Subscriber access provided by UNIV OF DURHAM
Kinetics, Catalysis, and Reaction Engineering
Two-Step Thermal Cracking of an Extra-Heavy Fuel Oil: Experimental Evaluation, Characterization, and Kinetics Mohammad Ghashghaee, and Samira Shirvani Ind. Eng. Chem. Res., Just Accepted Manuscript • DOI: 10.1021/acs.iecr.8b00819 • Publication Date (Web): 14 May 2018 Downloaded from http://pubs.acs.org on May 16, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Two-Step Thermal Cracking of an Extra-Heavy Fuel Oil: Experimental Evaluation, Characterization, and Kinetics
Mohammad Ghashghaee*, Samira Shirvani
Faculty of Petrochemicals, Iran Polymer and Petrochemical Institute, P.O. Box 14975-112, Tehran, Iran
ACS Paragon Plus Environment
1
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 47
Abstract This work deals with an efficient two-step thermal upgrading process for converting extraheavy fuel oil to light olefins (ethylene, propylene, and butenes) and fuels (gasoline and diesel fuel). In the first step, mild thermal pretreatment was implemented at different temperatures (360–440 °C) in the liquid phase to obtain a more suitable feedstock for an olefin production unit. Thanks to this cost-effective pretreatment, the upgraded feedstock demonstrated considerable flowability and crackability compared to the initial fuel oil, making the subsequent vapor-phase operation easier to handle at temperatures as high as 800 °C with no severe operational impediments. The quantitative 1H and
13
C NMR studies shed light on the enhanced
features of the thermally treated feedstock towards lighter and more valuable products. As a result, remarkable olefins production (74.7 or 55.1 wt% of light olefins based on the upgraded or the original feedstock) was accomplished in this two-step process. The process could be alternatively stopped at the first stage for maximum liquid fuels (69.3 wt%) with gasoline as the larger constituent. The detailed kinetic investigations of the thermal decomposition of the feedstock using several reliable approaches revealed that the activation energy predictions (42.3– 272.9 kJ/mol) by the Kissinger–Akahira–Sunose method almost perfectly matched the trend of a reference Starink model over the whole range of conversion. All model-free methods correlated with a coefficient of determination above 97.9%. The Avrami’s theory was further applied to determine the reaction order, and the values were slightly smaller than those from a five-lump kinetic model of the semi-batch operation. However, the apparent activation barrier in the reactor was in good correspondence with the range from the micro-scale nonisothermal decomposition. Key words: heavy fuel oil; upgrading; thermogravimetric analysis, pyrolysis; olefin; kinetics; activation energy; NMR; gasoline; diesel fuel.
ACS Paragon Plus Environment
2
Page 3 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
1. Introduction Over the years, heavy hydrocarbon cuts have been among the most challenging issues that petroleum refineries are confronted. The conversion of these low-ranked and poor-quality hydrocarbons to light olefins, viz. ethylene and propylene, and fuels with increasing demands in the market is then of an utmost importance.1-4 Nowadays, the bottom-of-the-barrel technologies have become a must in the petroleum industries as the crude sources are becoming heavier and also more contaminated with sulfur and metals. In addition, the production of propylene and butenes through conventional sources cannot meet their growing needs as two versatile organic chemicals.2 Although new on-purpose technologies, such as, propane dehydrogenation (PDH) have been developed for the production of key petrochemicals, namely propylene,4-5 it is quite beneficial and perhaps inevitable to utilize also the low-ranked heavy and extra-heavy hydrocarbon feedstocks for obtaining the desired petrochemicals. By far, several efficient or partially efficient ways have been proposed to tackle this problem. However, all are categorized in two main pathways: thermal and catalytic cracking. The former is frequently applied in petroleum industries by allocating more than 64% of residue processing technologies to itself,6-8 thanks to its obvious convenience, cost-effectiveness, and flexibility to work with.9-10 Although catalytic cracking provides more flexibility with the adjustment of the product slate to some extent, it suffers from immediate deactivation and poisoning of the catalyst due respectively to severe coking and the common presence of impurities in the extra-heavy feedstocks. Therefore, the application of a catalytic cracking process presents a small added value for the one-step upgrading of heavy and even extra-heavy hydrocarbons. The mentioned
ACS Paragon Plus Environment
3
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 47
problems are not seriously relevant to a thermal cracking process. In this sense, the heavier feedstocks can be upgraded conveniently through thermal cracking pretreatments.11-13 Up until now, several studies have been reported on thermal processing of heavy hydrocarbons,14-15 their chemistry,16-17 kinetics,6, 18-20 and economies.21 Depending upon the type of the desirable products of a thermal cracking process, different reaction conditions are employed. It is an established concept that for obtaining significant amounts of olefins, which are the lightest cracked products from every feedstock, the pyrolysis temperature should be high enough, e.g., beyond 700 °C. On the other hand, thermal processes occurring at lower temperatures are in favor of producing more liquid fuels, such as diesel fuel, gasoline, and so forth. Although increasing the reaction temperature in a pyrolytic process would be definitely helpful in increasing the olefins production, operating at such a high temperature range in terms of high viscosity and coking tendency of extra-heavy hydrocarbons is cumbersome. In the present study, we explore the potential of a two-step double-purpose thermal cracking process for the conversion of an extra-heavy fuel oil to light olefins (ethylene, propylene, and butenes) and fuels (gasoline, diesel fuel, and fuel gases) in which, instead of the dilution of the feedstock with a solvent, such as, toluene, gas oil, etc. for a better flowability, the heavy feedstock is treated in two separate steps. The goal of the first step which was accomplished in the liquid phase was to reject the prohibiting compounds and more importantly to enhance the rheological properties of the heavy feedstock. Thanks to this cost-effective improvement, higher yields of the desired products are expected to be achieved through the second step, which has been conducted in the vapor phase under more severe pyrolysis conditions without hindrance by the aforementioned obstacles. In addition to the experiments and structural assessment of the fractions using NMR, and owing to the importance of the kinetic studies,5 a detailed inspection
ACS Paragon Plus Environment
4
Page 5 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
of the thermal decomposition of the extra-heavy fuel oil has been implemented via several wellknown model-free methods, which are known to be more accurate in comparison with the model-fitting approaches to estimate the reaction parameters. At variance with the numerous kinetic studies that commonly assume first-order reactions, we investigated both the reaction order and activation energy from both micro-scale decomposition thermogravimetry and semibatch reaction test data, very scant attention has been paid in the literature to the comparative assessment of the key model-free methods. To the best of our knowledge and despite the apparent simplicity of the ideas, no similar study has appeared previously reporting these considerations.
ACS Paragon Plus Environment
5
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 47
2. Experimental 2.1
Feedstock
The extra-heavy fuel oil applied as a low-quality feedstock in this study was obtained from Tehran Refinery, and from now on is denoted as HFO. In Table 1 are illustrated the main features of this hydrocarbon feedstock. [Table 1] 2.2
Two-Step Pyrolysis
Thermal upgrading of the extra-heavy feedstock was implemented in two consecutive steps. The first stage of processing which is regarded as a pretreatment step was conducted in the liquid phase in a semi-batch reactor placed in an electrical heater equipped with a Ktype temperature sensor. To avoid any possible leakage, all directions were carefully sealed. Before the commencement of the reactions, the system was purged with nitrogen to provide an inert atmosphere. The nitrogen stream was also efficient in sweeping of the products from the reactor chamber. Then the reactor temperature rose to the reaction temperature (360–440 °C) with an approximate rate of 7 °C/min and remained at this temperature for a certain period of time and finally the reactor was quickly cooled down to near-ambient temperature. Nearly 50 g of the feedstock was introduced into the first step. The liquid products were collected in a cold trap placed in an ice bath and the gaseous products were sent to flare after sampling. The obtained liquid product from thermal cracking of HFO at 440 °C, which is called heavy cracked oil (HCO) in the ongoing sections, was applied as a high quality feedstock to the second step. At this stage, the vapor-phase thermal cracking
ACS Paragon Plus Environment
6
Page 7 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
reactions occurred in a tubular reactor located in an electrical furnace. The experimental reaction test apparatus is illustrated in Fig. 1. Five temperature levels from 600 °C to 800 °C were chosen to evaluate the effect of temperature on the performance of the reactions. Nitrogen flow was applied as the carrier and diluent with the flow rate of 80 mL/min with which the flow rate of hydrocarbons was maintained at 0.5 mL/min using a diffusion pump. [Figure 1]
2.3
Feedstock and Product Analysis
The hydrocarbon fractions were characterized with 1H and
13
C NMR spectroscopic
analyses at room temperature using CDCl3 as a solvent. The analysis was performed on a Bruker DRX 500 ANANCE spectrometer. In order to achieve precise proton spectra, an acquisition time of 1.59 s, a sweep width of 10330.58 Hz, a flip angle of 90° (a 5 µs pulse) and a recycle delay of 6 s were fixed on the apparatus. Nearly 128 scans with the broadening line of 0.3 MHz associated with residual CDCl3 resonance at 7.24 ppm were employed. The above-mentioned settings enable the accurate quantitative analysis of protonated species in any petroleum mixture. For the carbon spectra, the acquisition time of 1.04 s, the sweep width of 31446.54 Hz, and the flip angle of 90° were applied. These settings allowed the quantitative analysis of carbons with spin lattice relaxations of about 100 s for HCO and 30 s for the initial HFO feedstock. The nuclear Overhauser for the carbon signals was suppressed by the help of inverse gated decoupling. Total of 3000 scans were employed for the HCO carbon spectra; however, a total number of 7000 scans was collected in the
13
C
NMR analysis of the initial HFO feedstock in which the line broadening of 10 Hz have been employed. The CDCl3 solution was set to 77 ppm for all carbon spectra. The obtained
ACS Paragon Plus Environment
7
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 47
products (liquids and vapors) at each stage were analyzed by gas chromatography (GC) using flame ionization (FID) and thermal conductivity (TCD) detectors. The gasoline and diesel fuel products were defined as the C5–C11 and C12–C22 hydrocarbon ranges, respectively. Thermal decomposition studies (TG/DTG) were performed on a Perkin Elmer Pyris 1 apparatus with about 10 mg of each sample at different ramping rates (5, 10, 20, and 30) in nitrogen atmosphere. The spanning range of temperature was from room temperature to about 800 °C.
ACS Paragon Plus Environment
8
Page 9 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
3. Reaction Kinetics 3.1 Non-Isothermal Decomposition Kinetics Kinetic studies were implemented using TG/DTG spectra of HFO obtained at different temperature programs. Among the two main categories of techniques for analyzing the kinetics of solid state reactions (model-free and model-fitting), the model-free methods were preferred owing to their obvious advantages in estimating the kinetic parameters compared to the modelfitting methods in which the reaction model should be considered beforehand to determine the kinetic parameters.19, 22-24 As it is quite possible that a proposed reaction model fails to reveal the real decomposition events and that a satisfactory agreement between the experimental results and models cannot suffice to vote for a particular model for the reactions, the second technique may bring an undeniable ambiguity to the final kinetic data, thus making them incomparable.25-27 Moreover, model-free methods are more flexible to possible changes in the reaction mechanism, and the application of several heating rates for these types of models enable evaluations as much possibly free from thermal diffusion and mass transfer limitations.28 These altogether provoked us to consider only the model-free methods of Friedman, Flynn– Wall–Ozawa (FWO), Kissinger–Akahira–Sunose (KAS), Kissinger, and the more recent and accurate model of Starink for estimating of the kinetic parameters including the activation energy and pre-exponential factor during the non-isothermal decomposition of the heavy feedstock. In contrast to most of the previous reports in the field, no reaction order was taken for granted. Instead, the variation of the so-called intrinsic reaction order with the decomposition temperature was explored on the basis of the Avrami theory.29-31
ACS Paragon Plus Environment
9
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 10 of 47
Eq 1 describes the general formula for the rate of decomposition:
=
(1)
in which t is the time, T is the temperature, and α is the extent of reaction expressed as eq 3.
=
−
−
(2)
In this equation, mi and mf are the initial and final masses of the target substance, respectively, and mt is the mass of substance at any time of the decomposition test. Three main categories of reaction types are usually considered: accelerating, decelerating and sigmoidal.23 The cracking reactions of heavy oil are more likely to fall into the decelerating category, which illustrates its maximum rate at the beginning of the reaction
32
. In this type of reaction model, the rate of
decomposition decreases as long as the extent of reaction increases. The most common conversion function applied to this type, which is widely applied also to solid state reactions, can be written as follows:
= 1 −
(3)
As the rate constants generally comply with the Arrhenius Equation, one may replace k(T) in eq 1 with the following expression to derive eq 5 in which A is the pre-exponential factor and Ea is the activation energy.
= exp
−
(4)
ACS Paragon Plus Environment
10
Page 11 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
− = exp 1 −
(5)
As thermal analysis experiments in the present study were conducted under nonisothermal conditions with different heating rates, the temperature is defined at any time as follows:
= +
(6)
Substitution of eq 6 into eq 5 gives the final equation as described below:
− = exp 1 −
(7)
Upon integration and applying a variety of temperature dependent approximations, the following relationships were derived, which have been assigned to the Friedman, FWO, KAS, Kissinger, and Starink methods (please see the literature23, 28, 33 for further details).
ln = ln − ln = ln
(8)
− 5.331 − 1.052
(9)
ln = ln −
(10)
ln = ln[ ] −
(11)
ACS Paragon Plus Environment
11
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ln ).* = ln + 3.7545411 − 1.92 ln − 1.0008
Page 12 of 47
(12)
The Avrami’s theory reveals the variation of the degree of conversion with temperature and heating rate as follows:
− = 1 − exp / 0
(13)
Taking double natural logarithm of both sides and replacing k(T) with Aexp[–E/(RT)] yields the following equation:
ln1− ln11 − 22 = ln −
− 3 ln
(14)
Hence, the plots of ln(–ln(1–α)) vs. lnβ at isothermal conditions, give the order of reaction as the slop of regression lines. 3.2 Semi-Batch Reactor Kinetics According to an appropriate procedure developed previously,34 the residual fraction was envisaged as an ensemble of two parts:
= 4 + ∞
(15)
in which R signifies the weight fraction of the unconverted (residual) phase inside the reactor with respect to the initial amount of the heavy fuel oil, 4 refers to the convertible portion of the
ACS Paragon Plus Environment
12
Page 13 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
residual solid remaining at a specified reaction time, and R∞ represent the asymptotic (final) amount or the unconvertible part of the fuel oil. [Figure 2] A five-lump kinetic model shown in Fig. 2 was employed, which encompassed four pathways with different reaction orders towards gases, including olefins and paraffins, and liquid products, including gasoline and diesel fuel, on the basis of the available portion of the feedstock 1 − 4 . Correspondingly, the consumption of the residual fraction is modeled as follows:
= 5 + 4 exp−67
(16)
where t represents the reaction time and the subscript 0 refers to the initial state. After the content of R∞ is estimated, the following set of rate equations are to be solved:
8/ = 6: 4 ;
(17)
(18)
?/ = 6@ 4 A
(19)
B/ = 6C 4 D
(20)
where K stands for the equilibrium constant, O represents olefins, P denotes paraffins, N represents gasoline or naphtha, and D is diesel fuel; the superscripts p, q, r, and s show the
ACS Paragon Plus Environment
13
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 14 of 47
corresponding reaction orders. These constants are calculated using an iterative numerical optimization to minimize the error sum of squares for the model predictions.
ACS Paragon Plus Environment
14
Page 15 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
4. Results and Discussion Table 2 shows the liquid fractions obtained in the two steps. As expected, the first step of the process was more likely to produce fuels than olefins due to the mild cracking conditions. More precisely, the total liquid fuels from the pretreatment of the HFO feedstock amounted to 69.3 wt%, which included gasoline with the yield of 38.2 wt% and diesel fuel with the 31.1 wt% yield. At the same time, the yield of light olefins at this stage was only about 5 wt%, indicating that the virgin extra-heavy fuel oil was not an appropriate feedstock for olefins production. Instead, the first moderate-temperature thermal upgrading process produces maximum amounts of liquid fuels. As such, the overall conversion obtainable from the extra-heavy oil in the first step of the proposed upgrading process was about 81% including also the fuel gases. [Table 2] The graphs of instantaneous and accumulative gas yields related to the first step of the reaction (thermal pretreatment step) at the lowest and highest temperatures with respect to the reaction time are presented in Fig. S1. The accumulative gas yield after 205 min on stream was above 12.5 wt% at both temperatures. As it is evident, the maximum instantaneous gas yield of 8.3 wt% was obtained at 55 min time-on-stream for the temperature of 440 °C while it was 4.7 wt% after 108 min time-on-stream at the lowest temperature, which is quite sensible due to the fact that lighter compounds are obtained more abundantly at higher temperatures. After that, the production of light olefins declined rapidly and eventually reached almost zero at the 200 min time-on-stream. The instantaneous yields of light olefins in the first step at two temperature levels of 360 and 440 °C are shown in Fig. S2. As evident, for both cases, the olefin production started
ACS Paragon Plus Environment
15
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 16 of 47
immediately, then fell down to minimum at the time-on-stream of about 80 min and then again increased steadily up to the end of run. More precisely, it almost occupied the major part of the gas composition at approximately the end of the reaction time. This implied the high potential of this extra-heavy feedstock in producing light olefins. The quantitative NMR analysis of the carbon and proton spectra is presented in Table 3 for the extra-heavy fuel oil (HFO) and the improved cracked oil (HCO). In the same line, Fig. 3 depicts the assessment of paraffinic, olefinic, naphthenic, and aromatic (PONA) hydrocarbons according to the approach presented previously.34 As can be seen, the thermal pretreatment led to increased naphthenes and paraffins while reducing aromatics and olefins. This translates into the higher crackability and quality of the cracked oil for olefins production compared to the initial fuel oil. More precisely, HCO had both lower contents of protonated aromatic carbon and quaternary carbon compared to HFO, serving as a principal explanation for the higher cracking reactivity of the former. Moreover, considerable amounts of paraffins in both of the fractions were evolved as α-to-cycloparaffinic methyl groups whereas lower methyl groups were bound to the aromatic rings. Taking into account the abundant CH2 carbon types and the slightly higher proportion of long-chain paraffins in HCO, one can contemplate that longer chains of normal paraffins were present in the upgraded feedstock. Applying the ratio of long chain paraffinic CH2 and the terminal methyl groups of paraffinic chains as an indicator of the chain length35-36 (2.95 and 2.22 for HFO and HCO, respectively), however, one may note that the paraffinic chains in the pretreated fraction are slightly smaller than the original paraffins due mainly to the liquid-phase cracking reactions. At the same time, the paraffinic CH2 carbon was more pronounced in the upgraded feedstock, indicating the higher proportion of linear paraffins in HCO. These altogether explain the higher crackability and lower viscosity of the upgraded feedstock. The same
ACS Paragon Plus Environment
16
Page 17 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
alteration led to the increase in the H/C ratio while the sulfur content of HCO was slightly higher than that of HFO (Table 3). [Table 3] [Figure 3] Interestingly, the upgraded feedstock presented higher quantities of iso-paraffinic methyl groups that made it prone to propylene production. On the other hand, the amounts of paraffinic cycles in the two feedstocks were almost the same, implying that the cyclic paraffins underwent some alkylation during the thermal pretreatment. On the other hand, HCO contained a larger proportion of paraffinic CH carbon than HFO as the main indicator of the higher naphthenic content of the former feedstock. The presented structural changes also indicated that the original fuel oil contained a plethora of polyaromatics that remained in the residual phase, thus making the obtained cracked oil intact from the unwanted coke precursors. However, considerable amounts of mononuclear aromatics remained in the upgraded fuel oil (Fig. 3). Also shown in Fig. 3 is that the olefinic content of HCO was lower than that of HFO. In total, this indicates that the thermal pretreatment on the original feedstock did not bring about new olefinic (double or triple) bonds in the liquid phase that can escape from the reactor to the upgraded fuel oil. It should be noted, however, that the NMR data represent the atomic contribution rather than the molecular concentration and the latter is larger in case of olefins 34. Instead, the olefins produced as a consequence of the β-scission reactions evolved in the gases coming out of the reactor as pointed out previously. The liquid product of the pretreatment step (440 °C) was employed in the second step as a more convenient feedstock for olefin units. In the second step, an elevated range of temperatures could be applied to the thermal cracking to achieve olefin yields as high as possible in the final
ACS Paragon Plus Environment
17
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 18 of 47
products. Notably, the maximum yield of 74.7 wt% of olefins based on the HCO feedstock (55.1 wt% based on the initial HFO) was obtained from thermal cracking of HCO at 800 °C. This is an outstanding amount when compared to the values reported in the literature even with lighter feedstocks or even in the catalytic cracking routes.16,
34, 37-41
At the same time, however, the
methane-to-olefins (M/O) selectivity factor42 was maximized at 800 °C, indicating that the cracking severity was also maximized at this temperature. The remarkably high olefin yields obtained here indicated that although the feedstock at hand was an extra-heavy hydrocarbon fraction, it contained amenable components and structures in its composition that could be cracked to astonishing yields of light olefins at temperatures common in industrial olefin plants using a proper two-step process. Meanwhile, the lowest amount of liquid fuel was obtained at the highest temperature (Table 2) where the liquid fraction was only composed of gasoline (14.0 wt%). However, the liquid yield from HCO was highest at the lowest temperature with the contributions of 63.6 wt% of gasoline and 25.7 wt% of diesel fuel (Table 2). This indicates that the second step produces gasoline as the predominant liquid fuel, particularly at higher temperatures. The detailed composition of the olefins from HCO in the second step of the upgrading procedure is shown in Fig. 4. As obvious, by increasing the temperature of the cracking reactions, the total olefin yields increased as well. This increasing trend was kept until the reaction temperature of 800 °C. However, the steepest changes occurred at lower temperatures. More strictly, the changes indicated the substantial increase of light olefins from 10.1 wt% to 65.3 wt% on a change of operating temperature from 600 to 700 °C, respectively. The detailed composition of the produced olefins followed the order of butenes < ethylene < propylene at
ACS Paragon Plus Environment
18
Page 19 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
almost all of the temperatures except 700 °C at which the butenes yield surpassed that of ethylene and also 800 °C at which the ethylene yield was larger than that of propylene. [Figure 4] As it is clear, the production of ethylene increased constantly by increasing the reaction temperature in such a manner that it reached its maximum of 31.9 wt% at 800 °C. Meanwhile, the yield of propylene passed through its peak of 32.2 wt% at 750 °C which was followed by a slight decrease at higher temperatures (see Fig. 4). This temperature level is then regarded as an optimum temperature with respect to propylene production. However, the highest propylene-toethylene ratio (P/E severity factor)7 of 2.24 was obtained at 650 °C. The obtained olefin yields were quite higher than the reported values in the literature normally accomplished in a conventional thermal cracking process. The proposed two-step process then permits the thermal cracking to conveniently proceed to 800 °C without any significant coke formation or clogging of the tubes, which is a major problem with the thermal conversion of the extra heavy feedstocks in olefin plants. The other notable impact of the aforementioned cost-effective pretreatment step is that it makes the extra-heavy hydrocarbon feed easily flowable and facile to work with. The DTG curves of the extra-heavy fuel oil under nonisothermal conditions at different heating rates are shown in Fig. 5. The results revealed that by enhancing the heating rate from 5 to 30 °C/min, the value of Tmax for the decomposition of hydrocarbons shifted to higher values. This observation is confirmed by the reported findings in the previous works on thermal analysis.26 Besides, increasing the heating rate has led to strengthen the intensity of the small valley around 410 °C such that at the rate of 30 °C/min, it became the main valley of the degradation profile (see Fig. 5). The other notable impact of amplifying the heating rate is that the intensity of the
ACS Paragon Plus Environment
19
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 20 of 47
valley at about 670 °C decreased significantly. To provide an explanation to these observations, one may note that the heating rate has indispensible affects on both the decomposition temperature and the mechanism of thermal cracking reaction.26 [Figure 5] Table S1 reports the variation of the reaction order with temperature accomplished by adopting Avrami’s theory. As evident, the so-called intrinsic order of reaction ranged from 0 to 0.48. The Avrami’s theory fitted fairly well to the experimental data. Further, five model-free approaches including Friedman, FWO, KAS, Kissinger, and Starink (eqs 8 through 12) were employed to analyze the decomposition kinetics of HFO. The results of the analysis are shown in Fig. 6. To implement the kinetic identification, the intercept and slope of each straight line in this figure along with the aforementioned model equations were applied. The range of the obtained parameters at various extents of reaction (α=0.05–0.95) have been reported in Table 4. [Figure 6] [Table 4] As evident, the predictions from all models were in the good agreement with the experimental data, being certified by their high correlation coefficients, higher than 97.9% in all cases. Hence, from a goodness-of-fit point of view, all models were generally accepted for the degradation of this extra-heavy feedstock. As the reactions proceeded, the apparent activation energies varied by the alteration in the extent of decomposition. This observation alludes to the presence of a complex mechanism for the reactions. It can be observed from Table 4 that all models propose a wide range of activation energies from ca. 42 to 273 kJ/mol for the thermal cracking of the HFO
ACS Paragon Plus Environment
20
Page 21 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
feedstock. As long as the degradation reactions progressed, the activation energy increased according to the four models of Friedman, FWO, KAS, and Starink (see Fig. 7a). As evident, for all models a gradual increase in the activation energy was observed until α=0.6. After that, the activation energy was increased significantly to its maximum at α=0.95. This increase was more severe in the case of Friedman. This observation underlined the difficulty of the final (10–15% conversion) stages in the thermal decomposition of the extra-heavy fuel oil at hand, which is attributable to the aromatic rings. The activation energies predicted by KAS were approximately similar to those obtained by the basically more dependable Starink model,23, 28, 33 which implies that these two models worked with comparable precisions for the present system (see Fig. 7a). Although the estimated activation energy by FWO was close to those proposed by KAS and Starink at the final stages of the reaction (high conversion levels), it obviously deviated by ~7 kJ/mol from Starink at the first half of the temperature-programmed process (up to the α = 0.55) (see Fig. 7b). On the other hand, the Friedman approach predicted quite higher values of activation energy for the thermal decomposition of HFO at the second half of the process. At the same time, Kissinger predicted single activation energy of 54.6 kJ/mol for the whole decomposition range with the highest correlation coefficient of 99.7%. The predicted stepwise augmentation of the activation energy points to the presence of complex molecular structures in HFO, some of which could be easily degraded while others faced higher energy barriers to convert. Overall, the models shared the increasing trends of activation energy with comparable overall ranges. [Figure 7] The Arrhenius equation was applied to undertake kinetic studies of the thermal cracking reactions in the semi-batch reactor (see Fig. S3). The analysis predicted the kinetic parameters
ACS Paragon Plus Environment
21
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 22 of 47
with the correlation coefficient of higher than 98%. The obtained apparent activation energy for the conversion levels practiced in the reactor (64.68 kJ/mol) was quite rational and comparable with those obtained from the thermogravimetric modeling. The observed differences should be connected with the presence of transport limitations inside the reactor compared to the microscale thermogravimetric experiments and the resulting gradients in the macro-scale environment. [Table 5] [Figure 8] To obtain more insights into the (apparent) kinetics of the reactor operation, eqs 16–20 were employed for the kinetic modeling of the semi-batch reactor. The differential equations were applied to fit the yield profiles at 440 °C with time. The content of 5 at the investigated conditions was obtained as 17.03 wt%. The prediction curves demonstrated in Fig. 8 show excellent fits to the experimental data with the estimated parameters given in Table 5. The individual reaction rate constants varied from 0.33×10–3 to 18.00×10–3 min–1 with the largest magnitude estimated for the production of diesel fuel. As also evident, the reaction order with respect to the conversion pathways varied in the range of 0.81–1.34, being slightly larger than those from the thermogravimetric data (vide supra). This indicates that the reaction order in the real practice can be larger than those of micro-scale thermal analysis owing to the importance of the transport phenomena within the reactor.
5. Conclusions
ACS Paragon Plus Environment
22
Page 23 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
This article addressed the efficient two-step upgrading of an extra-heavy fuel oil in which the feedstock was first thermally treated to enhance its quality and flowability and then applied in a second vapor-phase thermal cracking process. The pretreatment step, as a cost-effective and engineer-friendly solution, was applied successfully to one of the most important challenges of petro-refining industries, i.e., the conversion of bottom-of-the-barrel components particularly those with prohibitive properties such as high viscosity and high sulfur and metal contamination. As found from the carbon and proton NMR analyses, the thermal pretreatment led to increased naphthenes and paraffins while reducing aromatics and olefins. These altogether explained the higher crackability of the upgraded fuel oil in which most of the methyl groups were bound to the cyclic paraffins rather than aromatic rings. However, the HCO contained more branched hydrocarbon species and smaller chains of normal paraffins, which are particularly suitable for propylene production. The reaction results illustrated more than 81% conversion of the extraheavy fuel oil and also more than 55% overall yield of olefins with propylene as interestingly the main olefinic compound. Whereas the highest propylene production appeared at 750 °C, the largest P/E severity factor was attributed to the temperature of 650 °C. However, the highest propylene-to-ethylene ratio (P/E severity factor) of 2.24 was obtained at 650 °C. In an alternative scenario, however, the process could be stopped at the first stage for the maximum liquid fuels of 69.3 wt% with gasoline as the larger part. The reaction order (0–0.48) was determined from the nonisothermal decomposition of the extra-heavy feedstock at 573–873 K using the Avrami’s theory. All types of model-free methods applied to evaluate the kinetic parameters, showed the correlation coefficient of higher than 97% along with a relatively wide range of activation barriers (42.3–272.9 kJ/mol) over the whole range of the extent of reaction in the decomposition of the extra-heavy fuel oil. However, the KAS model performed almost equally well compared to
ACS Paragon Plus Environment
23
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 24 of 47
the reference Starink model. Meanwhile, the FWO and Friedman methods most deviated from Starink at low and high conversion levels, respectively. A five-lump kinetic scheme with different reaction orders in each route could very successfully describe the apparent kinetics inside the reactor. The thus obtained reaction order and activation energy differed slightly from those by the micro-scale thermal analysis, due mainly to the interference of transport limitations inside the reactor.
AUTHOR INFORMATION Corresponding Author * Corresponding author. Tel.: +98 21 48662481; fax: +98 21 44787032. E-mail address:
[email protected]
Notes The authors declare no competing financial interest. Acknowledgments Technical assistance by Ms. Mahboobeh Balar is gratefully acknowledged as are the partial support received from Iran National Science Foundation (INSF) under Grant No. 94016123 and Iran Polymer and Petrochemical Institute under Grant No. 53793101.
ACS Paragon Plus Environment
24
Page 25 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Supporting Information. Further data for the performance of the reactions in the first step of the process supplied as Supporting Information.
Nomenclature Symbols A = pre-exponential factor [s–1 or the same as k] D = diesel fuel [wt%] Ea = apparent activation energy [kJ/mol] f(α) = conversion function [–] G = cracked gas [wt%] K = equilibrium reaction rate constant [min–1] k = reaction constant [depending upon the order of reaction] mf = final mass [g] mi = initial mass [g] mt = mass at any time [g] N = gasoline or naphtha [wt%] n = overall order of reaction [–]
ACS Paragon Plus Environment
25
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 26 of 47
O = light olefins [wt%] P = paraffins [wt%] p…s = reaction order R = fuel oil reaction residue [wt%] or gas constant [kJ/(mol.K)] R2 = coefficient of multiple determination [–] R∞ = unconvertible fraction of the feedstock [wt%] 4 = convertible portion of the residual phase [wt%] T = temperature [°C or K] t = time [s, min, or h] T0 = initial temperature [K] Tm = peak temperature [K] α = extent of reaction [–] β = heating rate [°C/min or K/min] Abbreviations DTG = derivative thermogravimetric FID = flame ionization detector FWO = Flynn–Wall–Ozawa
ACS Paragon Plus Environment
26
Page 27 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
GC = gas chromatography HCO = heavy cracked oil HFO = heavy fuel oil KAS = Kissinger–Akahira–Sunose NMR = nuclear magnetic resonance TCD = thermal conductivity detector TG = thermogravimetry Subscripts 0
initial
D = diesel fuel N = naphtha O = olefins P = paraffins R = fuel oil residue
References
ACS Paragon Plus Environment
27
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(1)
Page 28 of 47
Schmidt, L. D.; Siddall, J.; Bearden, M. New Ways to Make Old Chemicals. AIChE J.
2000, 46, 1492–1495. (2)
Ghashghaee, M. Heterogeneous catalysts for gas-phase conversion of ethylene to higher
olefins. Rev. Chem. Eng. 2017, DOI: 10.1515/revce-2017-0003. (3)
Ghashghaee, M.; Karimzadeh, R. Applicability of protolytic mechanism to steady-state
heterogeneous dehydrogenation of ethane revisited. Micropor. Mesopor. Mat. 2013, 170, 318– 330. (4)
Barghi, B.; Fattahi, M.; Khorasheh, F. The Modeling of Kinetics and Catalyst
Deactivation in Propane Dehydrogenation Over Pt-Sn/γ-Al2O3 in Presence of Water as an Oxygenated Additive. Pet. Sci. Technol. 2014, 32, 1139–1149. (5)
Fattahi, M.; Kazemeini, M.; Khorasheh, F.; Rashidi, A. Kinetic modeling of oxidative
dehydrogenation of propane (ODHP) over a vanadium–graphene catalyst: Application of the DOE and ANN methodologies. J. Ind. Eng. Chem. 2014, 20, 2236–2247. (6)
Ghashghaee, M. Predictive correlations for thermal upgrading of petroleum residues. J.
Anal. Appl. Pyrol. 2015, 115, 326–336. (7)
Ghashghaee, M.; Karimzadeh, R. Multivariable optimization of thermal cracking
severity. Chem. Eng. Res. Des. 2011, 89, 1067–1077. (8)
Ghashghaee, M.; Karimzadeh, R. Dynamic Modeling and Simulation of Steam Cracking
Furnaces. Chem. Eng. Technol. 2007, 30, 835–843. (9)
Singh, J.; Kumar, S.; Garg, M. O. Kinetic modelling of thermal cracking of petroleum
residues: A critique. Fuel Process. Technol. 2012, 94, 131–144. (10)
Speight, J. G. Visbreaking: A technology of the past and the future. Sci. Iran. 2012, 19,
569–573. (11)
Speight, J. G. The Refinery of the Future. William Andrew: 2010.
(12)
Speight, J. G. Heavy and Extra-Heavy Oil Upgrading Technologies. Gulf Professional
Publishing: 2013. (13)
Lee, S.; Speight, J. G.; Loyalka, S. K. Handbook of Alternative Fuel Technologies.
Taylor & Francis Group: Boca Raton, FL, 2007. (14)
Hamidi Zirasefi, M.; Khorasheh, F.; Ivakpour, J.; Mohammadzadeh, A. Improvement of
the thermal cracking product quality of heavy vacuum residue using solvent deasphalting pretreatment. Energy Fuels 2016, 30, 10322–10329.
ACS Paragon Plus Environment
28
Page 29 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
(15)
Urán Castaño, L. C. Coke formation during thermal cracking of a heavy crude oil;
Universidad Nacional de Colombia-Sede Medellín: Medellin, Colombia, 2015. (16)
León, A. Y.; Díaz, S. D.; Rodríguez, R. C.; Laverde, D. In Reactivity study on thermal
cracking of vacuum residues, J. Phys.: Conf. Ser., 2016; IOP Publishing: 2016; p 012094. (17)
Stratiev, D.; Shishkova, I.; Dinkov, R.; Nikolova, R.; Mitkova, M.; Stanulov, K.; Sharpe,
R.; Russell, C. A.; Obryvalina, A.; Telyashev, R. Reactivity and stability of vacuum residual oils in their thermal conversion. Fuel 2014, 123, 133–142. (18)
Montoya, T.; Argel, B. L.; Nassar, N. N.; Franco, C. A.; Cortés, F. B. Kinetics and
mechanisms of the catalytic thermal cracking of asphaltenes adsorbed on supported nanoparticles. Pet. Sci. 2016, 13, 561–571. (19)
Heydari, M.; Rahman, M.; Gupta, R. Kinetic Study and Thermal Decomposition
Behavior of Lignite Coal. Int. J. Chem. Eng. 2015, 2015, 9. (20)
Díaz, F. A.; Chaves, A.; Maradei, M. P.; Fuentes, D.; Guzmán, A.; Picón, H. Kinetic
analysis of the thermal decomposition of Colombian vacuum residua by thermogravimetry. Ing. Invest. 2015, 35, 19–26. (21)
Demirbas, A.; Bafail, A.; Nizami, A.-S. Heavy oil upgrading: Unlocking the future fuel
supply. Petrol. Sci. Technol. 2016, 34, 303–308. (22)
Sewry, J. D.; Brown, M. E. “Model-free” kinetic analysis? Thermochim. Acta 2002, 390,
217–225. (23)
Vyazovkin, S.; Burnham, A. K.; Criado, J. M.; Pérez-Maqueda, L. A.; Popescu, C.;
Sbirrazzuoli, N. ICTAC Kinetics Committee recommendations for performing kinetic computations on thermal analysis data. Thermochim. Acta 2011, 520, 1–19. (24)
Vyazovkin, S. A time to search: finding the meaning of variable activation energy. Phys.
Chem. Chem. Phys. 2016, 18, 18643–18656. (25)
Shirvani, S.; Ghashghaee, M. Mechanism Discrimination for Bimolecular Reactions:
Revisited with a Practical Hydrogenation Case Study. Phys. Chem. Res. 2017, 5, 727–736. (26)
Sokoto, M. A.; Singh, R.; Krishna, B. B.; Kumar, J.; Bhaskar, T. Non-isothermal kinetic
study of de-oiled seeds cake of African star apple (Chrosophyllum albidum) using thermogravimetry. Heliyon 2016, 2, e00172. (27)
Vyazovkin, S.; Wight, C. A. Isothermal and non-isothermal kinetics of thermally
stimulated reactions of solids. Int. Rev. Phys. Chem. 1998, 17, 407–433.
ACS Paragon Plus Environment
29
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
(28)
Page 30 of 47
Guida, M. Y.; Bouaik, H.; El, M. L.; Moubarik, A.; Aboulkas, A.; El, h. K.; Hannioui, A.
Utilization of Starink Approach and Avrami Theory to Evaluate the Kinetic Parameters of the Pyrolysis of Olive Mill Solid Waste and Olive Mill Wastewater. J. Adv. Chem. Eng. 2017, 7. (29)
Li, H.; Niu, S.; Lu, C.; Wang, Y. Comprehensive Investigation of the Thermal
Degradation Characteristics of Biodiesel and Its Feedstock Oil through TGA–FTIR. Energy Fuels 2015, 29, 5145–5153. (30)
Otero, M.; Gómez, X.; García, A. I.; Morán, A. Non-isothermal thermogravimetric
analysis of the combustion of two different carbonaceous materials. J. Therm. Anal. Calorim. 2008, 93, 619–626. (31)
Niu, S.; Liu, M.; Lu, C.; Li, H.; Huo, M. Thermogravimetric analysis of carbide slag. J.
Therm. Anal. Calorim. 2014, 115, 73–79. (32)
Guo, A.; Zhang, X.; Wang, Z. Simulated delayed coking characteristics of petroleum
residues and fractions by thermogravimetry. Fuel Process. Technol. 2008, 89, 643–650. (33)
Starink, M. J. The determination of activation energy from linear heating rate
experiments: a comparison of the accuracy of isoconversion methods. Thermochim. Acta 2003, 404, 163–176. (34)
Asgharzadeh Shishavan, R.; Ghashghaee, M.; Karimzadeh, R. Investigation of kinetics
and cracked oil structural changes in thermal cracking of Iranian vacuum residues. Fuel Process. Technol. 2011, 92, 2226–2234. (35)
Japanwala, S.; Chung, K. H.; Dettman, H. D.; Gray, M. R. Quality of Distillates from
Repeated Recycle of Residue. Energ. Fuel. 2002, 16, 477–484. (36)
Netzel, D. A.; McKay, D. R.; Heppner, R. A.; Guffey, F. D.; Cooke, S. D.; Varie, D. L.;
Linn, D. E. 1H- and 13C-n.m.r. studies on naphtha and light distillate saturate hydrocarbon fractions obtained from in-situ shale oil. Fuel 1981, 60, 307–320. (37)
Jafari Fesharaki, M.; Ghashghaee, M.; Karimzadeh, R. Comparison of four nanoporous
catalysts in thermocatalytic upgrading of vacuum residue. J. Anal. Appl. Pyrolysis 2013, 102, 97–102. (38)
Doronin, V. P.; Lipin, P. V.; Sorokina, T. P. Effect of process conditions on the
composition of products in the conventional and deep catalytic cracking of oil fractions. Catal. Ind. 2012, 4, 100–104.
ACS Paragon Plus Environment
30
Page 31 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
(39)
Zhao, L.; Gao, J.; Xu, C.; Shen, B. Alkali-treatment of ZSM-5 zeolites with different
SiO2/Al2O3 ratios and light olefin production by heavy oil cracking. Fuel Process. Technol. 2011, 92, 414–420. (40)
Corma, A.; Diaz-Cabanas, M. J.; Martínez-Triguero, J.; Rey, F.; Rius, J. A large-cavity
zeolite with wide pore windows and potential as an oil refining catalyst. Nature 2002, 418, 514– 517. (41)
Hauser, A.; AlHumaidan, F.; Al-Rabiah, H.; Halabi, M. A. Study on Thermal Cracking of
Kuwaiti Heavy Oil (Vacuum Residue) and Its SARA Fractions by NMR Spectroscopy. Energy Fuels 2014, 28, 4321–4332. (42)
Karimzadeh, R.; Ghashghaee, M.; Nouri, M. Effect of Solvent Dearomatization and
Operating Conditions in Steam Pyrolysis of a Heavy Feedstock. Energ. Fuel. 2010, 24, 1899– 1907.
ACS Paragon Plus Environment
31
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 32 of 47
Figure Captions Figure 1. Schematic diagram of the experimental setup for the liquid-phase (a) and vapor-phase (b) pyrolysis. Figure 2. Five-lump kinetic model for the semi-batch reactor. Figure 3. Hydrocarbon type (PONA) analysis for the HFO and HCO fractions. Figure 4. Yields of light olefins obtained from HCO at different temperatures. Figure 5. DTG graphs of the extra-heavy hydrocarbon feedstock at different heating rates. Figure 6. Kinetic plots of the decomposition of the extra-heavy fuel oil at different conversion levels using the (a) Friedman, (b) FWO, (c) KAS, and (d) Starink models. Figure 7. Variation of the activation energy with the extent of reaction for the Friedman, FWO, KAS, and Starink methods (a) and the enlarged view at the lower half of the conversion range (b). Figure 8. Prediction capability of the proposed five-lump kinetic model.
ACS Paragon Plus Environment
32
Page 33 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
For Table of Contents
ACS Paragon Plus Environment
33
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 34 of 47
Table 1. Properties of the extra-heavy feedstock. Characteristics
Value
Density at 15 °C, g/cm3
0.97
API gravity
14.05
Pour point, °C
32
Flash point, °C
65
CCR, mass%
17
Sulfur content, mass%
1.36
Ash, mass%
0.15
ACS Paragon Plus Environment
34
Page 35 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Table 2. Liquid fuels obtained in the upgrading process. Liquid fuel, wt%
1st step
2nd step (600 °C)
2nd step (800 °C)
Gasoline fraction
38.2
63.6
14.0
Diesel fraction
31.1
25.7
0
Total liquid fuels
69.3
89.3
14.0
ACS Paragon Plus Environment
35
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 36 of 47
Table 3. Quantitative results obtained from NMR analysis of HFO and HCO and their general specifications. HFO
HCO
Aromatic CH
11.6497
9.3942
Methyl of an aromatic ethyl
1.7854
2.0450
Terminal methyl of a paraffinic chain
4.4635
6.3906
Quaternary carbon of aromatics
19.8178
12.3339
Paraffinic CH
9.0832
10.8960
Long chain paraffinic CH2
13.1449
14.1552
Cycloparaffinic CH2
6.6506
6.5823
Oxygenated
1.6068
0.1917
Aromatic-attached ethyl CH2
1.7854
2.0450
α-to-aromatic naphthenes
3.3253
3.2912
Paraffinic CH2
6.6283
8.7871
Chain β-CH2
3.1244
4.1539
Chain γ-CH2 or β-to-aromatic CH2
4.4188
6.5823
Branched-chain CH3
0.1339
0.4473
α-to-cycloparaffinic methyl
8.5577
8.3025
α-to-aromatic ring methyl
1.5298
2.8172
Density, g/cm3
0.9701
0.8713
Viscosity at 37.8 °C, cSt
589.69
6.05
Sulfur content, mass%
1.36
1.83
H/C (atomic ratio)
1.51
1.69
Compositional data, wt%
General specifications
ACS Paragon Plus Environment
36
Page 37 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Table 4. Estimated kinetic parameters for the degradation of the extra-heavy fuel oil using four selected model-free methods.a Model
Ab
E
R2
Friedman
9.85E+02—6.34E+17
48.2–272.9
97.9
FWO
2.20E+04—6.23E+12
47.4–265.7
98.7
KAS
1.48E+03—4.06E+12
42.3–262.8
98.3
Kissinger
2.3E+04
54.64
99.7
Starink
1.84E+03—4.6E+12
42.6–263.3
98.3
a
Range of different parameters were attributed to α=0.05 to 0.95.
b
Assuming n=0.48 as the reaction order.
ACS Paragon Plus Environment
37
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 38 of 47
Table 5. Reaction rate constants and reaction orders obtained from the five-lump kinetic modeling of the reactor.a Constant
Value
Constant
Value
R∞
17.03
KD
18.00E–3
KR
25.90E–3
p
1.31
KO
0.33E–3
q
1.34
KP
0.82E–3
r
1.07
KN
9.12E–3
s
0.81
a
R∞ is in wt% and the reaction rate constants are in min–1.
ACS Paragon Plus Environment
38
Page 39 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Figure 1 (a)
(b)
ACS Paragon Plus Environment
39
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 40 of 47
Figure 2
ACS Paragon Plus Environment
40
Page 41 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Figure 3
ACS Paragon Plus Environment
41
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 42 of 47
Figure 4
ACS Paragon Plus Environment
42
Page 43 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Figure 5
ACS Paragon Plus Environment
43
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 44 of 47
Figure 6
ACS Paragon Plus Environment
44
Page 45 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Figure 6 (Contid.)
ACS Paragon Plus Environment
45
Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 46 of 47
Figure 7 (a)
(b)
ACS Paragon Plus Environment
46
Page 47 of 47 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Industrial & Engineering Chemistry Research
Figure 8
ACS Paragon Plus Environment
47
