearly the same (560 °C). Therefore, it could be seen that the active sites remained due to incomplete oxidation did not influence the subsequent coke oxidation process, and the coke samples obtained at varying O2 concentrations can be concluded to possess similar oxidative activities. E

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 4. Yield, element, and surface functional group of coke formed in oxidizing atmosphere with varying oxygen concentrations. (a) Coke yield and elemental analysis, (b) DRIFTS, and (c) DTG curve of coke oxidation.

alkyl groups remained in the coke sample obtained at 350 °C and under 0.1 MPa. Peak temperature of the DTG curve of all the samples approximated 544 °C, as shown in Figure 5c and the burn off temperatures are all around 560 °C. Therefore, oxidative activity of coke remained the same under different pressures, although light-off temperature varied among the samples owing to the existence of active groups, which were incompletely oxidized during LTO. 3.3. Analysis of Heavy-Oil Coking During LTO. According to the above-mentioned results, the influence of conversion conditions on heavy-oil coking in oxidizing atmosphere was mainly reflected on coke yield and characteristics but was observed minimally on oxidative activity. To comprehend heavy-oil coking and influence of conversion conditions, our study divided LTO into sections then analyzed the products of each portion. We heated oil-sand samples at a constant heating rate of 5 K/min until temperatures of 250, 300, 350, and 400 °C were reached in 5% O2 atmosphere. When a target temperature was achieved, the sample was held at this temperature for 1 h to allow all possible reactions to proceed. In such cases, the types of reaction that occurred below the holding temperature can be presented through analysis and comparison of solid residues. As some holding temperatures were low, and heavy oil may not completely coke, this experiment used residue proportion instead of coke yield to indicate the mass ratio of residue remained on the sand surface

to original oil. The same batch of oil-sand sample was used in all cases discussed in this section. Table 4 presents the changes in mass ratio and elemental composition of residues at varying holding temperatures. To easily analyze the released substances during different temperature stages, all the data in Table 4 regarded the total mass of C/H/O in original oil-sand sample as 100%. The contents of other elements in oil, such as N and S, are very low and negligible. For example, residue proportion is derived by the following formula: η=

⎛ αr ⎞ ⎛ αoil ⎞ mresidual moil / =⎜ ⎟ ⎟/⎜ ′ msand msand ⎝ 1 − αr ⎠ ⎝ 1 − αoil ⎠

(2)

where αr and αoil refer to the total mass percentage of C/H/O elements in solid residue and in original oil-sand mixture, respectively. As indicated in Table 4, with increasing holding temperature, weight loss of heavy oil increased. After treating the sample at 250 °C, weight loss reached 55.9%. According to the element equilibrium calculation, H/C ratio of the substance released to the gas phase is 2:1, corresponding to the chemical composition of light hydrocarbons. This result reveals that before 250 °C, heavy oil lost weight mainly by evaporation. At the same time, oxygen content increased by 5 times compared with that of original oil, proving the presence of oxygen-adding reactions F

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 5. Yield, element, and surface functional groups of coke formed in oxidizing atmosphere under 0.1/3 MPa. (a) Coke yield and elemental analysis, (b) DRIFTS, and (c) DTG curve of coke oxidation.

Table 4. Mass Ratio and Elemental Composition of Residues at Varying Holding Temperatures temperature °C

residue proportion (wt %)

C (wt %)

H (wt %)

O (wt %)

H/C/O mole ratio of released substance

room T 250 300 350 400

100.0 44.1 44.9 36.9 23.8

87.2 34.4 31.3 27.9 18.2

11.5 3.0 1.3 1.0 0.8

1.3 6.7 12.3 8.0 4.8

− 2/1/− 7/1/− 1/1/1 4/12/3

around 250 °C. Comparing the 300 °C residue with that obtained under 250 °C, residue proportion increased instead of decreasing. The mass ratio of C and H continually decreased, and H/C ratio of the released substance approximated 7:1, indicating that in addition to dealkylation, dehydrogenation also occurred. A certain amount of hydrogen was released as water formed by condensation, dehydration, and oxidation. Oxygen content continually increased as more active sites were exposed after cracking, and increasing temperature increased the extent of oxygen-adding reaction. Residue proportion decreased by 8% after increasing the holding temperature from 300 to 350 °C. H/C/O ratio of the released part was approximately 1:1:1. Decarbonylation possibly occurred, producing carbon oxides, small aldehydes and ketones, and steam.22 Residue proportion significantly decreased after increasing the holding temperature to 400 °C. H/C/O ratio of the released part was approximately 4:12:3. Consumption of H and O was lower than that of C, indicating that in addition to

decarbonylation, carbon oxidation in solid residue may have also occurred. Figure 6 shows the DRIFTS of solid residues. Compared with oil, alkyl infrared absorption intensity of residue decreased at 250 °C (2950, 2900, and 1400 cm−1), whereas carbonyl and hydroxyl absorption intensity increased (1700 and 3400 cm−1). These results proved the presence of oxygen-adding reactions. The decrement of alkyl absorption was related to the formation of light hydrocarbons caused by cracking in the presence of oxygen and the ensuing evaporation of light hydrocarbon. When holding temperature increased to 300 °C, the alkyl group was completely eliminated from the solid residue. In accordance with the results of the above-mentioned elemental analysis, carbon content varied slightly, implying that certain alkyl groups polymerized into cycloalkanes and aromatic rings instead of cracking and being released to the gas phase. DRIFTS also showed an increment in aromatic absorption (3100 and 1600 cm−1). Compared with the 300 °C residual, DRIFTS of the 350 and 400 °C residues showed that G

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Mild oxidation exhibits a more significant influence than cracking. In accordance with the above-mentioned analysis, during thermal conversion of heavy oil, light components of hydrocarbon first evaporated into the gas phase with increasing temperature. O2 partially oxidized a number of active hydrocarbons, forming several oxygen-containing functional groups. When temperature increased to 300 °C, dealkylation, dehydrogenation, and polymerization occurred, which eliminated the alkyl groups and increased the amount of aromatics. After holding the temperature at 350 °C, the amount of hydroxyl groups decreased because of polymerization and oxidation. A few unstable oxygen-containing functional groups cracked and were released into the gas phase. The functional groups of the solid residue were similar to those of coke, and the residue was toluene-insoluble, indicating that coke formation was nearly complete at 350 °C. At 400 °C, unstable oxygen-containing functional groups were further cleaved off and coke was partially oxidized. Conversion conditions affect heavy-oil thermal conversion at various stages. Hence, coke characteristics may differ. Heating rate influences the extent of oxygen-adding reactions, mild oxidation, and the amount of polymerization sites. With the increase in heating rate, oxygen-adding reaction time was shortened, and the extent of reaction decreased, thus reducing oxygen content. Reduction in oxygen content also decreased the amount of oxygen functional group, which served as the main polymerization site. Hence, coke yield decreased from 5 to 15 K/min. However, coke yield was also related to mild oxidation and cracking reactions. At a heating rate of 3 K/min, reaction time of these processes increased, thus reducing coke yield compared with that at 5 K/min. Therefore, coke yield increased initially then decreased with increasing heating rate from 3 to 15 K/min. As for the effect of O2 concentration, the rates of oxygen-adding reaction, dealkylation and mild oxidation of coke were influenced. With increasing O 2 concentration, oxygen content of coke increased owing to the increase in oxygen-adding reactions. Although the amount of oxygen surface functional group increased, the increase in O2 concentration significantly promoted dealkylation and mild oxidation reactions. Hence, coke yield decreased with increasing O2 concentration. Pressure exhibited an effect similar to that of O2 concentration. When total pressure was high, partial pressure of O2 increased, thus increasing the amount of dissolved O2 in the oil phase. Compared with 0.1 MPa, partial pressure of O2 increased by 30-fold at 3 MPa. Hence, the extent of oxygen-adding reactions increased significantly. With increasing amount of oxygen functional groups, which acted as polymerization sites, coke yield at 350 °C measured much higher than that at 0.1 MPa. At 400 °C, high partial pressure of O2 accelerated mild oxidation rate of coke. Given the oxidation of coke and functional groups, yield and oxygen content of coke obtained under 3 MPa at 400 °C decreased compared with that obtained under 0.1 MPa at 400 °C and under 3 MPa at 350 °C.

Figure 6. Surface functional groups of residues with varying holding temperatures.

absorption intensity of carbonyl groups gradually decreased as a result of decarbonylation. The same holds true with the hydroxyl groups because of polymerization and oxidation. DRIFTS were consistent with those of elemental analysis. On the basis of the results in section 3.2, coke yield decreased with increasing oxygen concentration or decreasing heating rate. This phenomenon is speculated to be caused by mild oxidation of coke in the high-temperature range during LTO. To verify the existence of this reaction, a solid residue sample, which was obtained in 5% O2 atmosphere at 5 K/min with a terminal temperature of 400 °C, was separately treated at 400 °C in 5% O2 and inert atmosphere for 1 h. Elemental composition and yield of solid residue were studied. As shown in Figure 7, after holding for 1 h in inert atmosphere, the yield

Figure 7. Yield and elemental analysis of residue treated in inert/ oxidizing atmosphere for 1 h.

decreased from 30% to 27%, whereas oxygen content decreased by 2%. This condition indicated that certain functional groups of the solid residue cracked at around 400 °C. After holding for 1 h in an oxidizing atmosphere, the yield decreased by 7.5% compared with the original amount. Carbon content decreased by 4%, whereas that of oxygen increased. These results indicated that mild oxidation of coke can occur at approximately 400 °C. Therefore, at around 400 °C during LTO, cracking of certain functional groups and mild oxidation of coke occurred, which caused the decrement in coke yield.

4. CONCLUSION This research studied the influence of conversion conditions on characteristics of coke obtained from heavy-oil thermal conversion. The main conclusions include the following. Yield and characteristics of coke obtained in inert and oxidizing atmospheres differed significantly. Compared with that obtained in inert atmosphere, coke yield increased by 5 times of that obtained in oxidizing atmosphere under 0.1 MPa. H

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

(5) Mahinpey, N.; Ambalae, A.; Asghari, K. In situ combustion in enhanced oil recovery (EOR): A review. Chem. Eng. Commun. 2007, 194 (8), 995−1021. (6) Freitag, N. P; Evidence that naturally occurring inhibitors affect the low-temperature oxidation kinetics of heavy oil. Journal of Canadian Petroleum Technology 2010, 49 (7), 36−41. (7) Vossoughi, S.; Bartlett, G. W.; Willhite, G. P. Development of a kinetic model for In-situ combustion and prediction of the process variables using TGA/DSC techniques. SPE Annual Technical Conference and Exhibition, New Orleans, LA, Sept 26, 1982; Society of Petroleum Engineers, 2013; p 11073.10.2118/11073-MS (8) Liu, D.; Song, Q.; Tang, J.; Zheng, R.; Yao, Q. Interaction between saturates, aromatics and resins during pyrolysis and oxidation of heavy oil. J. Pet. Sci. Eng. 2017, 154, 543−550. (9) Xu, Q.; Jiang, H.; Zan, C.; Tang, W.; Xu, R.; Huang, J.; Li, Y.; Ma, D.; Shi, L. Coke formation and coupled effects on pore structure and permeability change during crude oil in situ combustion. Energy Fuels 2016, 30 (2), 933−942. (10) Cinar, M.; Castanier, L. M.; Kovscek, A. R. Combustion kinetics of heavy oils in porous media. Energy Fuels 2011, 25 (10), 4438−4451. (11) Xu, Q.; Jiang, H.; Ma, D.; Chen, X.; Huang, J.; Shi, L. Pyrolysis of a Low Asphaltene Crude Oil under Idealized in Situ Combustion Conditions. Energy Fuels 2017, 31 (10), 10545−10554. (12) Drici, O.; Vossoughi, S. Study of the surface area effect on crude oil combustion by thermal analysis techniques. JPT, J. Pet. Technol. 1985, 37 (4), 731−735. (13) Guisnet, M.; Magnoux, P. Organic chemistry of coke formation. Appl. Catal., A 2001, 212 (1), 83−96. (14) Li, S.; Zhang, S.; Feng, Z.; Yan, Y. Coke Formation in the catalytic cracking of bio-oil model compounds. Environ. Prog. Sustainable Energy 2015, 34 (1), 240−247. (15) Martínez-Escandell, M.; Torregrosa, P.; Marsh, H.; RodríguezReinoso, F.; Santamaría-Ramírez, R.; Gómez-De-Salazar, C.; RomeroPalazón, E. Pyrolysis of petroleum residues: I. Yields and product analyses. Carbon 1999, 37 (10), 1567−1582. (16) Torregrosa-Rodriguez, P.; Martinez-Escandell, M.; RodriguezReinoso, F.; Marsh, H.; Gómez de Salazar, C.; Romero Palazón, E. Pyrolysis of petroleum residues: II. Chemistry of pyrolysis. Carbon 2000, 38, 535−546. (17) Guo, A.; Zhang, X.; Zhang, H.; Wang, Z.; Wang, Z. Aromatization of naphthenic ring structures and relationships between feed composition and coke formation during heavy oil carbonization. Energy Fuels 2010, 24, 525−532. (18) Wiehe, I. A. A phase-separation kinetic model for coke formation. Ind. Eng. Chem. Res. 1993, 32 (11), 2447−2454. (19) Santamaría-Ramírez, R.; Romero-Palazón, E.; Gómez-de-Salazar, C.; Rodríguez-Reinoso, F.; Martínez-Saez, S.; Martınez-Escandell, M.; Marsh, H. Influence of pressure variations on the formation and development of mesophase in a petroleum residue. Carbon 1999, 37 (3), 445−455. (20) Fassihi, M. R. Analysis of fuel oxidation in In-situ combustion oil recovery. Ph.D. Thesis, Stanford University, 1981. (21) Chen, Z.; Wang, L.; Duan, Q.; Zhang, L.; Ren, S. High-pressure air injection for improved oil recovery: low-temperature oxidation models and thermal effect. Energy Fuels 2013, 27 (2), 780−786. (22) Zhang, L.; Deng, J.; Wang, L.; Chen, Z.; Ren, S.; Hu, C.; Zhang, S. Low-temperature oxidation characteristics and its effect on the critical coking temperature of heavy oils. Energy Fuels 2015, 29 (2), 538−545. (23) Khansari, Z.; Gates, I. D.; Mahinpey, N. Low-temperature oxidation of Lloydminster heavy oil: Kinetic study and product sequence estimation. Fuel 2014, 115, 534−538. (24) Metzinger, T.; Huttinger, K. J. Investigation on the cross-linking of binder pitch matrix of carbon bodies with molecular oxygenPart 1. Chemistry of reactions between pitch and oxygen. Carbon 1997, 35 (7), 885−892. (25) Metzinger, T.; Huttinger, K. J. Investigations on the crosslinking of the binder pitch matrix of carbon bodies with molecular oxygenPart II. Kinetics and effects of cross-linking of the binder

Oxygen proportion increased, whereas that of carbon decreased. Oxygen existed in the form of oxygen-containing functional groups, such as hydroxyl and carbonyl. For coke produced by oxidation, DTG peak temperature was higher than that obtained through pyrolysis, whereas maximum reaction rate was lower, indicating that oxidative activity was poorer than that through pyrolysis. In oxidizing atmosphere, changes in conversion conditions, such as heating rate, O2 concentration, and pressure, significantly influenced characteristics of coke from heavy-oil thermal conversion. With the increase in heating rate, on one hand, coke yield increased initially then decreased. On the other hand, oxygen content of coke decreased gradually. With increasing O2 concentration, coke yield decreased, whereas oxygen content increased. Compared with that obtained at 0.1 MPa, coke yield and oxygen content increased at a holding temperature of 350 °C. However, these variables decreased at 400 °C. This phenomenon occurred as the increase in oxygen partial pressure reduced the reaction temperature during LTO. Although coke yield and characteristics showed several differences, oxidative activity was similar among the cokes produced with varying heating rates, O2 concentrations, and pressure. The influences of conversion conditions on heavy-oil coking are related to LTO. At 250 °C, evaporation of light components and oxygen-adding reactions dominated. At 300 °C, polymerization of the products of oxygen-adding reactions and dealkylation occurred. Further oxidation of hydroxyl groups, decarbonylation, and polymerization occurred at approximately 350 °C. At 400 °C, in addition to further decarbonylation, coke was mildly oxidized. Conversion conditions affected the progress of the above-mentioned reactions. Each reaction featured a different sensitivity to changes in conversion conditions. Thus, the coke obtained from heavy oil exhibited differing characteristics.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +86 10 62781740. ORCID

Qiang Song: 0000-0002-5484-3594 Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the fund from Xinjiang Oilfield Company, China. REFERENCES

(1) Moore, R. G.; Belgrave, J. D. M.; Ursenbach, M. G.; Laureshen, C. J.; Mehta, S. A.; Gomez, P. A.; Jha, K. N. N. In situ combustion performance in steam flooded heavy oil cores. J. Can. Pet. Technol. 1999, 38 (13), 1−9. (2) Xu, K. M. Basic research and application of in-situ combustion technology of oil production. Ph.D. Thesis, Northeast Petroleum University, 2012 (in Chinese). (3) Wang, Y.; Ren, S.; Zhang, L.; Peng, X.; Pei, S.; Cui, G.; Liu, Y. Numerical study of air assisted cyclic steam stimulation process for heavy oil reservoirs: Recovery performance and energy efficiency analysis. Fuel 2018, 211, 471−483. (4) Hascakir, B.; Glatz, G.; Castanier, L. M.; Kovscek, A. R. In-situ combustion dynamics visualized with x-ray computed tomography. SPE Journal 2011, 16 (03), 524−536. I

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels pitch matrix on the properties of the carbon bodies. Carbon 1997, 35 (7), 917−922. (26) Niu, B.; Ren, S.; Liu, Y.; Wang, D.; Tang, L.; Chen, B. Lowtemperature oxidation of oil components in an air injection process for improved oil recovery. Energy Fuels 2011, 25, 4299−4304. (27) Pei, X.; Hou, L. Effect of dissolved oxygen concentration on coke deposition of kerosene. Fuel Process. Technol. 2016, 142, 86−91. (28) Kok, M. V. Thermal behavior and kinetics of crude oils at low heating rates by differential scanning calorimeter. Fuel Process. Technol. 2012, 96, 123−127. (29) Varfolomeev, M. A.; Rakipov, I. T.; Isakov, D. R.; Nurgaliev, D. K.; Kok, M. V. Characterization and kinetics of Siberian and Tatarstan regions crude oils using differential scanning calorimetry. Pet. Sci. Technol. 2015, 33 (8), 865−871. (30) Kok, M. V.; Hughes, R.; Price, D. High pressure TGA analysis of crude oils. Thermochim. Acta 1996, 287, 91−99. (31) Anto-Darkwah, E.; Cinar, M. Effect of pressure on the isoconversional in situ combustion kinetic analysis of Bati Raman crude oil. J. Pet. Sci. Eng. 2016, 143, 44−53. (32) Adegbesan, K. O.; Donnelly, J. K.; Moore, R. G.; Bennion, D. W. Low-temperature oxidation kinetic parameters for in-situ combustion numerical simulation. SPE Reservoir Eng. 1987, 2 (4), 573−582. (33) Mailybaev, A. A.; Bruining, J.; Marchesin, D. Analytical Formulas for In-Situ Combustion. SPE Journal 2011, 16, 513−523. (34) Pang, K.; Xiang, W.; Zhao, C. Investigation on pyrolysis characteristic of natural coke using thermogravimetric and Fouriertransform infrared method. J. Anal. Appl. Pyrolysis 2007, 80, 77−84. (35) Kissinger, H. E. Reaction kinetics in differential thermal analysis. Anal. Chem. 1957, 29 (11), 1702−1706.

J

DOI: 10.1021/acs.energyfuels.8b00098 Energy Fuels XXXX, XXX, XXX−XXX