Article pubs.acs.org/EF
Seven-Lump Kinetic Model for Non-catalytic Hydrogenation of Asphaltene Qiang Sheng,† Gang Wang,*,† Qiyuan Zhang,† Chengdi Gao,‡ Ailin Ren,† Mengchao Duan,† and Jinsen Gao† †
State Key Laboratory of Heavy Oil Processing, College of Chemical Engineering and ‡College of Science, China University of Petroleum, Beijing 102249, People’s Republic of China S Supporting Information *
ABSTRACT: Non-catalytic hydrogenation with a hydrogen donor is a beneficial way for effective conversion of asphaltene to distillate with minimal coke formation. In this work, detailed product distribution, which includes gas, light oil [initial boiling point (IBP)−350 °C], middle oil (350−540 °C), heavy oil (>540 °C), asphaltene, and coke, obtained from non-catalytic hydrogenation of asphaltene with tetralin as a hydrogen donor, was investigated in an autoclave. The effects of reaction conditions, including reaction time, reaction temperature, and hydrogen donor/asphaltene weight ratio, on asphaltene conversion, detailed product distribution, liquid product yield, and liquid product selectivity were studied. Results showed that through controlling the reaction condition, asphaltene conversion and total liquid yield reached 72.72 and 70.34 wt %, respectively, and produced only 2 wt % coke and 0.34 wt % gas. We then developed a seven-lump kinetic model, including an active hydrogen lump to describe the reaction behaviors of asphaltene hydroliquefaction. Activation energies ranged from 106.07 to 237.50 kJ mol−1. The activation energies of the main reaction that asphaltene decomposed and hydrogenated by active hydrogen to produce heavy oil and middle oil were 106.07 and 109.06 kJ mol−1, respectively, which were lower than those of thermal cracking. The activation energy of distillate formation from active hydrogen combined with macromolecule radicals was 143.78 kJ mol−1. The detailed product yield predicted by the developed seven-lump kinetic model exhibited good consistency with the experimental data.
1. INTRODUCTION
process for efficiently converting asphaltene to distillates with less coke formation. Several researchers have been making efforts to improve the efficient conversion of asphaltene to distillates and investigate the structure of asphaltene before and after hydrogenation by hydrogen donors. Al-Samarraie et al.11 used a variety of analytical methods to characterize the structure of asphaltene after pyrolysis with tetralin (THN) at 450 °C. Ignasiak and coworkers12 investigated the thermal cracking of asphaltene under mild conditions, i.e., the reaction temperature ranging between 195 and 390 °C, to study the interaction of Athabasca asphaltene with THN as a hydrogen-donating solvent. RuedaVelásquez et al.13,14 characterized the building blocks of asphaltene by analyzing the structure of liquid products of asphaltene hydrogenated by THN under favorable conditions. Those studies predominantly focus on the variation of the asphaltene structure before and after reaction; however, the product distribution was neglected. The product distribution, including gas, maltene, asphaltene, and coke, in pyrolysis and hydropyrolysis with THN as the solvent was investigated by Savage et al.15 The product distribution of asphaltene under non-catalytic hydrogenation conditions was also investigated by Jin et al.6 However, the liquid products were regarded as a whole, which was not separated in detail by these researchers. However, the detailed product distribution is important for in-
The effective conversion of asphaltene has become the crucial problem of upgrading heavy oil. However, in traditional processes, the high molecular weight, high polarity, complicated structure, and high heteroatom contents result in asphaltene mainly converted to coke.1−3 Therefore, enhancing the reaction of asphaltene converted to distillate with less coke formation is the most valuable way to improve the conversion of heavy oil.4 At presents, asphaltene treatment processes include thermal cracking and hydrogenation. The hydrogenation process can be further divided into catalytic hydrogenation with gaseous hydrogen and non-catalytic hydrogenation with a hydrogen donor as the source of hydrogen. In the thermal cracking process, over 50 wt % asphaltene was converted to coke under 400 °C for 1 h.1,5 However, in the hydrogenation process, under the same reaction temperature of 400 °C, the coke yield can be significantly reduced and, at the same time, heighten the liquid product yield.6 Additionally, the contact efficiency of gaseous hydrogen and asphaltene extremely decreased because of the low solubility of hydrogen in the liquid phase during the process of catalytic hydrogenation.7 In comparison to catalytic hydrogen, in the non-catalytic hydrogenation process, the contact efficiency of active hydrogen with asphaltene is higher. Furthermore, in the non-catalytic hydrogenation process, the coke precursors can be dissolved by a hydrogen donor,8 enhancing the stability of the colloidal system9 and reducing coke formation.10 In conclusion, in comparison to thermal cracking, non-catalytic hydrogenation is a more competitive © 2017 American Chemical Society
Received: March 1, 2017 Revised: April 17, 2017 Published: April 19, 2017 5037
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels
temperature. This pressure was set to enhance the solubility of hydrogen in THN. The temperature was increased up to the set reaction temperature within 50 min and then kept stable for a set time while stirring at 300 revolutions/min. The reaction was terminated by introducing condensate water to the inside cooling coil.6 After reaction, the gas was collected by the displacement method and then released to a 1 L capacity gas bag for further analysis by gas chromatography. The liquid mixture was separated into liquid products (including light oil, middle oil, and heavy oil), asphaltene, and coke, which is displayed in Figure S1 of the Supporting Information. The liquid products were separated by dissolving the liquid mixture containing asphaltene and coke in n-heptane at a ratio of 40:1 (volume/weight) n-heptane/liquid mixture and stirring at 60 °C for 3 h. After 12 h, the sample was filtered under vacuum using three layers of medium-speed quantitative filter paper (pore size distribution of 30−50 μm) and rinsed with n-heptane until the filtrate was clarified. The solid mixture, including asphaltene and coke obtained in the filter paper, was dried off in an oven at 80 °C for 24 h. The solid-free sample was concentrated by a distillation apparatus to obtain a solvent-free sample (liquid product) for further analysis by simulated distillation for determination of the weight percentages of light oil, middle oil, and heavy oil. The distillation conditions were 100 °C under atmospheric pressure for n-heptane and 75 °C under 200 Pa for THN and its products. Inevitable evaporation of these solvents is joined with evaporation of parts of lighter end products. Coke was separated by dissolving the dried solid mixture in toluene at a ratio of 40:1 (volume/weight) toluene/solid sample and stirring at 60 °C for 3 h. After 12 h, the sample was filtered in vacuum using three layers of medium-speed quantitative filter paper (pore size distribution of 30−50 μm) and rinsed with toluene until the filtrate was colorless. Coke obtained in the filter paper was dried off in an oven at 100 °C for 24 h. The filtrate, including asphaltene and toluene, was distillated by a simple distillation apparatus to remove the large amount of toluene. Asphaltene with little toluene was dried off in an oven at 110 °C for 24 h. The coke yield was calculated according to the increase in the filter paper weight. The mass of asphaltene after reaction was calculated on the basis of the increase in weight of the conical flask. The yield of gas, asphaltene, liquid products (including light oil, middle oil, and heavy oil), and coke was calculated from eq 1
depth exploration of the reaction pathways of asphaltene conversion. Therefore, it should be given considerable attention. Great efforts have also been made to investigate the kinetic behavior of asphaltene conversion under different processes, including thermal cracking,16−21 thermal hydrocracking,22 catalytic hydrocracking,22,23 and thermal cracking in supercritical water (SCW).4 However, the kinetics of asphaltene conversion under non-catalytic hydrogenation conditions, such as selecting THN as a hydrogen donor, were seldom investigated. The kinetic lump model of thermal cracking or catalytic hydrocracking of asphaltene primarily includes threeand four-lump models. A three-lump model, which involved parallel reactions for generation of oil plus gas and coke, was proposed by Martinez et al.18 A modified three-lump model, which involved parallel reactions of asphaltene to liquid oil and gas plus coke and a consecutive reaction of liquid oil to gas plus coke, was proposed by Zhao et al.22 A four-lump model, which included asphaltene, maltene, gas, and coke, was proposed by Yasar et al.19 and Li et al.4 However, the products of maltene or oil derived from asphaltene were not separated further. Thus, the detailed product distribution is significant for accurately predicting the product distribution of asphaltene decomposition. Therefore, in this paper, we attempt to establish a noncatalytic hydrogenation condition by dissolving the asphaltene in THN and with a certain hydrogen pressure to enhance the solubility of hydrogen in THN. This process are referred to as non-catalytic hydrogenation of asphaltene. Additionally, to obtain a maximum liquid product yield and more clearly understand the reaction pathways, the properties of hydroliquefaction of asphaltene under non-catalytic hydrogenation conditions were investigated. The effects of operating conditions, including reaction time, reaction temperature, and hydrogen donor/asphaltene weight ratio (HD/Asp), on the detailed product distribution, such as gas, light oil [initial boiling point (IBP)−350 °C], middle oil (350−540 °C), heavy oil (>540 °C), and coke, were investigated. On the basis of these studies, a new seven-lump kinetic model, including a hydrogen radical lump, was developed to describe the reaction pathways of hydroliquefaction of asphaltene under non-catalytic hydrogenation conditions.
yieldi (wt %) =
mi × 100 masp,1
(1)
where i is gas, asphaltene, liquid product, light oil, middle oil, heavy oil, and coke, m is the mass of the sample, and asp denotes asphaltene. The mass of the liquid products was calculated by the subtraction method as eq 2
mliquid (g) = masp,1 − mgas − mcoke − masp,2
2. EXPERIMENTAL SECTION 2.1. Feedstock. In this study, paraffin-based asphaltene was obtained by precipitating propane de-oiled end cut (DOE) with nheptane. The propane DOE was mixed with n-heptane at a ratio of 1:40 (mass/volume) DOE/n-heptane, and then the mixture was stirred for 3 h with a magnetic stirrer. Asphaltene was filtered out under vacuum using quantitative filter paper. Asphaltene was dried at room temperature for 48 h to evaporate residual n-heptane. The properties of paraffin-based C7 asphaltene are listed in Table S1 of the Supporting Information. 2.2. Equipment and Procedure. The experiments were performed in a 250 mL stainless-steel autoclave reactor, which was equipped with a mechanical agitator, automatic temperature control system, and inside cooling coil. The schematic diagram of the experimental setup was reported previously.6 A typical experiment was performed as follows: the reactor was loaded with approximately 100 g of the total sample, including C7 asphaltene and THN, at a set ratio. The reactor was closed and tested for leakage at 8 MPa with nitrogen at room temperature. Once the reactor was leak-free, it was pressurized with hydrogen to 5 MPa and vented to the atmosphere 3 times. An appropriate amount of hydrogen was pressured into the reactor, so that the pressure can reach 7 MPa at the initial designed reaction
(2)
where 1 and 2 represent before and after reaction. The mass of distillates, including light oil, middle oil, and heavy oil, was calculated as follows:
mx = wxmliquid
(3)
where x is light oil, middle oil, and heavy oil. The conversion of asphaltene was calculated as eq 4. conversionasp (wt %) =
masp,1 − masp,2 masp,1
× 100 (4)
The concentration of hydrogen radical released by THN was calculated by the following steps: THN and its dehydrogenation products obtained from distillation under conditions of 75 °C and 200 Pa were analyzed by a gas chromatograph equipped with a flame ionization detector using argon as the carrier gas to measure the relative amount of the components of THN and anthracene. The conversion of THN was calculated according to the gas chromatograph data. The hydrogen radical released by THN was calculated by the conversion of THN as eq 5 5038
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels H0 =
m THNx THN × 4 M THN
(5)
where H0 is the mass of hydrogen radical released by THN, mTHN is the mass of THN, xTHN is the conversion of THN, and MTHN is the molecular weight of THN. 2.3. Analytical Methods. The gas sample was analyzed by an Agilent 6860N gas chromatograph (according to the ASTM D1946 standard), which was equipped with two detectors: one is the flame ionization detector, and the other is the thermal conductivity detector. Nitrogen was used as the carrier gas to measure the volume percentage of H2, C1−C6 hydrocarbons, CO, and CO2. The state equation of ideal gas was used for converting the obtained volume percentage data to mass percentage. The weight percentages of light oil (IBP−350 °C), middle oil (350−540 °C), and heavy oil (>540 °C) were obtained according to simulated distillation results. An Agilent 6890 gas chromatograph was used for simulated distillation according to the ASTM D7169-05 standard.
Figure 2. Influence of the reaction time on the product distribution.
reaction of generating gas and coke was enhanced and the yield of distillates tended to be steady, indicating that, with the increasing reaction time, side chains and small aromatic ring structures from the asphaltene molecule were broken and the remaining molecules obtained more aromatics with a high condensation degree.1,6,24 Therefore, the remaining molecules were easy to condense to coke rather than be converted to distillate. This occurrence demonstrates that, to maximize the distillates, the reaction time should be reasonably controlled. Additionally, the varying yield of liquid products and liquid product selectivity (Figure 3) indicated that the reaction time should be controlled within 7 h again.
3. RESULTS AND DISCUSSION In the process of inferior heavy oil, especially for the process of asphaltene conversion, suitable operating parameters are essential to maximize the total liquid yield and minimize the coke yield. Therefore, the effects of the reaction time, reaction temperature, and HD/Asp on the conversion of asphaltene, detailed product distribution, and selectivity of liquid products were investigated. 3.1. Effects of the Reaction Time. When the reaction temperature and HD/Asp were maintained at 400 °C and 3, respectively, the effects of the reaction time on the conversion of asphaltene, product distribution, and selectivity of liquid products were investigated at varied times from 1 to 9 h. The effects of the reaction time on asphaltene conversion and product distribution are plotted in Figures 1 and 2,
Figure 3. Influence of the reaction time on the liquid product yield and its selectivity.
3.2. Effects of the Reaction Temperature. When the reaction time and HD/Asp were kept at 7 h and 3, respectively, the effects of the reaction temperature on the conversion of asphaltene, product distribution, and selectivity of liquid products were investigated with the temperature varying from 380 to 420 °C. Figures 4 and 5 illustrate the effects of the reaction temperature on the asphaltene conversion and product distribution, respectively. The data indicated that the asphaltene conversion changed little when the reaction temperature increased from 380 to 390 °C; however, the asphaltene conversion went up rapidly when the reaction temperature exceeded 390 °C. The yield of gas varied little with changing the reaction temperature, but the yield of coke suddenly increased to 32.28 wt % with the temperature increasing from 400 to 420 °C, indicating that, at a higher reaction temperature, asphaltene was more likely to condense to coke. The yields of light oil and middle oil increased gradually with increasing the reaction temperature. Especially, the yield of heavy oil first
Figure 1. Influence of the reaction time on the conversion of asphaltene.
respectively. The asphaltene conversion and the yield of gas, coke, light oil (IBP−350 °C), and middle oil (350−540 °C) increased gradually with increasing the reaction time. This result is mainly because asphaltene not only decomposes to gas and distillates but also condenses to form coke by the combination of macromolecular radicals. Furthermore, the lighter oil (such as middle oil and light oil) can also be produced from the secondary cracking reaction of heavy oil (>540 °C), thereby resulting in the initial increase of the yield of the heavy oil and then kept steady with increasing the reaction time. However, the secondary cracking reaction of heavy oil was not serious, which did not cause a decrease of the heavy oil yield. When the reaction time was longer than 7 h, the 5039
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels
below 400 °C can maximize the liquid product yield with less than 1.50 wt % coke yield. 3.3. Effects of HD/Asp. When the reaction time and reaction temperature were maintained at 7 h and 400 °C, respectively, the effects of HD/Asp on asphaltene conversion, product distribution, and selectivity of liquid products were investigated within the range of 1−7. The effects of HD/Asp on asphaltene conversion and product distribution are shown in Figures 7 and 8, respectively.
Figure 4. Influence of the reaction temperature on the conversion of asphaltene.
Figure 7. Influence of the solvent ratio on the conversion of asphaltene.
Figure 5. Influence of the reaction temperature on the product distribution.
increased to the maximum and then gradually decreased with increasing the reaction temperature. This phenomenon can be illustrated by the famous parallel−sequential reaction mechanism;4 heavy oil, as the middle product, undergoes secondary cracking at higher reaction temperatures (>400 °C) to form light oil and middle oil.1,5,15,19,25 Variation in the yield and the selectivity of liquid products with an increased reaction temperature are given in Figure 6. Data showed that the yield of liquid products increased, while the selectivity of liquid products decreased when the reaction temperature was higher than 400 °C. This finding indicated that the maximum liquid product yield cannot be obtained at a high reaction temperature. Therefore, controlling the reaction temperature
Figure 8. Influence of the solvent ratio on the product distribution.
Data illustrated that asphaltene conversion and coke yield decreased with increasing HD/Asp from 1 to 3. This observation is mainly due to HD/Asp being equal to 1, active hydrogen in the reaction system is insufficient for combining with massive macromolecular radicals. Consequently, the macromolecular radicals condense with each other and form large amounts of coke. When HD/Asp increases to 3, asphaltene is well-dissolved in the hydrogen donor and the colloidal system is more stable than that formed at the ratio of 1. At the same time, the macromolecular radicals are fully captured by enough active hydrogen released by the hydrogen donor, resulting in the yield of coke decreasing and the conversion of asphaltene declining. When HD/Asp was increased to 7, the conversion of asphaltene and yield of distillates (light oil, middle oil, and heavy oil) gradually increased and then approached a stable level. The yield of light oil initially declined and then gradually increased to an equilibrium level with increasing HD/Asp from 1 to 7. This observation can be attributed to the fact that the secondary cracking reaction of middle oil and heavy oil is serious at the ratio of 1, leading to the yield of light oil being higher than that
Figure 6. Influence of the reaction temperature on the liquid product yield and its selectivity. 5040
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels
parallel−sequential reaction. Asphaltene can be decomposed to heavy oil, middle oil, and gas and can be condensed to coke. The secondary cracking reaction occurred in heavy oil and middle oil lumps. The yield of coke produced in the process of hydrogenation of asphaltene was obviously lower than that of asphaltene pyrolysis. To simplify the calculations, we assumed that (1) coke formed only from asphaltene condensation, (2) no gas was generated in the secondary cracking reaction of middle oil, (3) the active hydrogen radicals only acted on macromolecular radicals produced by asphaltene, (4) light oil was not further cracked, (5) without catalyst addition, gas hydrogen did not participate in asphaltene conversion, and (6) the volume of the reaction system was not changed. Additionally, some researchers found that the reaction of asphaltene thermal cracking was a first-order reaction,4,27,28 whereas other researchers found that it was a second-order reaction.18 In this research, all reaction pathways were assumed as first-order reactions. A seven-lump model with nine reactions was established to describe the reaction pathways of noncatalytic hydrogenation of ashphaltene with THN as a hydrogen donor in Figure 10.
at the ratio of 3. When HD/Asp exceeds 1, the secondary cracking reaction is significantly reduced as a result of the existence of large amounts of hydrogen radicals. The gas yield was less than 0.5 wt % and only slightly changed. As shown in Figure 9, the yield of total liquid products increased to the
Figure 9. Influence of the solvent ratio on the liquid product yield and its selectivity.
maximum at the ratio of 5 and the selectivity of liquid products remained stable beyond 5, indicating that this was the suitable HD/Asp. From the investigation about the effects of reaction conditions, including reaction time, reaction temperature, and HD/Asp, on the conversion of asphaltene, product distribution, and selectivity of liquid products, results showed that the reaction of non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor was a typically parallel−sequential reaction. Asphaltene as a reactant mainly underwent breakage of its side chains and weak bonds between aromatic sheets. Afterward, large free-radical fragments derived from the breakage of weak bonds were stabilized by the active hydrogen radicals released from the hydrogen donor THN and then formed distillates, including heavy oil, middle oil, and light oil. Additionally, the large free-radical fragments, which were not stabilized by the active hydrogen radicals, condensed to form coke. However, the reaction of forming coke was significantly weakened as a result of the existence of large amounts of active hydrogen radicals. Therefore, the yield of coke obviously decreased in comparison to that of the thermal cracking reaction.1,5 Furthermore, middle oil and light oil were also produced by the secondary cracking reaction of the middle product of heavy oil. Meanwhile, the secondary cracking reaction may also occur for middle oil.
Figure 10. Kinetic scheme for the seven-lump model of non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor.
4.2. Mathematical Models. The reaction was performed in an autoclave reactor with rapid stirring; hence, the batch reactor can be taken as the ideal reactor. In this batch reactor, the reactants or products during the reaction neither inflowed nor outflowed, that is
4. REACTION KINETICS 4.1. Model Description. A kinetic model is essential to describe asphaltene conversion behavior. Given the complexity of asphaltene and its produced products, a lump strategy was adopted for describing asphaltene conversion.26 Accordingly, the liquid products were classified into light oil (IBP−350 °C) lump, middle oil (350−540 °C) lump, and heavy oil (>540 °C) lump. Gas and coke were taken as gas lump and coke lump. Particularly, active hydrogen released by THN played an important role in inhibiting coke formation by quenching macromolecular radicals; therefore, active hydrogen was also considered as one lump. As a result of the existence of hydrogen radicals, the conversion reactions of asphaltene were considered as irreversible reactions. On the basis of the above research, the asphaltene conversion process was a typical
Fj0 = Fj = 0
(6)
where F is the mass flow rate. The resulting general mole balance on species j is
dNj dt
=
∫
V
r j dV
(7)
where N represents the amount of substance j, V represents the volume of the mixture, rj represents the reaction rate of j, and t represents the reaction time. Considering that the reactor interior was rapidly stirred, the mixture was perfectly mixed, so that no variation existed in the rate of reaction throughout the reactor volume; hence, rj can be 5041
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels Table 1. Kinetic Rate Constants of the Seven-Lump Kinetic Model reaction temperature (°C) reactiona
kinetic rate constant (h−1)
Asp → HO Asp → MO Asp → G Asp → C HO → MO HO → LO HO → G MO → LO H• + Asp• → O
k1 k2 k3 k4 k5 k6 k7 k8 k9
390 2.976 1.995 1.005 × 2.390 × 0.0496 5.667 × 1.940 × 0.177 0.0177
10−4 10−4 10−3 10−10
400 3.961 2.677 1.618 × 3.780 × 0.0705 8.614 × 3.680 × 0.268 0.0261
10−4 10−4 10−3 10−10
410 5.2275 3.5601 2.567 × 10−4 5.880 × 10−4 0.0993 0.0129 6.850 × 10−10 0.4011 0.0381
a Asp, asphaltene; HO, heavy oil; MO, middle oil; LO, light oil; G, gas; C, coke; H•, active hydrogen; Asp•, macromolecular radicals derived from asphaltene cracking; and O, oil, including heavy oil, middle oil, and light oil.
dy3
taken out of the integral. Then, after integration of eq 7, the mole balance can be written as eq 8.
dNj dt
= rjV
dt dy4
(8)
dt
The maximum yield of gas is less than 0.75 wt %, which has little effect on the volume of the reaction system. Therefore, the volume is taken as a constant; that is, V = V0, where V0 represents the initial volume of reactants. Equation 8 can be expressed as dCj d(NA /V0) 1 dNA rj = = = V0 dt dt dt
dy5 dt dy6 dt
dy7
(9)
dt
Additionally, the reaction rate rj also can be expressed as the power law model rj = −kCA aC Bb
dt
= −kCA aC Bb
(10)
(11)
Instead of using the molar concentration, the mass concentration was used for calculating the kinetic parameters. Hence, the rate equation can be written as dyi dt
= −kiyi m yj n
(12)
where yi and yj are the mass concentrations of i and j lumps, respectively, ki is the rate constant of the i reaction, and m and n are the reaction orders of reactions i and j, respectively. Therefore, according to the reaction network of the seven-lump model, the mathematical equations of the kinetic model can be written as follows: dy1 dt dy2 dt
= −(k1y1y7 + k 2y1y7 + k 3y1 + k4y1)
(13)
= k1y1y7 − (k5y2 + k6y2 + k 7y2 )
(14)
(15)
= k6y2 + k 8y3
(16)
= k 3y1 + k 7y2
(17)
= k4y1
(18)
= −k 9y7
(19)
We have to point out that the cracking reaction of THN was not considered. Even though this will restrict the application of the model, the model can still be used to describe the behaviors of asphaltene conversion. 4.3. Estimation of Kinetic Parameters. To solve the above kinetic equations, the software of MATLAB was employed and compiled the program for calling the fourthand fifth-order Runge−Kutta numerical method.8,29 The genetic algorithm method was used for optimizing the kinetic parameters.30 The rate constants were estimated according to the experimental data on the non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor. The rate constants for each reaction lump at 390, 400, and 410 °C are listed in Table 1. The reaction of asphaltene to produce heavy oil was the main reaction with the largest rate constant, followed by the reaction of asphaltene to form middle oil. This finding indicated that the existence of the hydrogen donor, THN, favored for asphaltene conversion to distillate. Notably, the rate constants of k3 and k4, which represent the reaction of the production of end products, gas and coke, are 4 orders of magnitude lower than k1 and k2, demonstrating that the condensation reaction (which may produce coke) was significantly weakened. Additionally, the reaction of heavy oil to form gas can be ignored because the rate constant of k7 was small than other rate constants. The values of k5 and k8 showed that the secondary reaction of middle oil cracking to form light oil was stronger than the formation of middle oil from heavy oil. The rate constants of k9 representing the function of active hydrogen to asphaltene were 0.0177, 0.0261, and 0.0381 h−1 at the temperatures of 390, 400, and 410 °C, respectively. The activation energies and frequency factors were calculated according to the Arrhenius equation, and the results are listed
where k denotes the reaction rate constant, CA and CB denote the molar concentrations of reactants A and B, respectively, and a and b represent the reaction orders of reactants A and B, respectively. Combining eqs 9 and 10, we can obtain eq 11. dCj
= k 2y1y7 + k5y2 − k 8y3
5042
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels
higher than that of asphaltene to maltene, indicating that the reaction rate of asphaltene formation was much faster than that of asphaltene decomposition. These results indicated that the non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor was more suitable for asphaltene conversion and the asphaltene lump favored the reaction path for generating heavy oil and middle oil. Table 4 lists the activation energies for different processes of asphaltene conversion, including thermal cracking and non-
in Table 2. The activation energies ranged from 106.07 to 237.50 kJ mol−1. The activation energies of the main reaction Table 2. Apparent Activation Energies and Frequency Factors of the Seven-Lump Model non-catalytic hydrogenation reactiona Asp → HO Asp → MO Asp → G Asp → C HO → MO HO → LO HO → G MO → LO H• + Asp• → O
activation energy (kJ mol−1)
frequency factor (h−1)
E1 E2 E3 E4 E5 E6 E7 E8 E9
A1 A2 A3 A4 A5 A6 A7 A8 A9
106.07 109.06 176.56 169.17 130.78 155.43 237.50 153.63 143.78
6.70 7.78 8.13 5.06 9.92 9.92 9.91 2.24 3.75
× × × × × × × × ×
108 108 109 109 108 109 108 1011 109
Table 4. Activation Energies of Asphaltene Conversion in Different Processes cracking under nitrogen19
a
Asp, asphaltene; HO, heavy oil; MO, middle oil; LO, light oil; G, gas; C, coke; and H•, active hydrogen.
that asphaltene cracking and macromolecule radicals combined with active hydrogen to form heavy oil and middle oil were 106.07 and 109.06 kJ mol−1, which were lower than those of thermal cracking17,19 but higher than those of catalytic hydrocracking.31 This finding can be attributed to the highly efficient hydrogenation method of non-catalytic hydrogenation with active hydrogen released by THN. The activation energy of active hydrogen combining with macromolecule radicals to form heavy oil and middle oil was 143.78 kJ mol−1. The activation energies of forming coke (E4) and gas (E3 and E7) were high, indicating that a high temperature favored coke formation. Therefore, the relatively low temperature benefited the formation of more liquid products. The rate constants of thermal cracking and non-catalytic hydrogenation of asphaltene under 400 °C were compared, and the results are listed in Table 3. The rate constant of thermal
rate constant (h−1)
reaction networkb
rate constant (h−1)
Asp → M Asp → C Asp → G M→G M → Asp
48.943 2.927 0.301 3.000 × 10−4 81.565
Asp → MHM Asp → C Asp → G MM → G MHM → Asp
2.677−3.961 3.780 × 10−4 1.618 × 10−4 3.680 × 10−10 did not occur
activation energy (kJ mol−1)
reaction networkb
activation energy (kJ mol−1)
Asp → M Asp → C Asp → G M→G M → Asp
163.16 144.71 162.71 160.03 100.00
Asp → MHM Asp → C Asp → G MM → G MHM → Asp
106.07−109.06 169.17 176.56 237.50 did not occur
Asp, asphaltene; M, maltene (n-heptane solube); C, coke; and G, gas. Asp, asphaltene; MHM, maltene (heavy oil and middle oil); MM, maltene (heavy oil); C, coke; and G, gas. b
catalytic hydrogenation. The activation energy of thermal cracking was obtained from the literature19 for comparison. The activation energy for generating heavy oil and middle oil (106.07−109.06 kJ mol−1) during the process of non-catalytic hydrogenation was lower than that of thermal cracking (163.16 kJ mol−1). Moreover, the activation energies of generating coke and gas from asphaltene during the process of non-catalytic hydrogenation (169.17 and 176.56 kJ mol−1) were higher than those of thermal cracking (144.71 and 162.71 kJ mol−1). Additionally, the activation energy of generating gas from middle oil during non-catalytic hydrogenation (237.50 kJ mol−1) was much higher than that of thermal cracking (160.03 kJ mol−1). Furthermore, the data above showed that the reaction that most easily proceeded was maltene condensing to form asphaltene in thermal cracking with the lowest activation energy (100.00 kJ mol−1). However, this reaction did not occur under the conditions of non-catalytic hydrogenation. These results substantiated the fact that asphaltene had diverse reaction characteristics under different processes. Therefore, a suitable process, such as non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor, can achieve the goal of effectively converting asphaltene to liquid products with little coke and gas formation. 4.4. Comparison of Experimental and Predicted Values. The comparison of experimental data and estimated data according to the developed kinetic model is presented in Figure 11. The comparison results showed that the estimated values were consistent with experimental data, indicating that the developed kinetic model was able to estimate the experimental values. Figure 12 showed the estimated and experimental yields of each lump as a function of the reaction time (1−9 h), at a constant reaction temperature, HD/Asp, and hydrogen pressure of 390 °C, 3, and 7 MPa, respectively. Expect for the coke yield after 5 h, the experimental values were close to the estimated yields. Because of the lower coke yield than asphaltene conversion, despite a little deviation from coke yield prediction, the model can also predict the product distribution well. In conclusion, the kinetic model can be used
non-catalytic hydrogenation
reaction networka
reaction networka
a
Table 3. Kinetic Rate Constants for Different Processing of Asphaltene at 400 °C cracking under nitrogen19
non-catalytic hydrogenation
a Asp, asphaltene; M, maltene (n-heptane soluble); C, coke; and G, gas. bAsp, asphaltene; MHM, maltene (heavy oil and middle oil); MM, maltene (heavy oil); C, coke; and G, gas.
cracking of asphaltene was cited from the literature.19 In comparison to asphaltene thermal cracking, the rate constants of the main reaction of asphaltene to heavy oil and middle oil were much lower than those of asphaltene converting to maltene. However, the rate constant of asphaltene to coke was 4 orders of magnitude lower than that in thermal cracking. Additionally, the reaction of heavy oil and middle oil condensing to form asphaltene was not discovered in the non-catalytic hydrogenation process, but the rate constant of maltene to form asphaltene in thermal cracking was much 5043
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels
Figure 11. Parity plots for experimental and calculated values of each component produced by the reaction of asphaltene hydroliquefaction under non-catalytic hydrogenation conditions with the temperature ranging from 390 to 410 °C.
cracking. The activation energies of forming coke and gas are higher than that of thermal cracking of asphaltene. The activation energy of the process of active hydrogen combining with macromolecule radicals to form heavy oil and middle oil was 143.78 kJ mol−1. The detailed product yields predicted by the seven-lump model were in good agreement with the experimental data.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.energyfuels.7b00608. Properties of paraffin-based asphaltene (Table S1) and separation scheme of the liquid mixture (Figure S1) (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*Telephone: 8610-8973-3085. E-mail:
[email protected]. ORCID
Qiang Sheng: 0000-0001-7382-886X Gang Wang: 0000-0001-7315-2070
Figure 12. Predicted (lines) and experimental (plots) yields as a function of the reaction time.
Notes
The authors declare no competing financial interest.
■
to describe the reaction behavior of non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor.
ACKNOWLEDGMENTS The authors acknowledge the financial support provided by the National Natural Science Foundation of China (21476259), the State Key Program of National Natural Science Foundation of China (21336011), the National Natural Science Foundation of China Petrochemical Joint Fund (Class A) Project (U1662105), and the Science Foundation of China University of PetroleumBeijing (2462015YQ0310).
5. CONCLUSION The aim of effectively converting asphaltene to distillates with minimal coke formation can be achieved by controlling the reaction conditions. Under suitable conditions, the asphaltene conversion and total liquid yield can reach 72.72 and 70.34 wt %, respectively, and with only 2 and 0.34 wt % coke and gas yield, respectively. The reaction of non-catalytic hydrogenation of asphaltene with THN as a hydrogen donor was a typical parallel−sequential reaction. The developed seven-lump kinetic model, including active hydrogen, can accurately predict the reaction behavior. The model contained asphaltene, heavy oil, middle oil, light oil, gas, coke, and hydrogen radical as lumps and included nine kinetic constants. The reaction rate constants (390, 400, and 410 °C) and apparent activation energies were estimated. The activation energies ranged from 106.07 to 237.50 kJ mol−1. The reactions of asphaltene to produce heavy oil and middle oil were the main reactions with activation energies of 106.07 and 109.06 kJ mol−1, respectively, which were lower than those of thermal
■
REFERENCES
(1) Akmaz, S.; Gurkaynak, M. A.; Yasar, M. The effect of temperature on the molecular structure of Raman asphaltenes during pyrolysis. J. Anal. Appl. Pyrolysis 2012, 96 (96), 139−145. (2) Ayala, M.; Hernandez-Lopez, E. L.; Vazquez-Duhalt, R.; Perezgasga, L. Reduced coke formation and aromaticity due to chloroperoxidase-catalyzed transformation of asphaltenes from Maya crude oil. Fuel 2012, 92 (1), 245−249. (3) Gonçalves, M. L. A.; Ribeiro, D. A.; Teixeira, A. M. R.; Teixeira, M. A. G. Influence of asphaltenes on coke formation during the thermal cracking of different Brazilian distillation residues. Fuel 2007, 86 (4), 619−623. 5044
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045
Article
Energy & Fuels (4) Li, N.; Yan, B.; Xiao, X. M. Kinetic and reaction pathway of upgrading asphaltene in supercritical water. Chem. Eng. Sci. 2015, 134, 230−237. (5) Savage, P. E.; Klein, M. T.; Kukes, S. G. Asphaltene reaction pathways. 1. Thermolysis. Ind. Eng. Chem. Process Des. Dev. 1985, 24 (4), 1169−1174. (6) Jin, N.; Wang, G.; Han, S.; Meng, Y.; Xu, C.; Gao, J. Hydroconversion Behavior of Asphaltenes under Liquid-Phase Hydrogenation Conditions. Energy Fuels 2016, 30 (4), 2594−2603. (7) Tayakout, M.; Ferreira, C.; Espinat, D.; Arribas Picon, S.; Sorbier, L.; Guillaume, D.; Guibard, I. Diffusion of asphaltene molecules through the pore structure of hydroconversion catalysts. Chem. Eng. Sci. 2010, 65 (5), 1571−1583. (8) Del Bianco, A.; Panariti, N.; Prandini, B.; Beltrame, P.; Carniti, P. Thermal cracking of petroleum residues: 2. Hydrogen-donor solvent addition. Fuel 1993, 72 (1), 81−85. (9) Alemán-Vázquez, L. O.; Cano-Domínguez, J. L.; GarcíaGutiérrez, J. L. Effect of Tetralin, Decalin and Naphthalene as Hydrogen Donors in the Upgrading of Heavy Oils. Procedia Eng. 2012, 42, 532−539. (10) Gray, M. R.; McCaffrey, W. C. Role of Chain Reactions and Olefin Formation in Cracking, Hydroconversion, and Coking of Petroleum and Bitumen Fractions. Energy Fuels 2002, 16 (3), 756− 766. (11) Al-Samarraie, M. F.; Steedman, W. Pyrolysis of petroleum asphaltene in tetralin. Fuel 1985, 64 (7), 941−943. (12) Ignasiak, T. M.; Strausz, O. P. Reaction of Athabasca asphaltene with tetralin. Fuel 1978, 57 (10), 617−621. (13) Rueda-Velásquez, R. I.; Freund, H.; Qian, K.; Olmstead, W. N.; Gray, M. R. Characterization of Asphaltene Building Blocks by Cracking under Favorable Hydrogenation Conditions. Energy Fuels 2013, 27 (4), 1817−1829. (14) Rueda Velásquez, R. I. Characterization of asphaltene molecular structures by cracking under hydrogenation conditions and prediction of the viscosity reduction from visbreaking of heavy oils. Ph.D. Thesis, University of Alberta, Edmonton, Alberta, Canada, 2013. (15) Savage, P. E.; Klein, M. T.; Kukes, S. G. Asphaltene reaction pathways. 3. Effect of reaction environment. Energy Fuels 1988, 2 (5), 619−628. (16) Rahmani, S.; Gray, M. R. Dependence of Molecular Kinetics of Asphaltene Cracking on Chemical Composition. Pet. Sci. Technol. 2007, 25 (1), 141−152. (17) Zhao, Y.; Gray, M. R.; Chung, K. H. Molar Kinetics and Selectivity in Cracking of Athabasca Asphaltenes. Energy Fuels 2001, 15 (3), 751−755. (18) Martinez, M. T.; Benito, A. M.; Callejas, M. A. Thermal cracking of coal residues: Kinetics of asphaltene decomposition. Fuel 1997, 76 (9), 871−877. (19) Yasar, M.; Trauth, D. M.; Klein, M. T. Asphaltene and Resid Pyrolysis. 2. The Effect of Reaction Environment on Pathways and Selectivities. Energy Fuels 2001, 15 (3), 504−509. (20) Wang, J.; Anthony, E. J. A study of thermal-cracking behavior of asphaltenes. Chem. Eng. Sci. 2003, 58 (1), 157−162. (21) Alvarez, E.; Marroquín, G.; Trejo, F.; Centeno, G.; Ancheyta, J.; Díaz, J. A. Pyrolysis kinetics of atmospheric residue and its SARA fractions. Fuel 2011, 90 (12), 3602−3607. (22) Zhao, Y.; Wei, F.; Li, D. Kinetic of the Thermal Cracking, Thermal Hydrocracking and Catalytic Hydrocracking of Asphaltene. Acta Pet. Sin., Pet. Process. Sect. 2011, 27 (5), 753−759. (23) Zhao, Y.; Lin, X.; Li, D. Catalytic Hydrocracking of a BitumenDerived Asphaltene over NiMo/γ-Al2O3 at Various Temperatures. Chem. Eng. Technol. 2015, 38 (2), 297−303. (24) Lababidi, H. M.; Sabti, H. M.; AlHumaidan, F. S. Changes in asphaltenes during thermal cracking of residual oils. Fuel 2014, 117, 59−67. (25) Trauth, D. M.; Yasar, M.; Neurock, M.; Nigam, A.; Klein, M. T.; Kukes, S. G. Asphaltene and Resid Pyrolysis: Effect of Reaction Environment. Fuel Sci. Technol. Int. 1992, 10 (7), 1161−1179.
(26) Nace, D. M.; Voltz, S.; Weekman, V., Jr Application of a kinetic model for catalytic cracking. Effects of charge stocks. Ind. Eng. Chem. Process Des. Dev. 1971, 10 (4), 530−538. (27) Wiehe, I. A. A phase-separation kinetic model for coke formation. Ind. Eng. Chem. Res. 1993, 32 (11), 2447−2454. (28) Trejo, F.; Rana, M. S.; Ancheyta, J. Thermogravimetric determination of coke from asphaltenes, resins and sediments and coking kinetics of heavy crude asphaltenes. Catal. Today 2010, 150 (3−4), 272−278. (29) Carnahan, B.; Luther, H. A.; Wilkes, J. O. Applied Numerical Methods; Wiley: New York, 1969. (30) Houck, C. R.; Joines, J.; Kay, M. G. A Genetic Algorithm for Function Optimization: A MATLAB Implementation; North Carolina State University: Raleigh, NC, 1998; Technical Report NCSU-IE-TR95-09. (31) Trejo, F.; Ancheyta, J. Kinetics of asphaltenes conversion during hydrotreating of Maya crude. Catal. Today 2005, 109 (1), 99−103.
5045
DOI: 10.1021/acs.energyfuels.7b00608 Energy Fuels 2017, 31, 5037−5045