Removal of Refractive Sulfur and Aromatic Compounds from Straight

May 29, 2015 - Sayles , S.; Bailor , J.; Ohmes , R. ULSD problems and solutions PTQ Autumn 2004, 107– 115. There is no corresponding record for this...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/EF

Removal of Refractive Sulfur and Aromatic Compounds from Straight-Run, Fluidized Catalytic Cracking, and Coker Gas Oil Using N‑Methyl-2-pyrrolidone in Batch and Packed-Bed Extractors Sunil Kumar,†,‡ Vimal Chandra Srivastava,*,† Rohit Raghuvanshi,‡ Shrikant Madhusudan Nanoti,‡ and Nisha Sudhir‡ †

Department of Chemical Engineering, Indian Institute of Technology Roorkee, Roorkee, Uttarakhand 247667, India Indian Institute of Petroleum, Dehradun, Uttarakhand 248005, India

‡

ABSTRACT: In the present study, removal of sulfur and aromatic compounds from straight-run gas oil (SRGO), light cycle oil (LCO), coker gas oil (CGO), and their mixture, termed as mixed gas oil (MGO), was studied using N-methyl-2-pyrrolidone (NMP) solvent in a single-stage batch extractor and continuous countercurrent packed-bed extraction column. The effect of the extraction temperature (TE), solvent/feed ratio (S/F), and antisolvent concentration (Wc) on the degree of sulfur removal (Dsr), yield of extracted gas oil (Y%), and performance factor (Pf,α), which combines both Dsr and yield, was studied in a single-stage batch extractor. After optimization of the operating conditions for SRGO, LCO, and CGO in a single-stage batch extractor, studies on MGO were carried out in both a single-stage batch extractor and continuous countercurrent packed bed at estimated optimized values of TE, S/F, and Wc. Issues related to the loss of valuable hydrocarbons with extract and value addition to extract hydrocarbon have been addressed in the present study by performing quantitative evaluation of distillate products from processing of the pseudo-raffinate (generated from the extract phase using antisolvent) in hydrocracker and fluidized catalytic cracking (FCC) processes. The present study also demonstrates the approach for improving the quality of extract as a carbon black feedstock (CBFS). The benefits and befitting of disposal of extract hydrocarbon in a delayed coker as a blending stream with vacuum residue (VR) is also discussed in detail.

1. INTRODUCTION The deep removal of sulfur and polyaromatics from gas oil is the need of the hour to minimize the emissions of oxides of sulfur, polyaromatic hydrocarbons (PAHs), and particulate fines, to minimize corrosion and wear of engine systems, and to improve the performance of emission control technologies used for exhaust after treatment systems, which are adversely affected by sulfur poisoning.1−4 Hydrotreatment is the well-established process for removal for sulfur and polyaromatics from gas oil in the refineries. It is known that dibezothiophene, benzonaphthothiophenes, and their alkylated derivatives are refractory sulfur compounds and remain in hydrotreated gas oil as well.5,6 The condensed polyaromatics in gas oil act as an inhibitor during hydrotreatment of refractive sulfur compounds because of strong competition among aromatic and sulfur compounds for adsorption on the catalyst active site. It is also reported that nitrogen compounds, such as carbazole and acridine (dibenzo[b,e]pyridine), retard the performance of hydrotreatment even at low concentrations.1,7,8 Therefore, to overcome the inhibitory action of polyaromatics and nitrogen compounds and the refractive nature of sulfur compounds during removal of these impurities using hydrotreatment, very severe operating conditions of temperature and pressure, expensive catalysts, significantly increased H2 consumption, and H2S impurity-free H2 recirculation are required. This results in a significant increase in operating cost and huge capital investment requirement to revamp the existing facilities.8−12 Therefore, refiners are looking for optimizing the desulfurization process © XXXX American Chemical Society

and investment cost for meeting the new stringent ultra lowsulfur specification of gas oil. The desulfurization process can be made more economical using a combination of processes. The selective solvent extraction process can remove the refractive sulfur compounds, polyaromatics, and nitrogen compounds to a very large extent.13,14 Its use prior to the conventional hydrotreatment process can help in attaining the goal of deep desulfurization at lower capital and operating costs than required in a highseverity standalone hydrotreatment method.8,15 Solvent extractive removal of sulfur from synthetic diesel (using bezothiophene, dibezothiophene, and their alkylated derivates) has been well-reported in the literature.16−18 Removal of sulfur, aromatic, and nitrogen compounds from actual liquid fuels using solvent extraction has also been reported in various studies.13,15,19−29 In one of our previous studies,13 extraction of sulfur and polyaromatic impurities from actual straight-run gas oil (SRGO) was studied using various solvents, such as acetonitrile, N,N-dimethylformamide, furfural, N,N-dimethylacetamide, and dimethyl sulfoxide. In the other study,14 a strategy for screening of solvents for sulfur, nitrogen, and aromatic compound removal from gas oil was developed. In refinery, SRGO from the atmospheric distillation column (ADC), light cycle oil (LCO) from fluidized catalytic cracking (FCC), and coker gas oil (CGO) from fractionators of the Received: April 20, 2015 Revised: May 29, 2015

A

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels coker unit are used to generate the mixed gas oil (MGO). In the refinery, FCC and coker units process heavy fractions of crude oil, such as heavy vacuum gas oil (HVGO), deasphlated oil (DAO), and long residue (LR). Therefore, gas oil streams generated from cracking of these heavy feedstocks in FCC and coker have significantly more polyaromatics and refractive sulfur compounds in comparison to SRGO. Therefore, operating parameters and the solvent system to be used in solvent extraction for sulfur and aromatic removal from LCO, CGO, SRGO, and MGO will be sufficiently different. There is no comprehensive solvent extraction study that optimizes the operating conditions and solvent system for removal of sulfur and polyaromatics from the LCO, CGO, and mixture of LCO, CGO, and SRGO streams in the literature. In the present study, removal of sulfur and aromatic compounds from SRGO, LCO, CGO, and MGO (mixture thereof) was studied using the well-proven and stable industrial N-methyl-2-pyrrolidone (NMP) solvent. First, the effect of the extraction temperature (TE), solvent/feed ratio (S/F), and antisolvent concentration (Wc) in the main solvent was studied in a single-stage batch extractor for SRGO, LCO, and CGO. The degree of sulfur removal (Dsr), yield of extracted gas oil (Y %), and performance factor (Pf,α), which combines both Dsr and yield, were considered as responses. Further studies were carried out for MGO (mixture of SRGO, LCO, and CGO) in single-stage and continuous countercurrent packed-bed extractors using the estimated values of TE, S/F, and Wc by summation of multiplication of volumetric composition of MGO and optimized operating conditions for individual SRGO, LCO, and CGO streams. The major issues associated with solvent extraction for gas oil desulfurization are to minimize the loss of valuable hydrocarbon with extract and value addition to extract hydrocarbon. Both of these issues have been addressed in the present study by generating pseudoraffinate from the extract phase using water as the antisolvent. Quantitative estimation of distillate products generated from processing the pseudo-raffinate in hydrocracker and FCC processes was performed to demonstrate use of pseudoraffinate in existing facilities. The strategy for disposal of extract hydrocarbon has also been discussed.

Table 1. Properties of SRGO, LCO, and CGO Feed Samples parameter

SRGO

sulfur (wt %) 1.360 density at 20 °C 0.85184 RI at 20 °C 1.4727 monoaromatics (wt %) 13.9 diaromatics (wt %) 10 polyaromatics (wt %) 5.5 non-aromatics (wt %) 70.6 boiling range, ASTM D86 (°C) initial boiling point (IBP) 236.5 30 vol % 285.2 50 vol % 305.1 70 vol % 330.3 95 vol % 383.5 final boiling point (FBP) 386.7 distillate 97.7 residue 1.9 loss 0.4

LCO

CGO

0.676 0.93835 1.5426 17.5 27.1 17.9 37.5

0.268 0.85291 1.4802 13.5 11.7 6.7 68.1

191.3 282.2 308.2 335.6 376.9 382.2 98 1.8 0.2

169 220.3 261.1 291.5 343.1 348.3 97.8 1.7 0.5

glass mixer settler provided with a stirrer thermostatic bath to generate the oil-rich phase (raffinate phase) and solvent-rich phase (extract phase). The raffinate phase was washed with water to remove a small quantity of solvent present because of equilibrium solubility of the solvent in hydrocarbon. Water-washed raffinate phase termed as raffinate was used for calculation of the yield, as defined in our previous study.13 Moisture of the solvent-free raffinate was removed using anhydrous calcium chloride. Thereafter, it was analyzed for its sulfur and aromatic concentrations. A schematic diagram of the continuous countercurrent extraction experimental setup is shown in Figure 2. Continuous countercurrent extraction of feed oil was carried out in a jacketed Pyrex glass column of 20 mm internal diameter. Column was filled up to 140 mm of its height with a ceramic intalox saddle packing of 7−9 mm size. The settling zones of 16 mm height were provided on both side of the packing zone at the top and bottom of the column. Feed and solvent were pumped using the metering pumps at the bottom and top of the column, respectively. Flow rates of feed and solvent were 5 and 7 mL/ min, respectively, to obtain a desired solvent/feed ratio of 1.4. In extraction runs, feed was used as the dispersed phase and the interface was observed at the top of the column. Level of interface was kept constant in the settling zone at the top of the column during the run. The temperature of the column was maintained by circulating the hot water in the jacket of the column. Steady state of the column was confirmed by a constant value of the RI measured for top hydrocarbon samples time to time before collecting the sample for analysis. Raffinate phase and solvent-rich phase were obtained from top and bottom of the column, respectively. Raffinate phase was further treated in the same way as in single-stage equilibrium experiments. It may be mentioned that a small quantity of NMP will dissolve in the raffinate phase coming from the extractor top. Because NMP is highly water-soluble, therefore, raffinate phase washing is generally performed in the industry using water in a subsequent column to remove NMP. Finally, nitrogen-free raffinate is obtained. In the present study, raffinate was washed with warm water to remove the solvent (Figure 2). A trace amount of moisture present in the waterwashed solvent was removed using anhydrous CaCl2, and thereafter, it was used for further analysis. 2.4. Generation of Pseudo-raffinate and Extract Hydrocarbon. Pseudo-raffinate is the hydrocarbon-rich phase generated by the addition of water in the extract phase obtained from the continuous extraction column. Because the amount of solvent and hydrocarbon in the raffinate phase are known, the amount of solvent and hydrocarbon in the extract phase can be estimated using the material balance. Amount of water to maintain the given percentage of water in the solvent of the extract phase was estimated. A total of 500 mL of the extract phase along with the estimated amount of water for

2. EXPERIMENTAL SECTION 2.1. Materials. SRGO, LCO, and CGO samples were obtained from an Indian refinery. Physicochemical properties of these samples are given in Table 1. The sulfur-type speciation analyses of feed samples are given in Figure 1. NMP (minimum assay of 99.7%) was used as the solvent. Anhydrous calcium chloride was used to remove water from the hydrocarbon phase. 2.2. Methods of Analysis. Density was determined using an apparatus manufactured from a Metller Toledo Japan DE45 densitometer at a temperature of 20 °C. Refractive index (RI) was determined using an Abbe refractometer RE45 at 20 °C. The total sulfur content of the feed and product samples was estimated by the Xray fluorescence (XRF) method using an ASOMA ED XRF analyzer Spectro Phoenix II make. Pulsed-flame photometric detector (PFPD) inbuilt gas chromatography (GC) was used to detect the sulfur-type speciation analysis. ASTM D86 method was used for determining the boiling range of samples. An ultraviolet (UV) spectrophotometric technique was used for estimation of mono-, di-, and polyaromatic contents in samples. Conradson carbon residue (CCR), which provides some indication of the relative coke-forming tendency of stream, was estimated using the ASTM D4530-03 method. 2.3. Apparatus and Procedure. Details of the single-stage equilibrium extraction system used in this study are given in our previous study.13 Gas oil and NMP solvent were charged in a jacketed B

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 1. PFPD spectra of SRGO, LCO, and CGO. the corresponding water percentage in the solvent was retained in the batch single-stage equilibrium extraction system for 15 min at 40 °C. Mixture was stirred for 5 min. A settling time of 45 min was provided for phase separation. Pseudo-raffinate was collected and water-washed. Traces of water from pseudo-raffinate were removed using anhydrous CaCl2. Hydrocarbon phase from the extract phase was generated by adding the excess amount of water and hexane in the extract phase in a 3 L separating funnel. After separation of the hydrocarbon phase, the remaining extract phase was treated 4 times in the same way as described above to remove the hydrocarbon from the solvent. Finally, hexane from the collected hydrocarbon phase was recovered using the distillation to generate the extract hydrocarbon. Reported results are an average of triplicate experimental runs, and a maximum deviation of 5% was observed from the average value.

3. RESULTS AND DISCUSSION 3.1. Analysis of Feed Samples for Solvent Selection. Properties of SRGO, CGO, and LCO feed samples shown in Table 1 suggest that these feeds are quite different from each other. Intensity of peaks in PFPD spectra of these samples shown in Figure 1 indicates that SRGO has a much higher concentration of sulfur compounds than LCO and CGO. The increasing trend of density and RI are in order: SRGO < CGO < LCO, which is the same as for the total aromatic content and the reverse of the non-aromatic content. This clearly suggests that the RI and density of the sample are a function of its aromatic and non-aromatic contents. Analysis of feed samples

Figure 2. Schematic diagram of the experimental setup for continuous countercurrent extraction.

C

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

section 2.3 at S/F of 1.0 and TE of 45 °C for analyzing the effect of the water content in the solvent (Wc) on the yields of raffinate (Y%), Dsr, and Pf,0.5. Properties of raffinate samples generated during the extraction experiments are given in Table 2.

also reveals that SRGO and CGO are comparable in term of their composition but the sulfur content of SRGO is significantly higher than that of CGO. It is reported that cracked gas oil has significantly more aromatics than SRGO.9 Concentration of polyaromatics increases with an increase in the boiling point of the feed. Comparable composition of SRGO and CGO may be attributed to the significantly lower final boiling point of CGO (343.1 °C) against 383.5 °C of SRGO. LCO composition, i.e., concentration of di- and polyaromatic contents, is much higher than those of SRGO and CGO. When feed is contacted with a polar solvent in the solvent extraction process, the feed is distributed in two phases. One is the hydrocarbon-rich phase (raffinate phase), and the second is the solvent-rich phase (extract phase). When a feed is contacted with a given polar solvent and does not form two phases, solvent extraction with that solvent is not possible. In single-stage solvent extraction, the two-phase formation largely depends upon availability of the sufficient density difference between the extract phase and the raffinate phase. The density of the raffinate phase is governed by the density of feed, whereas the density of the extract phase is governed by the density of the solvent and amount of dissolved hydrocarbon in the solvent. In the countercurrent solvent extraction process, the feed comes in contact with the extract phase, and to avoid the single-phase formation at the feed inlet, there should be a sufficient density difference between the feed and the extract phase. It is known that aromatic compounds are more soluble than non-aromatic compounds in a polar solvent. This implies that the higher the aromatic content of feed, the larger the density difference required between the feed and the solvent for twophase formation to compensate for the larger density reduction of the extract phase in comparison to the feed containing a lower amount of aromatics. In general, the feed having the higher concentration of aromatics has a higher density (Table 1). It may be mentioned that, in our previous study,14 DMF, NMP, DMA, and furfural seemed to be the best solvents for sulfur and aromatic removal from gas oil stream. However, in view of the oxidation stability and better density, NMP was selected to carry out solvent extraction of SRGO, LCO and CGO. 3.2. Single-Stage Extraction of SRGO, LCO, and CGO. The solvent extraction process works on the principle of the solubility difference between the solute and the carrier phase in the solvent. It is apparent that solubility of compounds in solvent during the extraction process can be modified to a certain extent by modifying the operating conditions, such as S/ F ratio, TE, and more particularly, solvency power of the solvent using antisolvent (water). The results of experiments carried out for studying the effect of these operating parameters on the performance of the extraction process for each sample gas oil stream are discussed in subsequent sections. During extractive desulfurization of gas oil, maximization of the degree of sulfur removal (Dsr) and raffinate yield (Y%) is desired. However, the operating conditions affect the Dsr and Y% in opposite ways. Performance factor (Pf,α) represents a single factor, which couples both Dsr and Y%, along with the weight factor (α), to sulfur removal. It is calculated using following equation:13 Pf, α = αDsr + (1 − α) × yield (%)

Table 2. Effect of the Water Content in the Solvent (Wc) on the Sulfur Content in Raffinate and Yields of Raffinate (Y%) at S/F of 1.0 and TE of 45 °C sulfur (wt %)

yield (vol %)

water content

SRGO

LCO

CGO

SRGO

LCO

CGO

0.0 3.0 5.0 7.0

0.704 0.790 0.881 1.003

S.F. 0.359 0.373 0.388

0.110 0.112 0.119 0.131

80.0 88.8 90.0 93.6

S.F 42.0 48.4 56.0

48.0 68.8 74.0 79.2

It can be seen from the results (Table 2) that, within the selected range (0.0−7.0 vol %) of the water content in the solvent, Dsr of the raffinate decreases (shown by an increase in sulfur weight percent in raffinate), whereas the yield (Y%) of raffinate increases with an increase in the water content of the solvent. Y% value for these streams followed the order of SRGO > CGO > LCO, which is the same order as the nonaromatic content in these samples. It is important to note that an increase in the yield of raffinate by the addition of the first 3.0 vol % of water in the solvent is much higher than the same amount of subsequent water addition. It was observed that LCO does not form the two phases with pure NMP. Most probable reasons for this phenomenon are high solvency power of NMP than other solvents and a high concentration of aromatic in LCO, which leads to dissolution of a huge quantity of hydrocarbon in the solvent and decreases the density difference between the extract and raffinate phases, and a large concentration of aromatics in the extract phase does not leave the non-aromatic compounds to form the raffinate phase at the operating condition of single-stage extraction. The addition of water to NMP decreases its solvency power for hydrocarbons and results in two-phase formation. In view of this, application of NMP without water dilution cannot be used for LCO extraction. Dsr and Pf,0.5, which combine the contradicting effect of the water content on the sulfur content and yield of raffinate for the equal weight factor assigned to Y% and Dsr, are given in Figure 3. It is observed that a gradient of change in Dsr and Pf,0.5 with an increase in the water content is different for different SRGO, LCO, and CGO samples. For a selected range of Wc, Dsr follows the order: CGO > LCO > SRGO. It may be seen that absolute and slope changes in Dsr are a maximum for SRGO and a minimum for LCO. Pf,0.5 values for SRGO first increase and then decrease with an increase in Wc. For LCO stream, Pf,0.5 continuously increased with an increase in Wc, but the change in the slope of Pf,0.5 is very small. However, for CGO, Pf,0.5 increases with an increase in Wc; however, a change in Pf,0.5 after 5% value of Wc is marginal. Pf,0.5 values for SRGO are higher than those for LCO until the Wc value of 3%, beyond that, the reverse trend is observed. This is attributed to a higher SRGO raffinate yield, with this water content superseding the higher Dsr of CGO. In view of above, Wc values of 3, 7, and 5% seem to be the best for SRGO, LCO, and CGO solvent extraction for studying the effect of other operating conditions on Y%, Dsr, and Pf,0.5.

(1)

3.2.1. Effect of the Water Concentration in the Solvent (Wc). The solvent extraction of actual SRGO, LCO, and CGO was carried out using NMP as per the procedure described in D

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 3. Effect of the water content in the solvent (Wc) on the degree of sulfur removal (Dsr) and performance factor (Pf,0.5).

3.2.2. Effect of the Solvent/Feed Ratio (S/F). It is known that a higher S/F ratio enhances the extent of removal of impurities from the feed but also results in the loss of the desired molecules along with impurity molecules in the solvent. To analyze the effect of S/F for the SRGO, LCO, and CGO feed samples, single-stage solvent extraction of these feeds at various S/F ratios was carried out with aqueous NMP containing a respective optimized Wc value and at TE of 45 °C. The analysis of raffinate samples is given in Table 3. As

observed that, at S/F value of 1.0, Dsr of SRGO and LCO are comparable, but for S/F ratio higher than 1.0, Dsr of the SRGO sample is much higher than that of LCO. This may be attributed to the much higher sulfur content of SRGO than LCO and their much different aromatic contents. The Pf,0.5 values for all samples do not show the similar and unidirectional trend with the increase in S/F. The change in the Pf,0.5 value for SRGO and LCO streams is marginal after S/ F values of 1.5 and 1.0, respectively, whereas the Pf,0.5 value for CGO is at a maximum at the S/F value of 1.5. This change illustrates that, after a certain value of S/F, the trade-off between the increase in Dsr and decrease in Y% is established. In the solvent extraction process, with solvent being an expensive material, it is recovered from the extract and raffinate phases for its reuse to sustain the economics of the process. For gas oil to have the boiling range up to 380 °C, a small amount of solvent from the raffinate phase can be recovered using the water washing but solvent needs to vaporize for its recovery from the solvent-rich extract phase. In view of the latent heat of vaporization requirement, it is easy to understand that solvent recovery is an energy-extensive step. The energy requirement will depend upon the amount of solvent to be recovered from the extract phase, which, in turn, depends upon the S/F used in the process. Considering this, the S/F values of 1.5, 1.0, and 1.5 seem to be best for use in solvent extraction of SRGO, LCO, and CGO, respectively. 3.2.3. Effect of the Extraction Temperature (TE). It was observed in our previous solvent extraction study for SRGO13 that TE affects the selectivity and solvency power of solvents in the solvent extraction process, which basically governs Dsr and

Table 3. Effect of the Solvent/Feed Ratio (S/F) on the Sulfur Content in Raffinate and Yields of Raffinate (Y) at SRGO Wc = 3.0, LCO Wc = 7.0, CGO Wc = 5.0, and TE = 45 °C sulfur (wt %)

yield (Y) (vol %)

S/F

SRGO

LCO

CGO

SRGO

LCO

CGO

1.0 1.5 2.0 2.5

0.790 0.526 0.470 0.422

0.388 0.318 0.287 0.253

0.119 0.086 0.076 0.069

88.8 77.5 75.0 70.0

56.0 47.5 45.0 41.5

74.0 70.0 62.5 60.0

expected, S/F is found to affect the sulfur content in raffinate and yields of raffinate (Y%) and that the sulfur content of raffinate and yield of raffinate decrease with the increase in S/F. However, the percent change in the sulfur content and yield is not the same for all feeds and depends upon the feed composition. Dsr and Pf,0.5 for SRGO, LCO, and CGO samples are given in Figure 4. It is clear that the gradient for change in Dsr removal is highest for the first 0.5 addition to the base S/F ratio of 1.0. It is

Figure 4. Effect of S/F on the degree of sulfur removal (Dsr) and performance factor (Pf,0.5). E

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels Y. To analyze the effect of TE on Y%, Dsr, and Pf,0.5, extraction experiments for SRGO, LCO, and CGO streams were carried out with selected Wc and S/F values in the previous sections 3.2.1 and 3.2.2 at 55 and 65 °C. The results of the above experiments are given in Table 4. Results suggest that the sulfur

Table 5. Analysis of Feed and Products Obtained MGO Extraction (Feed, SRGO/LCO/CGO Volumetric Ratio = 70:20:10, S/F = 1.4, TE = 45 °C, Wc = 4.0%, and α = 0.5) properties of stream

sulfur (wt %)

45.0 55.0 65.0

SRGO

1.171 0.512 0.8692 0.8386 1.4872 1.4661 14.6a 10.2 13.9a 7 8.3a 2.8 63.2a 80 Calculated Response yield (%) 72.1 sulfur removal, Dsr (%) 56.3 performance factor, Pf,0.5 64.2 monoaromatics removal (%) 30.3 diaromatics removal (%) 49.5 polyaromatics removal (%) 66.2

LCO

yield (vol %) CGO

0.526 0.388 0.086 0.465 0.378 0.082 0.437 0.362 0.084 Degree of Sulfur Removal 61.3 42.6 67.9 65.8 44.1 69.4 67.9 46.4 68.7

45.0 55.0 65.0

SRGO

LCO

CGO

77.5 56.0 70.0 71.0 50.0 67.0 68.5 42.0 62.5 Performance Factor 69.4 49.3 70.2 68.4 47.0 68.2 68.2 44.2 65.6

i

0.275 0.82164 1.4554 7.2 0.9 0.4 91.5 63.3 76.5 69.9 50.8 93.5 95.2

Estimated values using the property of mixing streams and their mass fraction in MGO. R = raffinate; E = extract.

Table 6. Boiling Range of MGO and Its Raffinate and Extract Products (ASTM D86)a volume percent

MGO

R-EPCE

E-EPCE

IBP 30 50 70 FBP distillate residue loss

201.3 279.6 302.6 330.6 383.7 97.7 1.9 0.4

219.4 286 307.5 334.2 385.3 97.7 1.9 0.4

185.3 272.3 297.1 328.5 383.6 97.6 2 0.4

a MGO, mixed gas oil; R-EPCE and E-EPCE, raffinate and extract obtained from continuous extraction of MGO, respectively.

continuous countercurrent extraction is significantly higher than the batch extraction process. This is attributed to the availability of more than one equilibrium stage in the continuous extractor and an improved concentration gradient between the contacting portion of the solvent and feed. In the continuous countercurrent extraction, the raffinate phase leaves the extractor at a location where it is in contact with the solvent, having zero impurities, whereas in batch extraction, the raffinate phase is in equilibrium with the extract phase, having significant impurities to be removed from feed. This results in a higher concentration gradient in the continuous column than batch extraction. It is clear that sulfur, diaromatic, and polyaromatic compounds can be removed from the MGO up to the extent of 76.5, 93.5, and 95.2% using the continuous extraction column. Pf,α values of the continuous column are higher than batch extraction, which reveals the importance of a more theoretical separation stage available in continuous extraction. For a given separation requirement of impurities from feed, Pf,α is a function of α, which denotes the weight factor assigned to Dsr. The extent of the sulfur content in the raffinate will depend upon requirement of the end-use process for raffinate. Therefore, selection of the value of α to evaluate Pf,α may vary depending upon the importance of the sulfur content and yield of desulfurized MGO in that situation. Considering this, the sensitivity analysis of Pf,α with respect to the α value for MGO

n

∑ viOPi ,j

R-continuous

a

content in raffinate and Y% decrease, whereas Dsr increases with an increase in TE. Also, a decrease in Pf,0.5 with an increase in TE is observed. This suggests that the rate of decrease in the raffinate yield is higher than the corresponding increment in Dsr with an increase in TE. It can be understood that, upon increasing the weight factor to Dsr, the Pf,α value increases. However, the raffinate yield may become important after meeting the given specification of the sulfur content in raffinate. Therefore, an increase in Dsr with an increase in TE will affect the process economics and performance of the process adversely. Therefore, selection of the temperature will greatly depend upon the demand of the situation for compromising the yield and sulfur contents of raffinate. 3.3. Single-Stage and Continuous Extraction of MGO. In the refinery, gas oil streams from ADU, FCC, and delayed coking unit (DCU) are blended to make MGO, which is generally hydrotreated to meet the sulfur and aromatic specifications required for its sale in the open market. In general, the MGO contains 60−75% SRGO, 30−15% LCO, and 10−20% CGO depending upon the configuration of the refinery and type of processed crude oil. In the present study, MGO was prepared by mixing SRGO, LCO, and CGO in the volumetric ratio of 70:20:10. Batch and continuous extraction of MGO was carried out using the values of TE, S/F, and Wc estimated using the correlation given below

OPe, j =

R-batch

sulfur (wt %) density at 20 °C RI at 20 °C monoaromatics (wt %) diaromatics (wt %) polyaromatics (wt %) non-aromatics (wt %)

Table 4. Effect of the Solvent/Feed Ratio (S/F) on the Sulfur Content in Raffinate and Yields of Raffinate (Y) at SRGO with Wc = 3.0 and S/F = 1.5, LCO with Wc = 7.0 and S/F = 1.0, and CGO with Wc = 5.0 and S/F = 1.5 extraction temperature (°C)

feed

(2)

where OPe,j is the estimated operating parameter for MGO, with j being Wc, S/F, or TE, and OPi,j and vi are optimized parameters and volume fraction, respectively, for i = SRGO, LCO, and CGO. Analysis of MGO and products obtained from the batch and continuous extraction of MGO are given in Table 5, whereas the boiling range of MGO, extract, and raffinate obtained from the MGO continuous extraction is given in Table 6. It can be clearly seen from the results (Table 5) that the solvent extraction process is capable of removing the refractive sulfur compounds and aromatic compounds to a significant extent. The sulfur and aromatic removal efficiency of a F

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Table 7. Properties of Pseudo-raffinate and Extract Samplesa

batch and continuous extraction was estimated and is given in Figure 5. The value of Pf,α for batch extraction decreases, whereas for continuous extraction, it increases with the increase in the α value assigned to Dsr.

pseudo-raffinate PR1 Wc = 8.0%

properties sulfur (wt %) density at 20 °C RI at 20 °C yield on feed basis (%)

1.1215 0.86147 1.4826 5.34 extract hydrocarbon

PR2 Wc = 12.0% 1.5821 0.889 1.4993 4.48

properties

E-EPCE

E-EPSRG

sulfur (wt %) density at 15.5 °C specific gravity at 15.5 °C average boiling point (°C) BMCI CCR (wt %)

2.376 0.95230 0.95280 299.3 79.5 NT

2.531 0.98144 0.98194 299.3 93.3 0.18

a

E-EPCE, extract obtained from the extract phase of continuous extraction (EPCE); E-EPSRG, extract obtained from the extract phase of pseudo-raffinate generation (EPSG); RI, refractive index; BMCI, Bureau of Mines correlation index; and NT, not tested.

Figure 5. Sensitivity analysis of the performance factor (Pf,α) with respect to the α value.

3.4. Generation of Pseudo-raffinate and Extract. In solvent extraction of gas oil, it is desirable to maximize the yield and minimize the sulfur content of desulfurized gas oil. There is no sharp separation between impurity compounds and desired compounds in the extraction process. There is a loss of desired compounds with sulfur and aromatic compounds in the extract phase. Even sometimes, the loss of desired paraffinic compounds with extract becomes undesired compounds because their presence deteriorates the quality of extract to be used as carbon black feedstock (CBFS). Therefore, removal of paraffinic-rich material from the extract phase obtained from the continuous extraction column will be very useful to produce the hydrocarbons, which can be further treated in the secondary process to generate the distillate product and improve the quality of extract to be used as CBFS. Pseudo-raffinate and extract hydrocarbon were generated from the extract phases using the procedure described in section 2.4. Pseudo-raffinate (PR1) was generated from the extract phase of the continuous extraction column by addition of a given amount of water corresponding to a given Wc in the extract phase solvent equal to 8%. Another pseudo-raffinate (PR2) was generated from the extract phase after generating the PR1 by further addition of a given amount of water to increase its concentration in the extract phase solvent from 8 to 12%. The properties of pseudoraffinates and extracts are given in Table 7. Marginal lower value of density and RI values for PR1 and higher value for PR2 in comparison to MGO suggest that PR1 contains a marginal lower aromatic content and PR2 contains a marginal higher aromatic content than MGO. The yield of the combined mixture of PR1 and PR2 is 9.82%. Properties of their mixture can be estimated using their mass or volume fraction in the mixture and their properties. Properties of pseudo-raffinate suggest that this is a valuable material and can be processed in secondary conversion processes, such as a hydrocracker and FCC, for generating the distillate products with a reduced sulfur content. Detail of their use in potential options is discussed in the subsequent section. 3.5. Use of Pseudo-raffinate and Extract. 3.5.1. Use of Pseudo-raffinate. The yield of pseudo-raffinate would vary in the range of 5−15%. It is common practice to design a unit with 15−20% over the design margin. Considering this, it is quite possible to process the generated pseudo-raffinate in the

existing hydrocracker and FCC. The hydrocracker and FCC are designed for processing heavier feedstocks having a boiling range up to ∼450−550 °C and operate at high severe operating conditions. PR1 and PR2 have been generated from the gas oil, having the boiling range up to ∼380 °C. Metal content and viscosity of crude distillate fractions increase with the increase in their boiling range. Metal content of PR1 and PR2 will be much less than the heavy feedstocks used in the hydrocracker and FCC; therefore, their blending with conventional feedstock would be helpful in diminishing deactivation of the catalyst. The quantitative estimation of distillate products for processing of PR1 and PR2 in the hydrocracker and FCC was carried using the correlations given in the literature.30 Because PR1 and PR2 are generated from the extract phase obtained from continuous extraction of MGO, their average boiling point will be close to extract hydrocarbon obtained from the extract phase of MGO continuous extraction (E-EP). Therefore, distillation data of EEP was used to represent the average boiling point of PR1, PR2, and PR1&2 streams in the correlations used for quantitative estimation of distillate products for processing of PR1 and PR2 in the hydrocracker and FCC. The values of the product yield generated in the hydrocracker and FCC units for PR1, PR2, and their mixture PR1&2 are given in Table 8. 3.5.2. Use of Extract. The application of the extract as CBFS in a carbon black generation unit is one of the potential options for its use without any further processing. It is known that the quality of CBFS is measured in terms of its Bureau of Mines correlation index (BMCI) value. Therefore, to evaluate the feasibility of extract use as CBFS, BMCI of extract products was estimated using the correlation given below31 BMCI = 473.7Sg − 456.8 + (48460/Tb)

(3)

where Sg is the liquid specific gravity at 60 °F and Tb represents the average boiling point (K). The average boiling point is the arithmetic average of temperatures at 10% interval from 20 to 80%. Sg was obtained by converting the density of extract stream from 20 to 15.5 °C and then density at 15.5 °C to specific gravity using the petroleum measurement tables. The values of estimated BMCI values along with the estimated Sg of extract samples are tabulated in Table 8. It can be seen that the BMCI value (Table 6) of the extract obtained G

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

raffinate. This has resulted in drastic improvement in the quality of extract hydrocarbon to be used as CBFS. It may be mentioned that the BMCI values and sulfur content of E-SR are in the range for CBFS, which is being marketed by refineries in India.32 DCU can convert extract into light and middle distillates. Delayed coking is the process where carbon is rejected to meet the requirement of hydrogen in distillate products. Properties of E-SR suggest that DCU can be another potential option for extract use. It may important to mention that the metal content, CCR, and viscosity of E-SR would be much lower in comparison to regular feedstocks, such as vacuum residue (VR), thermal tar, pyrolysis tar, and pitch streams to DCU. The lower viscosity and metal content of E-SR would improve the quality of coke generated. Further, it is vital to point out that coke formation in the DCU is a strong function of the CCR value [coke (wt %) = 1.6CCR] of feedstock.33 The CCR value (0.18%) of E-SR is much lower than the CCR value of VR (varies in the range from 8.3 to 21.84%),34,35 one of the most commonly used feedstocks in DCU. A lower CCR value of ESR indicates that its blending with the regular feedstock of DCU will reduce the amount of coke formation and increase the distillate yield.

Table 8. Product Yield Distribution for Pseudo-raffinate Processing in Hydrocracker and FCC Unitsa product name

PR1

Hydrocracker Unit H2 (wt %) 2.38 H2S (wt %) 1.19 light gasoline (wt %) 9.38 refinery fuel gas (wt %) 1.88 C4 LPG (wt %) 5.53 naphtha (wt %) 29.08 diesel (wt %) 55.31 FCC Unit coke (wt %) 3.6 flue gases (wt %) 1.2 H2S (wt %) 0.1 gasoline (wt %) 41.6 LPG (wt %) 13.4 LCO (wt %) 24.4 HCO (wt %) 15.6

PR2

PR1&2

2.96 1.68 7.97 1.75 4.70 24.70 62.16

2.60 1.42 8.84 1.83 5.22 27.42 57.87

3.9 1.3 0.2 41.6 13.0 23.8 16.2

3.7 1.3 0.1 41.6 13.2 24.1 15.9

a

LPG, liquefied petroleum gas; LCO, light cycle oil; and HCO, heavy cycle oil.

from the extract phase of pseudo-raffinate generation (E-SR) is significantly higher in comparison to the extract obtained from the extract phase of continuous extraction (E-EP). It suggests that the addition of water in the extract phase has increased the concentration of aromatic compounds in extract hydrocarbon by concentrating the non-aromatic compounds in pseudo-

4. CONCLUSION Currently, refineries are looking for a cost-effective process that can produce low-sulfur diesel. It is well-known that severity of the hydrotreater (hydrodesulfurization unit) depends upon the

Figure 6. Schematic of the novel strategy for removal of sulfur and aromatic compounds from various gas oil streams and extract use. SRGO, straight-run gas oil; LCO, light cycle oil; CGO, coker gas oil; Des-GO, desulfurized gas oil (raffinate); and FCC, fluidized catalytic cracking. H

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

(3) Pryor, P.; Cooper, R. Clean Fuel Technology; http://www2. dupont.com/Clean_Technologies/en_US/assets/downloads/ Hydrocarbon_Engineering_article_June_2009.pdf (accessed March 10, 2015). (4) Kumar, S.; Srivastava, V. C.; Badoni, R. P. Oxidative desulfurization of dibenzothiophene by zirconia-based catalysts. Int. J. Chem. React. Eng. 2014, 12, 295−302. (5) Choudhary, T. V. Structure−reactivity−mechanistic considerations in heavy oil desulfurization. Ind. Eng. Chem. Res. 2007, 46, 8363−8370. (6) Shiraishi, Y.; Hirai, T.; Komasawa, I. Petrochemical desulfurization and denitrogenation process for vacuum gas oil using an organic two-phase extraction system. Ind. Eng. Chem. Res. 2001, 40, 293−303. (7) Skov, E. R.; Dolbear, G. E. Synergistic extractive desulfurization processes. In Practical Advances in Petroleum Processing; Hsu, C. S., Robinson, P. R., Eds.; Springer: New York, 2006; Chapter 12, pp 373− 379. (8) Srivastava, V. C. An evaluation of desulfurization technologies for sulfur removal from liquid fuels. RSC Adv. 2012, 2, 759−783. (9) Sharma, M.; Sharma, P.; Kim, J. N. Solvent extraction of aromatic components from petroleum derived fuels: A perspective review. RSC Adv. 2013, 3, 10103−10126. (10) Palmer, R. E.; Torrisi, S. P. Hydrotreater revamps for ULSD fuel. PTQ Revamp Oper. 2003, 15−18. (11) Sayles, S.; Bailor, J.; Ohmes, R. ULSD problems and solutions. PTQ Autumn 2004, 107−115. (12) Smagilov, Z.; Yashnik, S.; Kerzhentsev, M.; Parmon, V.; Bourane, A.; Al-Shahrani, F. M.; Hajji, A. A.; Koseoglu, O. R. Oxidative desulfurization of hydrocarbon fuels. Catal. Rev. Sci. Eng. 2011, 53, 199−255. (13) Kumar, S.; Srivastava, V. C.; Nanoti, S. M.; Nautiyal, B. R.; Siyaram. Removal of refractive sulfur and aromatic compounds from straight run gas oil using solvent extraction. RSC Adv. 2014, 4, 38830− 38838. (14) Kumar, S.; Srivastava, V. C.; Nanoti, S. M.; Kumar, A. Solvent evaluation for desulfurization and denitrification of gas oil using performance and industrial usability indices. AIChE J. 2015, DOI: 10.1002/aic.14809. (15) Gaile, A. A. Development and improvement of extraction processes for separation and purification of petroleum products. Russ. J. Appl. Chem. 2008, 81, 1311−1324. (16) Huang, C.; Chen, B.; Zhang, J.; Liu, Z.; Li, Y. Desulfurization of gasoline by extraction with new ionic liquids. Energy Fuels 2004, 18, 1862−1864. (17) Tian, Y.; Yao, Y.; Zhi, Y.; Yan, L.; Lu, S. Combined extraction− oxidation system for oxidative desulfurization (ODS) of a model fuel. Energy Fuels 2015, 29, 618−625. (18) Ke-dra-Krolik, K.; Fabrice, M.; Jaubert, J. Extraction of thiophene or pyridine from n-heptane using ionic liquids. Gasoline and diesel desulfurization. Ind. Eng. Chem. Res. 2011, 50, 2296−2306. (19) Yordanov, D.; Petkov, P.; Yankov, V. Optimization of the extraction of sulphur and polycyclic arene hydrocarbons from middle distillate petroleum fractions with a selective solvent. J. Univ. Chem. Technol. Metall. 2009, 44 (1), 29−33. (20) Petkov, P.; Tasheva, J.; Stratiev, D. Extraction approach for desulphurization and dearomatization of middle distillates. Pet. Coal 2004, 46 (2), 13−18. (21) Gaile, A. A.; Somov, V. E.; Zalishchevskii, G. D.; Kaifadzhyan, E. A.; Koldobskaya, L. L. Extractive refining of atmospheric gas oil with N-methyl pyrrolidone. Russ. J. Appl. Chem. 2006, 79, 590−595. (22) Gaile, A. A.; Saifidinov, B. M.; Kolesov, V. V.; Koldobskaya, L. L. Extractive refining of high-sulfur diesel fraction to remove organic sulfur compounds and aromatic hydrocarbons. Russ. J. Appl. Chem. 2010, 83 (3), 464−472. (23) Gaile, A. A.; Saifidinov, B. M.; Kolesov, V. V.; Koldobskaya, L. L. Multistep countercurrent extraction of organic sulfur compounds and arenes from the high-sulfur diesel fraction. Russ. J. Appl. Chem. 2010, 83 (3), 473−476.

concentration of refractive sulfur and polyaromatic compounds in gas oil. Solvent extraction can remove these compounds selectively to produce a very clean feed to the hydrotreater to generate the ultra-desulfurized diesel. In the literature, extraction studies are based on SRGO and do not include the LCO and CGO. It becomes important to understand the behavior of extraction for LCO and CGO streams. The present study targeted solvent extraction of all three streams (SRGO, LCO, and CGO) used in the blending pools of gas oil. Figure 6 shows a schematic of the conceptualized strategy used in this study for the removal of sulfur and aromatic compounds from various gas oil streams using solvent extraction and use of extract stream. The study reveals that operating conditions, such as solvent/feed ratio (S/F), extraction temperature (TE), and anti/cosolvent concentration, in main the solvent to be used in the solvent extraction process depend upon the gas oil selected and also that the selection of the operating condition depends upon the selected gas oil streams. Continuous countercurrent solvent extraction provides significantly better results for the degree of sulfur removal (Dsr) and performance factor (Pf,α) in comparison to single-stage extraction. Loss of paraffinic material with the extract hydrocarbon can be reduced to a great extent along with improvement in the quality of extract hydrocarbon to be used as CBFS by generating the pseudo-raffinate from the extract phase. Composition of pseudo-raffinate consisting of a minor portion of feed is similar to MGO and can be used in existing refinery processes, such as hydrotreater, hydrocracker, and FCC units, to generate the distillate products with a reduced sulfur content to diminish the severity of the hydrotreater for cost-effective production of lowsulfur gas oil. Raffinate produced from the continuous extraction was found to contain 76.5, 93.5, and 95.2% lower sulfur, diaromatics, and polyaromatics, respectively, in comparison to feed. This implies the potential of extraction to remove the refractive sulfur compounds and inhibitor polyaromatics from gas oil. It can be said that application of the extraction process as a complementary process to a conventional hydrotreatment process for deep desulfurization of gas oil may result in huge savings in energy and operating cost. Use of extract as either a CBFS or blending feed stream in DCU will compensate for the economic loss because of the reduced feed to hydrotreatment unit. However, it seems that the integrated process may be overall more economical than the single- or multi-stage hydrotreatment processes to be operated at very severe operating conditions to meet the strict sulfur and cetane number of transported gas oil.

■

AUTHOR INFORMATION

Corresponding Author

*Telephone: +91-1332-285889. Fax: +91-1332-276535. E-mail: [email protected] and/or [email protected]. Notes

The authors declare no competing financial interest.

■

REFERENCES

(1) Koltai, T.; Macauda, M.; Guevara, A.; Schulz, E.; Lemaire, M.; Bacaud, R.; Vrinat, M. Comparative inhibiting effect of polycondensed aromatics and nitrogen compounds on the hydrodesulfurization of alkyldibenzothiophenes. Appl. Catal., A 2002, 23, 253−26. (2) Kumar, S.; Srivastava, V. C.; Badoni, R. P. Oxidative desulfurization by chromium promoted sulfated zirconia. Fuel Process. Technol. 2012, 93, 18−25. I

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels (24) Toteva, V.; Topalova, L.; Manolova, P. Extractive dearomatization and desulphurization of a distillate gasoil cut with imethylformamide. J. Univ. Chem. Technol. Metall. 2007, 42, 17−20. (25) Abd El-Aty, D. M.; Sif El-Din, O. I.; Hassa, S. I.; Tawfik, S. M.; Hanafi, S. Evaluation of some organic solvents for refining diesel fuel fraction. Pet. Sci. Technol. 2009, 27, 861−863. (26) Hassan, S. I.; Omar, I.; Sif El-Din, S.; Tawfik, M. Solvent refining of straight run diesel fuel by various solvents phase equilibrium. J. Appl. Sci. Res. 2009, 5 (5), 515−521. (27) Otsuki, S.; Nonaka, T.; Takashima, N.; Qian, W.; Ishihara, A.; Imai, T.; Kabe, T. Oxidative desulfurization of light gas oil and vacuum gas oil by oxidation and solvent extraction. Energy Fuels 2000, 14, 1232−1239. (28) Dharaskar, S. A.; Wasewar, K. L.; Varma, M. N.; Shende, D. Z.; Yoo, C. K. Extractive desulfurization of liquid fuels by energy efficient green thiazolium based ionic liquids. Ind. Eng. Chem. Res. 2014, 53, 19845−19854. (29) Lo, T. C.; Baird, M. H. I.; Hansan, C. Handbook of Solvent Extraction; Wiley-Interscience: New York, 1983. (30) Maples, R. E. Petroleum Refining Process Correlations; HPI Consulatants, Inc.: Geneva, Switzerland, 1987; Chapters 3, 4, and 7. (31) Mektta, J. J.; Cunningham, W. A. Encyclopedia of Chemical Processing and Design; Marcel Dekker, Inc.: New York, 1990; Vol. 6, p 200. (32) Indian Oil Corporation Limited (IOCL). CBFS Specification; http://www.iocl.com/Products/CarbonBlackFeedStockSpecifications. pdf (accessed March 10, 2015). (33) HPI Consultants, Inc. Petroleum Refining Process Correlations; HPI Consulatants, Inc.: Geneva, Switzerland, 1996. (34) Kulkarni, R. P.; Pandit, A. B.; Joshi, J. B.; Kataria, K. L.; Tandon, D.; Kumar, M. M. Visbreaking studies in the presence of soaker internals. Ind. Eng. Chem. Res. 2010, 49, 11221−11231. (35) Parra, M. J.; León, A. Y.; Hoyos, L. J. Separation of fractions from vacuum residue by supercritical extraction. CT&F, Cienc., Tecnol. Futuro 2010, 4, 83−90.

J

DOI: 10.1021/acs.energyfuels.5b00834 Energy Fuels XXXX, XXX, XXX−XXX