Removal of refractory sulfur and aromatic compounds from straight run gas oil using solvent extraction

Abstract

In the present study, extraction of sulfur and polyaromatic impurities from actual straight run gas oil (SRGO) containing 1.3 wt% sulfur was studied using various solvents such as acetonitrile, N,N-dimethylformamide, furfural, N,N-dimethylacetamide and dimethyl sulfoxide. Effects of water as antisolvent, extraction conditions and types of extraction operation (batch, single and multi-stage and continuous) have been studied. Performance of solvent extraction process, which is governed by the degree of sulfur removal (D_sr) and yield of extracted SRGO (ESRGO), has been evaluated in terms of a performance factor (P_f,α), which has been defined in terms of weight factor (0 < α < 1) as: P_f,α = αD_sr + (1 − α) yield. DMF was found to be a better solvent in terms of P_f,α and regeneration ability. The possibility of utilization of extract as carbon black feed stock (CBFS) has also been discussed based upon the calculated bureau of mines correlation index (BMCI) values.

---

Introduction

Gas oil is one of the most commonly used transportation fuel and contributes significant harmful emission of NOx, SOx, hydrocarbons and particulate matter to the environment. This leads to serious environmental and health concerns, such as smog, global warming, water pollution, acid rain, cancer and neurotoxicity.^1–5 The quantity of these emissions significantly depends on the concentration of sulfur, nitrogen and aromatic compounds in the gas oil being used.⁵ Environmental regulations have been implemented across the globe to limit the sulfur and aromatic content of gas oil for improving the air quality.^6,7

Currently, refining industry is facing a serious challenge of meeting the increasing demand for gas oil along with the required stringent specifications. This challenge shall become more serious in future due to the necessity of processing the sour and heavy crudes due to their increased availability. Hydrodesulfurization is the most developed and commonly used process in refineries for removing sulfur aromatic compounds from gas oil.

The cost of clean gas oil production from sour crude using conventional hydrotreatment method is increasing and has become drastically high due to the requirement of severe operating conditions, more amount of hydrogen, noble metal based very active and expensive catalysts and huge revamp and capital depreciation cost. Refiners are looking forward to adopt a methodology to reduce the production cost of clean diesel to address the challenge of reduced profit margin in refining. Selective solvent extraction appears to be an economic solution to modify high straight run gas oil (SRGO) by removing maximum amount of refractory sulfur prior to further treating it by conventional hydrotreatment method thereby reducing the sulfur removal load and the cost of hydrotreatment or other alternative methods.

There are many methods for the removal of sulfur and aromatic compounds, such as extraction, oxidative-extraction, adsorption and oxidative-adsorption, which can be used either individually or as complementary with hydrotreatment to produce ultra clean gas oil from straight run gas oil economically. Desulfurization using ionic liquids is widely being researched all over the world.^8,9

Extraction with selective solvent is well proven and widely adapted process in modern refining industries for production/removal of aromatic hydrocarbons from various hydrocarbon streams.^5,10 There are some studies which reported the removal of sulfur and heteroatomic sulphur, and nitrogen compounds from boiling range hydrocarbon stream of gas oil using the selective solvent extraction process.^11–18 These studies evaluated the performance of various solvents for sulfur and aromatics removal from gas oil using single stage and multiple stage batch extraction systems. However, to our knowledge there are no reported studies on continuous counter-current extraction column for extractive desulfurization. Similarly, application of antisolvent like water in sulfur extraction process is reported in very few studies.^14,15 Extraction process is a trade-off between the degree of sulfur removal (D_sr) and yield of extracted SRGO (ESRGO). No study is reported in the literature to the best of the authors' knowledge for the study of the performance of an extraction process in terms of a factor which combines both D_sr and yield.

Considering the above discussions, the present study includes the evaluation of the performance of industrially proven and viable solvents for the removal of sulfur compounds from SRGO containing high sulfur content (1.3 wt%). The effect of extraction temperature, solvent to feed ratio, antisolvent concentration and number of stages (during batch operation) on the degree of sulfur and aromatics removal and yield were evaluated for the batch and continuous counter-current extraction systems. A strategy on reutilization of the extract from industrial point of view has also been suggested.

Theory

Liquid–liquid solvent extraction is based on the principle of difference in the solubility of solute compounds in a solvent. The degree of solubility of solutes in solvent depends upon their chemical nature. Solvent extraction process involves the removal of impurities via scrubbing the hydrocarbon stream by solvent and recovery of the solvent from the scrubbed impurities for its reuse. The major challenge in the solvent extraction is to tackle the problem of the loss of desired hydrocarbons with the removed impurities. This loss depends on the capacity and selectivity of the solvent which can be adjusted by addition of co/antisolvent and changing the extraction temperature.

A number of solvents have been reported for the extraction of sulfur and aromatic compounds from gas oil.^11–18 However, from techno-economic point of view, the selection of the solvent depends on physico-chemical characteristics of the solvent and feedstock, e.g. boiling point/boiling range, density, viscosity, melting point, miscibility, the capacity and selectivity of solvent. The desirable features for extraction solvent have been summarized in literature.^19–24

The following physical properties can be used as preliminary tools to screen the solvents: (1) sufficient difference in the density between solvent and feed for allowing two phase formation and avoiding the flooding condition during extraction; (2) sufficient boiling point difference between solvent and feed to facilitate easy recovery of solvent for its reuse; (3) high thermal and chemical stability of solvent to avoid loss due to degradation; (4) low melting point to avoid steam tracing; (5) low viscosity for the high rate of mass transfer; (6) no zoetrope formation with components in feed to facilitate easy recovery; (7) non-toxicity for safe operation; and (8) noncorrosive to reduce capital investment. Moreover, the selected solvent should further be evaluated for their high capacity for solutes to reduce the required solvent to feed (S/F) ratio and high selectivity to reduce the height of extractor and improve the quality of extract, and to increase the yield of raffinate. The capacity and selectivity of solvent can be adjusted by changing the quantity of co- and anti-solvent in the main solvent and the extraction temperature.

The usefulness of a solvent in liquid–liquid extraction can be represented by the extraction factor for sulfur, sulfur distribution coefficient, yield of extracted gas oil, degree of sulfur removal, and performance factor.

Extraction factor is used to represent the capacity of a solvent. Extraction factor (ε_s) for single stage solvent extraction is defined as.²⁵

The distribution coefficient of solute (K_s) is defined as the ratio of composition for solute in the extract phase to that in the raffinate phase and is represented as follows:

K_s = y_s/x_s (2)

where y_s and x_s denote the concentration of sulfur (g/g) in the extract and in the raffinate phase, respectively. Yield of extracted straight run gas oil (ESRGO) is defined as follows:

The capacity of a solvent is a measure of its ability to dissolve the hydrocarbon. Considering this volumetric yield of ESRGO can also be used to represent the capacity of solvent.

Material and component balance equations which are required to estimate the unknown value of variable in raffinate/extract phase are defined as follows:

F = R + E (4)

x_f,iF = x_r,iR + x_e,iE (5)

where F, R and E are the masses of feed, raffinate and extract, respectively. x_f,i, x_r,i, and x_e,i are the respective mass fractions of component in feed, raffinate and extract, respectively.

Degree of sulfur removal (D_sr) used in this study was estimated using the following expression:

where S_SRGO and S_ESRGO denote the concentration of sulfur in the SRGO and in ESRGO, respectively. Moreover, the degree of aromatics removal can also be used to understand the effect of solvent extraction on the quality of the extracted gas oil obtained under different operating conditions. Degree of aromatic removal (Dar) was calculated by the following expression:where A_SRGO,i and A_ESRGO,i denote the concentration of aromatics (mono, di and poly) in the SRGO and in ESRGO, respectively.

Since capacity and selectivity of solvent show opposite trends, it can be understood that the yield of ESRGO would decrease with the increase in the degree of removal of sulfur and aromatic compounds. However, from economic point of view of a process, it is desirable to obtain the maximum yield of ESRGO with maximum removal of sulfur and poly aromatic compounds. To combine the effect of these two important parameters in a single factor, performance factor (P_f) of solvent, first time used in this study, is defined as follows:

P_f,α = αD_sr + (1 − α) × Yield(%) (8)

where α denotes the weight factor assigned to the degree of sulfur removal.

Experimental

Materials

Straight run gas oil (SRGO) was obtained from an Indian refinery. Physicochemical properties of SRGO are given in Table 1. Acetonitrile (AcN: 99.5%+, MERCK), N,N-dimethylformamide (DMF: 99.5%, MERCK), furfural (FF: 98%: SD Fines), N,N-dimethylacetamide (DMA: 99.5%+, MERCK) and dimethyl sulfoxide (DMSO: 99.8%, MERCK) were used as extraction solvents. All compounds mentioned above were used without any pretreatment except furfural. Furfural was distilled prior to use as a solvent.

Table 1 Physico-chemical properties of straight run gas oil (SRGO)

Parameter	Value
Total sulfur (wt%)	1.3
Non aromatics	66.8
Mono-aromatics	18.8
Di-aromatics	8.2
Poly-aromatics	6.2
Refractive index nD20	1.4762
Density at 20 °C (kg m^−3)	853.24
Kinematic viscosity at 70 °C (cSt)	2.17
Kinematic viscosity at 100 °C (cSt)	1.44

---

ASTM D-86
Volume%	Temperature, °C
IBP	222.1
5	244.5
10	251.2
20	259.5
30	266.8
40	276.1
50	287.2
60	300.3
70	315
80	332.1
90	351.8
95	369.7
FBP	380.9
Distillate	97.0
Residue	2.5
Lighter	0.5

---

Methods of analysis

Density was determined using DE45 densitometer manufactured from Metller Toledo, Japan at temperature of 20 °C. Refractive index was determined using Abbe Refractometer RE45 at 20 °C. Total sulfur content of SRGO and ESRGO were measured by X-ray fluorescence (XRF) method by using ASOMA ED XRF analyzer Spectro Phoenix II make. ASTM D86 method was used for determining the boiling range of gas oil. Ultraviolet (UV) spectrophotometry was used for the estimation of mono, di and poly aromatics content of SRGO and ESRGO.

Apparatus and procedure

Single stage equilibrium extraction. Known amounts of gas oil and solvent were added in a jacketed glass mixer settler provided with a stirrer. The extraction temperature was maintained within ±0.5 °C with the help of a thermostatic bath. The mixture was stirred for 30 min which was sufficient for establishing equilibrium. After mixing, settling time of 30 min was provided to separate the oil rich raffinate phase from the solvent rich extract phase. Due to the considerable solubility of solvent in hydrocarbon, the raffinate phase was washed with water to remove the solvent. The solvent-free raffinate was used for the calculation of the yield (as defined earlier). The moisture of the solvent-free raffinate was removed using solid ammonium sulphate. Thereafter, it was analyzed for sulfur and aromatics concentration.

Continuous counter-current extraction. Continuous counter-current extraction of gas oil was carried out in a jacketed pyrex glass column of 10 mm internal diameter. The column was filled up to 140 mm of its height with cannon packing of 2.3–3.0 mm. Settling zones of 15 mm were provided at the top and bottom of the column. The feed and solvent were pumped using metering pumps at the bottom and the top of the column, respectively. The flow rates of feed and solvent were fixed so as to get the desired solvent-to-feed ratio. During extraction runs, feed was used as dispersed phase, the interface was observed at the top of the column due to the feed being lighter. The 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 hot water through the jacket of the column. The steady state of the column was confirmed by the constant value of RI measured for hydrocarbon samples at the top from time to time before collecting the sample for analysis. The gas oil raffinate phase and solvent rich extract phase were obtained from the top and bottom of the column, respectively. The raffinate phase was further treated in the same way as in the single stage equilibrium experiments.

Results and discussion

Batch equilibrium extraction

Evaluation of solvents with SRGO. The extraction of actual SRGO was carried out using industrially viable polar solvents such as AcN, DMF, furfural, DMA and DMSO with the volumetric gas oil to solvent ratio of 1 at 45 °C using the procedure described in the earlier section. The yields of ESRGO and the degree of sulfur removal (D_sr) were estimated using eqn (3) and (6), respectively. The refractive index (RI) value of aromatic compounds is higher than that of the paraffinic materials. The lower value of RI of ESRGO indicates the higher removal of aromatic compounds.

The results (Table 2) clearly indicate that the sulfur removal and yield of ESRGO using solvent extraction strongly depends on the type of solvent used. Percent sulfur removal using DMF, DMA and furfural solvents are much higher than AcN and DMSO. Among DMF, DMA and furfural, DMA removes maximum sulfur. However, yield values for DMF, DMA and furfural solvents are lower than AcN and DMSO. This indicates that there is a trade-off between sulfur removal and yield value. From the economic point of view of the process, maximum sulfur removal with maximum yield value is desirable.

Table 2 Straight run gas oil extraction with selected solvents at 45 °C, S/F = 1.0 and α = 0.5

	ACN	DMF	DMA	Furfural	DMSO
Raffinate properties
Refractive index@20 °C	1.469	1.4635	1.4623	1.4657	1.4701
Density@20 °C, g ml^−1	0.84304	0.83521	0.83280	0.83468	0.84255
Sulfur in ESRGO (%)	1.18	0.81	0.76	0.8	1.03
Calculate responses
Gas oil yield %	87.5	81	72.5	82.5	88.5
Extraction factor	0.26	0.98	1.36	0.97	0.43
Distribution coefficient	0.22	0.62	0.73	0.53	0.28
Sulfur removal (%)	9.2	37.7	41.5	38.5	20.8
Performance factor	48.4	59.3	57	60.5	54.6

---

Considering this, the performance factor (P_f) for each solvent was estimated using eqn (8) and weight factor (α) = 0.5, thus giving equal importance to the yield and D_sr. The values of P_f are given in Table 3. It is observed that when the sulfur removal and yield were assigned the same value of α, values of P_f follow the order: furfural > DMF > DMA > DMSO > AcN.

The extraction factor and distribution coefficient follow the order: DMA > DMF > furfural > DMSO > AcN. Although the values of extraction factor for DMF and furfural are comparable, there is significant difference in their distribution coefficients. It may be attributed to the difference in the density of the solvents.

The effect of the value of α (in the range of 0.3–0.9) on the P_f value for each solvent is shown in Fig. 1. It is clear that P_f values for DMF and furfural solvents are very close to each other and they decrease over the whole range of α. For AcN and DMSO, the values of P_f are always much less as compared to other solvents at all values of α. P_f values of DMF are higher than that of DMA for α < 0.7, however, for α ≥ 0.7 P_f values of DMA become higher than that of DMF. P_f values for furfural are slightly higher than that of DMF, however, its value becomes lower than that of DMA for α ≥ 0.8. Therefore, DMF, furfural and DMA appear to be comparable solvents in terms of P_f for the whole range of α.

Fig. 1 Effect of weight (α) on solvent's performance factor (P_f,α) during the extractive desulfurization of SRGO by various solvents.

However, considering the low oxidative and thermal stability of furfural as reported in literature,⁵ although DMF and DMA have lower values of extraction factor and distribution coefficient, they were selected as solvents for further study.

Effect of extraction temperature and water concentration. It is well known that extraction temperature affects the capacity and selectivity of any solvent. The results for experiments with varying temperatures (with no water in solvent) have been summarized in Fig. 2. It may be noted that during the extraction of SRGO with DMA as solvent, a single phase formation was observed when the temperature was increased to 75 °C, therefore, experiments were carried out in the range of 45 to 65 °C for DMA and in the range of 45 to 75 °C for DMF. It is clear from Fig. 2 that an increase in the extraction temperature increases the degree of sulfur removal, however, it simultaneously decreases the ESRGO yield. Therefore, P_f,0.5 appears to be unaffected by the variation of temperature. However, P_f,0.7 increases with increasing temperature while P_f,0.3 decreases with increasing temperature.

Fig. 2 Effect of extraction temperature on volume yield%, degree of sulfur removal (D_sr) and solvent's performance factor (P_f,α) for solvent to feed ratio of 1.

Water behaves as an antisolvent during extraction process as it decreases the solubility of hydrocarbons in solvents. Concentration of antisolvent in the main solvent can change the performance of solvent.¹⁵ Considering this, extraction of SRGO with DMF and DMA solvents was carried out by varying the water content in the range of 0–3.0 volume% at 65 °C to understand the impact of the amount of water on D_sr and P_f of DMF and DMA. Results are shown in Fig. 3. As expected, an increase in water amount in the solvent, increased the ESRGO yield, however, it decreased the degree of sulfur removal for both the solvents. However, the extent of change in P_f,0.5 for DMA solvent was slightly higher than that of DMA with increasing water amount in solvent.

Fig. 3 Effect of anti-solvent (water) concentration on volume yield%, degree of sulfur removal (D_sr) and performance factor (P_f,α) for solvent to feed ratio = 1.0 and Temperature = 65 °C.

The P_f values of DMA solvent are lower than that of DMF over the whole range of studied temperature and water amount. It can be inferred that when the degree of sulfur removal is more important than the yield, it is better to do the extraction at higher temperature and lower water amount and vice versa.

It should be noted that P_f values for DMF are higher than the DMA values for α ≤ 0.7. Considering the significance of ESRGO yield and the degree of sulfur removal in the economic point of view of the extraction process, DMF can be considered as more efficient than DMA. Moreover, the boiling points of DMF and DMA are 153 and 165 °C, respectively. Therefore, it is easier to recover DMF from the extract using distillation. In brief, considering all the above points, DMF can be considered as a better solvent than DMA, therefore, it was selected for further studies.

Effect of solvent to feed ratio (S/F). In this study, SRGO was extracted with DMF at 65 °C with a volumetric solvent to feed ratio of 1.0, 1.5, 2.0, and 3.0 using single stage equilibrium extractor. The results are shown in Fig. 4. It is observed that sulfur removal increased with increase in the solvent to feed ratio, while the ESRGO yield decreased. P_f,0.5 values are nearly constant up to solvent to feed ratio of 2.0, however, it decreases with an increase in solvent to feed ratio beyond 2. It is important to note that energy and capital requirement of the extraction process also increase with increase in the solvent to feed ratio. Hence, the solvent to feed ratio of 1.0 appears to be better for continuous extraction.

Fig. 4 Effect of solvent to feed ratio on volume yield%, the degree of sulfur removal (D_sr) and solvent's performance factor (P_f,α) for batch extraction using DMF as solvent for α = 0.5 and temperature = 65 °C.

Multistage stage extraction. To understand the importance of number of equilibrium stages required to meet the desired amount of sulfur removal, multistage stage solvent extraction of SRGO was carried out with DMF at volumetric S/F ratio of 1.0 and an extraction temperature of 65 °C. ESRGO from 1^st stage was used as feed for 2^nd stage solvent extraction and that from the second stage was used as feed in the 3^rd stage solvent extraction. The volume of ESRGO obtained in each stage was used to estimate the yield value. Cumulative yield was also estimated for second and third stages. The sample of the extracted oil in each stage was analyzed for sulfur content. The results obtained for all the three stages of extractions are given in Table 3. It is observed that sulfur removal decreases appreciably in the 2^nd and 3^rd stages, whereas stage wise volumetric yield increases slightly in subsequent extraction stages. However, the cumulative sulfur removal rapidly increases at the cost of significant loss in cumulative yield. Cumulative sulfur removal and yield were estimated using the amount of sulfur and the volume of SRGO. This implies that it is possible to increase the degree of sulfur removal at the expense of the lower yield value of ESRGO using multistage solvent extraction. The performance factor (depending upon the value of α) of subsequent stage was lower than the previous one. It might be possible to reduce the feed sulfur from 1.3% to 0.36% using three equilibrium stages which is equivalent to 72.3% of sulfur removal.

Table 3 Multistage solvent extraction with DMF

Parameter	1st stage	2nd stage	3rd stage
Stage wise volume yield (%)	75.6	76.5	78.1
Cumulative volume yield (%)	75.6	57.8	45.2
Sulfur in ESRGO	0.74	0.53	0.36
Dsr (%)	43.1	28.4	32.1
Cumulative Dsr (%)	43.1	59.2	72.3
Performance factor (Pf)	59.3	52.4	55.1

---

Continuous counter current extraction

ESRGO product. Continuous counter current extraction of SRGO with DMF as described in experimental section was carried out under different operating conditions by varying the temperature from 55 °C to 45 °C and water amount in solvent from 0.0 to 5.0%. Four cases were formed by selecting the two values for each temperature and water amount in solvent. The analysis of ESRGO obtained for these four cases, a batch extraction case and SRGO has been summarized in Table 4. The code of different runs along with experimental conditions is also given in Table 4. Sulfur, di-aromatics and poly-aromatics are major impurities in gas oil which should be removed to produce a clean gas oil. Therefore, degrees of removal of these parameters are shown in Fig. 5.

Table 4 Analysis of feed and products obtained during batch and continuous extraction with DMF

Parameters	SRGO	B-ESRGO-55T-0W	C-SRGO-55T-0W	C-ESRGO-45T-0W	C-ESRGO-45T-3W	C-ESRGO-45T-5W
Experimental Conditions
Batch or continuous		Batch	Continuous	Continuous	Continuous	Continuous
Extraction temp. (°C)	—	55	55	45	45	45
Water content (vol%)	—	0	0	0	3	5
Experimental results
Density at 20 °C (g ml^−1)	853.24	836.60	824.10	829.03	829.80	833.95
Volume yield (%)	—	79.1	74.3	78.1	84.5	86.7
Sulfur (wt%)	1.3	0.79	0.37	0.46	0.65	0.78
Non-aromatics (wt%)	71.3	80.9	89.4	88.4	85.6	83.5
Mono-aromatics (wt%)	16.8	12.5	8.7	8.9	9.6	10.5
Di-aromatics (wt%)	7.2	4.4	1.3	1.9	3.1	3.9
Poly-aromatics (wt%)	4.7	2.2	0.6	0.8	1.7	2.1
Performance factor (Pf)		31.0	53.2	50.5	42.3	34.7

---

Fig. 5 Comparative analysis of degree of sulfur, di-aromatics and poly-aromatics removal during batch and continuous extraction using DMF as solvent under various experimental conditions.

It may be seen from Table 4 and Fig. 5 that the percent removal of these impurities is significantly greater by continuous extraction than batch extraction. It is attributed to the availability of more than one equilibrium stage in the column and increased concentration gradient across the counter current extraction column. The degree of removal of these impurities and the yield of ESRGO increased with increase in temperature and decrease in water content in solvent. The removal of undesired compounds followed the order: poly-aromatics > di-aromatic > sulfur > mono-aromatics. Removal of these compounds (which have very low cetane number) would significantly increase the cetane value of ESRGO.

It will also facilitate the easier deep desulfurization of gas oil in hydrotreater under less severe operating conditions due to the removal of refractory sulfur compounds and poly-aromatics responsible for slowing down the hydrotreatment reaction.²⁶

Results suggest that the maximum extraction temperature and zero percent water concentration are desired for the maximum removal of these impurities. However, the yield of valuable ESRGO decreases with an increase in temperature and a decreasing water content.

P_f values for each case were estimated with different α values to understand the overall impact of extraction temperature and water amount on solvent extraction process. The results are shown in Fig. 6. It may be seen that the P_f,α value is significantly higher for the continuous extraction process in comparison to the batch extraction, irrespective of the value of α. For C-ESRGO-55T-0W case (extraction temperature = 55 and zero water content), the difference in P_f values is the least because the sulfur removal and yield number are close to each other. It may be clearly seen that the decrease in the extraction temperature spreads the P_f values i.e. there is large variation in P_f values for various α values (P_f,0.3 = 68.7 to P_f,0.7 = 74.1). Whereas, an increase in water content increases the spread of P_f values. For C-ESRGO-45T-3W case (extraction temperature = 45 and water content = 3 vol%), P_f,0.3 and P_f,0.7 values were 60.4 and 74.2, respectively. Further increase in water content for C-ESRGO-45T-5W case, increased the spread of the P_f values (P_f,0.3 = 54.0 to P_f,0.7 = 72.7). Overall for α = 0.7, there is marginal effect because of the variation of extraction temperature, however for α = 0.5, P_f values decreases with an increase in water content and decrease in temperature.

Fig. 6 Effect of weight factor (α) on solvent's performance factor (P_f,α) during batch and continuous extraction using DMF as solvent under various experimental conditions.

Utilization of extract. Since percent removal of non-aromatics, mono-aromatics, di-aromatic and poly-aromatics, yield and the density of raffinate were known, therefore, the composition of the extract phase for all cases was estimated using mass and component balance; the results are shown in Fig. 7. The concentration of desirable non-aromatic compounds in ESRGO decreases in the extract with decrease in temperature and increase in water amount in the solvent. Considering the composition of extract, it should be taken into account that the extract should not be considered as a waste because this stream can be used either as a co-product, such as carbon black feed stock (CBFS), rubber processing oil, fuel oil blending stream, or as a high quality feed stock for secondary conversion process units, such as fluid catalytic cracker unit (FCCU) and delayed coking unit (DCU) to convert it to light and middle distillates.

Fig. 7 Composition of extract phase obtained during batch and continuous extraction using DMF as solvent under various experimental conditions.

The FCC and delayed coker are processes where carbon is rejected to meet the requirement of hydrogen in distillate products. The metal content, Conradson carbon residue (CCR) and viscosity in extract stream obtained after continuous extraction would be very low and hydrogen to carbon (C/H) ratio would be significantly higher in comparison to vacuum residue (VR), thermal tar, lube extracts, pyrolysis tar and pitch streams which are used as feed in DCU. Therefore, the blending of extract stream with VR and pitch stream will not only increase the yield of distillates, which is inversely dependent on the CCR and proportionally dependent on C/H, but will also improve the quality of coke.²⁷

Considering its utilization as a CBFS, bureau of mines correlation index (BMCI) value, which is indication of quality of black carbon feed stock for extract products, was estimated using the following correlation:²⁸

BMCI = 473.7S_g − 456.8 + (48 460/T_b) (9)

where, S_g is liquid specific gravity at 60 °F and T_b represents the average boiling point (K). Average boiling point is the arithmetic average of temperatures at 10% interval from 20 to 80%. Since the extract is obtained from gas oil stream, its average boiling point will be close to that of the gas oil. Therefore, the distillation data of gas oil was used to represent the average boiling point of the extract stream. S_g was obtained by converting the density of extract stream from 20 °C to that at 15.5 °C and then density at 15.5 °C to the specific gravity using petroleum measurement tables.²⁹

The values of estimated BMCI values along with the estimated S_g of extract streams are tabulated in Table 5. Results indicate that BMCI value increases with decrease in extraction temperature and increase in water content in solvent as the selectivity of solvent for aromatics increases with respect to increase in paraffin compounds (Fig. 7). As we know, higher the BMCI, better the quality of CBFS; the solvent extraction should be carried out with a solvent containing significant amount of water. However, the increasing trend of the density of extract stream with decrease in temperature and water content also suggests decrease in the C/H ratio of extract. Therefore, there is a possibility of adjusting the operating conditions of the extraction unit considering the requirement of further downstream operation to be used for raffinate and extract stream processing. For example, temperature and S/F can be adjusted to the higher side with zero percent water in solvent to maximize the recovery of sulfur compounds to debottleneck the hydrotreatment unit to be used for raffinate processing in order to bring the sulfur to ppm level. While, the water content can be increased to increase the aromatic concentration in extract streams so that it can be used as a CBFS feed-stock and to increase the yield of ESRGO. Therefore, operating conditions of extraction is to be adjusted depending on the further considered process/application for extract and raffinate streams. It may be noted that the BMCI values obtained for the extracts are in the range of CBFS, which are being marketed by various refineries in India.^29,30

Table 5 Properties and BMCI of extract streams

Parameter	B-ESRGO-55-0W	C-SRGO-55-0W	C-ESRGO-45-0W	C-ESRGO-45-3W	C-ESRGO-45-5W
Density @ 20 °C	0.9162	0.9375	0.9396	0.9810	0.9853
Density @ 15.5 °C	0.9194	0.9407	0.9428	0.9842	0.9885
Specific gravity @ 15.5	0.9199	0.9412	0.9433	0.9848	0.9891
Average boiling point, °C	291.00	291.00	291.00	291.00	291.00
BMCI	65.2	75.3	76.3	95.9	98.0

---

Conclusions

Based on the results presented in the study, N,N-dimethylformamide (DMF) solvent appears to be the best solvent for selective solvent extraction among the five most widely used solvents in hydrocarbon industries. The degree of sulfur removal greatly depends on the nature of the solvent and operating conditions used during the extraction. Water amount in the solvent significantly changes the values of ESRGO yield and removal of impurities. Extraction temperature and water content in solvent give the flexibility to adjust the yield and the degree of removal of impurities to maximize the benefit under a given situation. Continuous counter current extraction is considerably more effective than the single stage extraction. 71.5% sulfur can be removed from SRGO using continuous counter current extraction at reasonable ESRGO yield. The selection of weight factor for sulfur removal and yield affects the performance factor of extraction process and needs the utmost care in its value selection. There is a great possibility of utilization of the extract as carbon black feed stock (CBFS) as shown by comparison of the calculated bureau of mines correlation index (BMCI) values which is similar to those CBFS which are already being sold in the market.

References

1. V. C. Srivastava, An evaluation of desulfurization technologies for sulfur removal from liquid fuels, RSC Adv., 2012, 2, 759–783.
2. T. A. Koch, K. R. Krause, L. E. Manzer, M. Mehdizadeh, J. M. Odom and S. K. Sengupta, Environmental challenges facing the chemical-industry, New J. Chem., 1996, 20(2), 163–173.
3. S. Kumar, V. C. Srivastava and R. P. Badoni, Oxidative desulfurization by chromium promoted sulfated zirconia, Fuel Process. Technol., 2012, 93, 18–25.
4. S. Kumar, V. C. Srivastava and R. P. Badoni, Oxidative desulfurization of dibenzothiophene by zirconia-based catalysts, Int. J. Chem. React. Eng., 2014, 12, 1–8.
5. M. Sharma, P. Sharma and J. N. Kim, Solvent extraction of aromatic components from petroleum derived fuels: a perspective review, RSC Adv., 2013, 3, 10103–10126.
6. European emission standards, Wikipedia, http://en.wikipedia.org/wiki/European_emission_standards, accessed 22 March 2012.
7. Clean fuel technology, http://www2.dupont.com/Clean_Technologies/en_US/assets/downloads/Hydrocarbon_Engineering_article_June_2009.pdf.
8. R. Anantharaj and T. Banerjee, Aromatic sulfur-nitrogen extraction using ionic liquids: experiments and predictions using an a priori model, AIChE J., 2013, 59(12), 4806–4815.
9. J. Gao, H. Meng, Y. Lu, H. Zhang and C. Li, A carbonium pseudo ionic liquid with excellent extractive desulfurization performance., AIChE J., 2013, 59(3), 948–958.
10. Y. Shiraishi, K. Tachibana, T. Hirai and I. Komasawa, Desulfurization and denitrogenation process for light oils based on chemical oxidation followed by liquid-liquid extraction., Ind. Eng. Chem. Res., 2002, 41, 3662–4375.
11. P. Petkov, J. Tasheva and D. Stratiev, Extraction approach for desulphurization and dearomatization of middle distillates., Pet. Coal, 2004, 46, 13–18.
12. A. A. Gaile, Development and improvement of extraction processes for separation and purification of petroleum products., Russ. J. Appl. Chem., 2008, 81, 1311–1324.
13. A. A. Gaile, V. E. Somov, G. D. Zalishchevskii, E. A. Kaifadzhyan and L. L. Koldobskaya, Extractive refining of atmospheric gas oil with N-methylpyrrolidone, Russ. J. Appl. Chem., 2006, 79, 590–595.
14. A. A. Gaile, B. M. Saifidinov, V. V. Kolesov and L. L. Koldobskaya, Extractive refining of high-sulfur diesel fraction to remove organic sulfur compounds and aromatic hydrocarbons, Russ. J. Appl. Chem., 2010, 83(3), 464–472.
15. A. A. Gaile, B. M. Saifidinov, V. V. Kolesov and L. L. Koldobskaya, Multistep countercurrent extraction of organic sulfur compounds and arenes from the high-sulfur diesel fraction, Russ. J. Appl. Chem., 2010, 83(3), 473–476.
16. V. Toteva, L. Topalova and P. Manolova, Extractive dearomatization and desulphurization of a distillate gasoil cut with imethylformamide, J. Univ. Chem. Technol. Metall., 2007, 42, 17–20.
17. D. M. A. El-Aty, O. I. S. El.-Din, S. I. Hassa, S. M. Tawfik and S. Hanafi, Evaluation of some organic solvents for refining diesel fuel fraction, Pet. Sci. Technol., 2009, 27, 861–863.
18. S. I. Hassan, O. I. S. El-Din and S. M. Tawfik, Solvent refining of straight run diesel fuel by various solvents phase equilibrium, J. Appl. Sci. Res., 2009, 5(5), 515–521.
19. B. S. Rawat, S. K. Ghosh and I. B. Gulati, Basic considerations in the selection of solvents for aromatics extraction, Pet. Hydrocarbons, 1972, 6, 203–210.
20. B. S. Rawat and I. B. Gulati, Solvent for aromatics extraction and criteria for selection, J. Sci. Ind. Res., 1976, 35(6), 383.
21. M. J. Hampe, Selection of solvents in liquid-liquid extraction according to physicochemical aspects, Chem. Eng. Technol., 1985, 57(8), 669–681.
22. R. W. Cusack, P. Fremeaux and Y. N. V. Otto, A fresh look at liquid-liquid extraction, part 1: Extraction systems, Chem. Eng., 1991, 98(2), 66–76.
23. R. W. Cusack and P. Fremeaux, A fresh look at liquid-liquid extraction, part 2: Inside the extractor, Chem. Eng., 1991, 98(3), 132–138.
24. S. K. Ghosh, Solvents for petroleum industry, Hydrocarb Technol., 1993, 26, 13–19.
25. L. Alders, Liquid-Liquid Extraction–Theory and Laboratory Practice, Elsevier Publishing, 2nd edn, 1959.
26. Z. Ismagilov, S. Yashnik, M. Kerzhentsev, V. Parmon, A. Bourane, F. M. Al-Shahrani, A. A. Hajji and O. R. Koseoglu, Oxidative desulfurization of hydrocarbon fuels, Catal. Rev.: Sci. Eng., 2011, 53, 199–255.
27. HPI Consultants Inc, Petroleum Refining Process correlations, 1996.
28. Encyclopedia of Chemical. Processing and Design, ed. J. J. Mektta and W. A. Cunningham, Marcel Dekker, Inc., New York, 1990, vol. 6, p. 200.
29. CBFS specification, IOCL (Indian Oil Corporation Limited), http://www.iocl.com/Products/CarbonBlackFeedStockSpecifications.pdf, accessed 11 March 2014.
30. CBFS specification, RIL (Reliance Industries Limited), http://www.ril.com/downloads/pdf/carbonstocksspecification.pdf, accessed 11 March 2014.

---