Optimal HDS for lower-sulfur gasoline depends on several factors

Tek Sutikno
Fluor Daniel Williams Bros. Inc.
Tulsa

• Hydrotreating equations [67,802 bytes]

Optimal selection of a gasoline desulfurization technology for a particular refinery depends on the characteristics of sulfur compounds in gasoline, kinetic parameters of hydrodesulfurization (HDS) reactions, catalysts, and processing options.

The overall refining scheme and marketing objectives are also important factors.

Proposed U.S. Environmental Protection Agency (EPA) regulations to further reduce allowable sulfur contents in gasoline make consideration of deep desulfurization necessary. Deep desulfurization refers to the further reduction of sulfur in gasoline to comply with forthcoming EPA regulations.

A number of U.S. domestic refineries will need to modify their refining operations to comply with the new sulfur regulations as well as new requirements for gasoline vapor pressure and benzene content. Proper applications of technologies will minimize capital costs and loss of product value, such as gasoline octane.

The specifications for reduced gasoline sulfur have been issued for Canada (30 ppm for 2005) and California (40 ppm starting in 1996).

On May 1, 1999, the EPA proposed regulations limiting the gasoline sulfur level to an average of 30 ppm, to be applicable in 2004-2006. The maximum sulfur content cannot exceed 80 ppm in 2005 (OGJ, May 10, 1999, p. 32).

Currently, the average sulfur content for U.S. gasoline outside of California ranges from 150 ppm (Phase II RFG) to 340 ppm, with a maximum of 1,000 ppm.

Most U.S. refineries marketing gasoline outside California will need to add a new desulfurization scheme. Some of the refineries already equipped with hydroprocessing units will need to modify their operating conditions to reach the new sulfur target.

Other options include refining only sweet or low-sulfur crudes and reducing gasoline production. Although these options may comply with regulations, they may not be economical.

Characteristics of gasoline sulfur

Reduction of gasoline sulfur primarily requires removal of fluid-catalytic cracking (FCC) gasoline sulfur or sulfur in the FCC feed. Typically, over 90% of sulfur compounds in the refinery-gasoline pool comes from the FCC unit (FCCU).

In a refinery that produces a substantial amount of gasoline, FCC gasoline makes up about 40% of the overall refinery gasoline pool. The remaining gasoline is typically derived from straight-run naphtha, coking, hydrocracking, and other molecular conversion units such as alkylation and reforming operations.

Feeds to FCCUs are gas oils from crude and vacuum units. The sulfur contents of untreated FCC feeds can be as high as 2.0 wt % or even higher for refineries processing very sour crudes.

Fig. 1 [102,856 bytes] shows about 4.5 wt % sulfur in the 690° F.+ fraction of a sour crude with a total sulfur content of 2.78%.1 A number of studies have been made to evaluate the correlations between FCC feed sulfur and that of cat gasoline. One example is Equation 1 in the accompanying box.²

As a rule of thumb, the sulfur content of cat gasoline is about one-tenth of that in the FCC feed. For example, an FCC feed with 20,000 ppm sulfur will yield about 2,000 ppm sulfur cat gasoline.

Cat gasoline produced from hydrotreated FCC feeds, however, contains about one-twentieth of the sulfur in the hydrotreated feed. For example, if the 20,000 ppm FCC feed is first hydrotreated to reduce the sulfur content to 2,000 ppm, the sulfur content of the cat gasoline is about 100 ppm.

Higher FCC-reactor temperatures increase the sulfur content of cat gasoline, but this content decreases at higher conversions. Typically, about 40-55 wt % of the FCC feed is converted to cat gasoline.^{3 4} These general rules are useful when actual sulfur-distribution data for refining a particular crude are not available.

In cat gasoline, approximately 60% of sulfur compounds are thiophenes and their alkyl derivatives. The remaining 40% are mercaptans and sulfides (alkyl, naphthenic, and aromatic).

Relative to these cat gasoline sulfur compounds, sulfur in the FCC feed streams is in the forms of heavier sulfur compounds-mainly alkyl benzothiophenes and alkyl dibenzothiophenes (DBT). These compounds reportedly show the following relative desulfurization reactivities, represented by k, the effective rate constant.¹

^ksulfide ^kmercaptans ^kthiopheneskbenzothiophenes ^kDBT

For cat gasoline, the sulfur content typically does not distribute linearly with the boiling points in the cut. A large fraction of the total sulfur is in the heavy fraction.

Lowering the end point of cat gasoline is a potential option to comply with the new sulfur specification. This option may be favorable where demand for kerosine is high and justified where the maximum T-90 (ASTM D-86) of gasoline is restricted, such as in the case of CARB (California Air Resources Board) gasoline.

Sulfides and mercaptans in the low boiling point ranges may be removed by extraction or absorption processes using solvents such as caustic. Desulfurization of high boiling sulfur compounds such as thiophenes and its derivatives, however, requires catalytic hydroprocessing.

The difficulty of sulfur removal increases in the order of paraffins, naphthalenes, and aromatics.³

Reaction parameters

The principal reaction parameters in catalytic hydroprocessing vary depending on the feed characteristics and the catalyst activity, to be discussed later. These parameters are temperature, pressure, reaction time (commonly defined as liquid hourly space velocity, LHSV, the reciprocal of reaction residence time), and order of reaction.

Hydrotreating processes for sulfur removal vary in operating pressure from 290 to 500 psig for naphtha hydrotreaters and 400 to about 1,800 psig for gas oil hydrotreaters. Typical operating temperatures are 525-800° F.

Hydroprocessing of heavy fractions, such as resid compounds, may require pressure as high as 3,000 psig.³ Highly aromatic feedstocks, such as cycle oil from the FCCU, require high hydroprocessing severities.

The target-sulfur content of the outlet stream also affects the required hydroprocessing pressure and temperature. High hydroprocessing temperatures give lower residual sulfurs but also decrease the catalyst's lifetime, resulting in higher operating cost.

Feedstocks with high boiling end points require higher process severities.

Separating the feedstocks to lower the boiling end point not only reduces the inlet sulfur content but improves the percentage of sulfur removal under the same set of reaction conditions.

For example, one study on middle distillate showed that an end point reduction of about 100° F. decreased the inlet sulfur concentration by 50% and reduced the outlet residual sulfur from 0.04% to 0.004%.⁵ In this case, a 100° F. end point reduction gave 10 times lower residual sulfur.

For gasoline from catalytic cracking of an unhydrotreated feed, in another example, lowering the end point of heavy FCC naphtha (250-430° F.) from 430 to 400° F. reduced the total sulfur in this fraction by more than 40%.⁶

The effect of pressure on sulfur removal is quantitatively less obvious.

While the extent of HDS is increased by a higher hydrogen partial pressure, it is interdependent with the total reaction pressure, hydrogen-to-hydrocarbon ratio, and recycle rate. The hydrogen partial pressure required for an HDS reactor defines the total pressure and the hydrogen-recycle rate requirements of the HDS system.

For gas oil or cat-feed hydrotreating, the required hydrogen partial pressure increases as the gas oil becomes heavier. Above a hydrogen partial pressure of 1,500 psig, however, pressure variations exert only a very slight influence on the rate of HDS, and hydrogenation of the aromatic rings of the alkylated DBT compounds becomes significant.⁵

FCC naphtha typically contains olefinic compounds generated from the cracking reaction. In FCC naphtha hydrotreating, these olefinic compounds are hydrogenated in parallel to the HDS reactions.

The extent of olefin saturation reaction varies directly with the required severity of HDS. Olefin saturation not only reduces the octane rating of the FCC gasoline but significantly increases the consumption of hydrogen and the hydrogen partial pressure required for a target residual sulfur.

Moreover, the higher end of the normal pressure range increases the chance of hydrogenating high octane aromatic molecules to lower octane naphthenic rings. This is a negative side effect of reducing the hardest-to-react aromatic sulfur compounds in the last ppm range.

Reaction models

The LHSV required for a target-sulfur removal is normally calculated from an empirical kinetic-reaction model. For deep desulfurization reactions, however, these calculations may not be reliable before the reaction order and the rate constant of the model have been correlated with the actual deep desulfurization data.

Conventional HDS data have been correlated satisfactorily with pseudo first-order reaction kinetics, but the desulfurization kinetics actually involves a sum of HDS rates of different kinds of sulfur compounds with different reactivities.

In a naphtha hydrotreater typically upstream of a catalytic reforming unit, the first-order reaction kinetics is typically recommended to analyze the extent of sulfur removal even with the target-residual sulfur level less than 0.5 ppm.

For hydrotreating of cat gasoline (cracked naphtha) containing a significant percentage of olefins, a pseudo reaction order of 1.5 for desulfurization has been correlated from a pilot study.⁷

Deep desulfurization of middle distillates and heavier fractions such as gas oil (cat feed) and residuum involves reactions with less-reactive sulfur compounds. The pseudo (or apparent) order of reaction rate has been reported to increase from first order to within the range of 1.7 to 2.0 and as high as 3.0.^{5 8}

For deep desulfurization, a higher reaction order means a larger incremental reactor size to achieve a marginal decrease in the target-residual sulfur. The accuracy of reaction orders may be somewhat limited, however, as they are normally derived by empirically correlating a limited set of sulfur-removal data in a given set of operating conditions.

Consequently, these simplified models with the correlated reaction orders may not apply to deep desulfurization calculations and need to be verified against the relevant deep desulfurization test data.

Equations 2 and 3 show the commonly proposed reaction kinetic model for hydrotreating. When the values of k_eff and the associated order of reaction have been derived properly for the intended range of operating conditions, Equations 2 and 3 can be used to determine the required changes in the reaction parameters for achieving a new residual sulfur target.

For example, k_eff can be correlated with the reaction temperatures to evaluate its effect on desulfurization. For deep desulfurization with ultra low residual sulfur targets, however, k_eff and the reaction order need to be derived from or verified against the deep desulfurization data.

Equations 2 and 3 have a number of limitations. First, the well-known inhibitory effect of H₂S on the HDS rate is not accounted for. For example, one source reported that the HDS rate decreases by about 5-15% for each mole percentage of H₂S in the reaction mixture.⁵ Second, k_eff represents only an average, apparent value typically derived from a set of sulfur-removal data under a given set of operating conditions. As such, the value of k_eff and the apparent order of reaction may vary with inlet and outlet concentrations of sulfur compounds.

Without properly derived values of k_eff and n (the order of reaction) for the intended set of HDS conditions, kinetic models such as Equations 2 or 3 have a very restricted application for deep desulfurization calculations.

The actual reaction-rate constants are dependent upon the types of sulfur compounds in addition to temperature, pressure, and the mass transfer characteristics of the catalyst bed.

Recent published work on these areas is exemplified in References 9, 10, and 11.

Another recent kinetic study included the effect of hydrogen partial pressure and evaluated the relative HDS reactivity for various types of DBT derivatives compounds.¹² This study identified, for example, that in the benzothiophene family, methyl groups in certain positions reduce the hydrogenolysis rate with respect to that of benzothiophene.

Catalysts

New generations of hydrotreating catalysts have better desulfurization activities than that of former generations. It is questionable, however, if advances in catalysts can eliminate the needs for a higher process severity or complexity required for deep desulfurization.

Nevertheless, the selection of catalysts plays a decisive economic role in HDS, and commercially available and performance-guaranteed catalysts should be extensively reviewed to maximize the economic benefits from catalyst development.

Table 1 [65,536 bytes] compares the relative reactivities of common metal elements of commercial catalysts for hydroprocessing. Co-Mo catalyst is ranked the highest in HDS reactivity for normal or conventional desulfurization.

For deep desulfurization involving removals of certain DBT type compounds typically found in gas oil, however, Ni-Mo catalysts are more effective than Co-Mo.^{13 14}

Table 1 also indicates the cracking strengths of supporting materials commonly used for hydroprocessing catalysts; zeolite has the highest cracking activity.

As will be indicated later, the introduction of zeolite HY into Co-Mo catalysts significantly increased the HDS rate for some DBT-type compounds.⁵ One explanation for this increase is that the acidic zeolite improves cracking of the substituents in DBT compounds and promotes HDS reaction.

The improvement in HDS activities of catalysts primarily involves two categories:

1. Increasing the reactivity per unit catalyst volume
2. Changing the compositions of the active metals and/or the supporting materials.

The first category involves complicated manufacturing steps related to support materials, methods for introduction of the precursor of the active metals, thermotreatment, sulfidation, selection of special promoters, and others. ⁵ For the well-known Co-Mo catalysts, this approach resulted in a new generation of Co-Mo catalysts with up to 40% increase in HDS reactivity, depending on the application and operating conditions. ⁵

Examples of development in the second approach include the introduction of zeolite HY into Co-Mo-Al composition and the addition of new noble metals such as ruthenium and rhodium. These have reportedly increased the catalyst reactivities for DBT-type compounds.^{5 15}

As gasoline deep desulfurization mainly involves reactions of DBT-type compounds, successful commercialization of new catalysts with noble metals could have a major impact on the economics of deep desulfurization.

Processing options

Sulfur reduction of FCC gasoline for meeting the forthcoming sulfur regulations generally involves two processing alternatives ( Fig. 2 [101,960 bytes]):

1. Hydrotreating FCC gasoline
2. Hydrotreating FCC feed.

Hydrotreating FCC gasoline or cat gasoline requires less capital investment. It does not produce, however, any economic benefits except for compliance with environmental needs. Additionally, the high-octane olefin compounds in cat gasoline are concentrated in the light fraction.

Hydrotreating the full stream of cat gasoline results in octane number reduction as a result of hydrogenation of these olefins in the light fractions. A 7-10 loss in research octane number (RON) or 3-4 loss in motor octane number (MON) have been reported.¹⁶

This reduction can be minimized to 3.5 RON or 1.6 MON when only the heavy fraction (60%) of cat gasoline is hydrotreated and then blended back with the remaining light fraction to reach 92% sulfur reduction in the overall gasoline stream.¹⁶

About two thirds of the sulfur compounds in cat gasoline are in the last 15% heavy fraction. The loss of RON and MON can be further reduced to 1.0 and 0.6, respectively, when hydrotreating of cat gasoline heavy fraction is coupled with caustic treating of the light fraction for mercaptan removal.

Catalyst-distillation technology involving hydrotreating the light and heavy fraction ends at different reaction temperatures has been reported to minimize the loss of average RON and MON in the hydrotreated gasoline to 0.6.¹⁷ Licensed hydrotreating processes for cat gasoline with minimized losses of octane are also available.^{18 19}

Hydrotreating the FCC feed costs more than hydrotreating cat gasoline but results in economic benefits in addition to compliance with the sulfur regulations. These benefits vary depending on the characteristics of the cat feed, severity of hydrotreating, and operating conditions of the FCC.

One report indicated a 7-9 wt % increase in gasoline yield with constant FCC riser temperature, a 10% relative decrease in coke yield, and a 90% relative decrease in SO₂ emission from the regenerator.¹⁶ Other benefits include significantly lower Conradson carbon content and higher yield and better quality of middle distillates.

These benefits may be further maximized by simultaneously optimizing the operating conditions and catalyst selections for both the FCC feed hydrotreater and the FCCU. An economic analysis showed this synergy resulted in an economic benefit of $5.5 million/year (1992) for a 30,000 b/d FCCU.²⁰

Hydrocracking the cat feed for desulfurization results in higher increases in the FCCU gasoline yield than hydrotreating. Compared to the nontreated feed, an 18 vol % increase in gasoline yield has been reported by once-through hydrocracking of the FCC feed. At 60% conversion (40 vol % of feed is 390° C.+), the sulfur content was reduced from 1.4 wt % in the untreated FCC feed to 0.002 wt % in the 390° C.+ fraction of the hydro cracked feed.²¹

As discussed earlier, cracking promotes HDS of the DBT-type sulfur compounds. Nevertheless, cat-feed hydrocracking requires higher consumption of hydrogen than cat-feed hydrotreating.

Achieving the hydrocracking benefits in terms of added FCC conversion and selectivity is a delicate matter that will be impacted by the FCC hardware, catalyst, and raw gas-oil feedstock. The choice for hydrocracking the cat feed has impacts well beyond those driven by the issues of sulfur removal.

Other processing options to reach a target-sulfur level in the gasoline pool may be identified by an integrated, overall evaluation of the entire refining scheme, especially those units that contribute to the reduction of sulfur in the final gasoline pool. Streams that contribute to the sulfur in gasoline pool include unhydrotreated straight run naphtha, light thermally cracked streams that bypass the catalytic reformer, and poly-gasoline.

Reactor design

The design of a hydrotreating reactor is essential for the success of a hydrotreating unit. The most widely used reactor design is the cocurrent down-flow, trickle-bed reactor.

Table 2 [29,304 bytes] indicates the significant difference in the characteristics of catalysts used in different reactor designs.²²

While reactor designs other than the typical trickle-bed reactor design may offer added possibilities for deep desulfurization, probably without major investment, the number of commercial units with these novel reactor designs is much smaller than those with the trickle-type reactor. These novel designs may require special catalysts dedicated for a particular design.

Moreover, the smaller the catalyst volume fraction in the novel reactor design, the higher the level of operability difficulty.²² This is mainly due to the operation of continuous catalyst renewal in high temperature and pressure conditions.

In trickle-bed reactors, certain design features are essential. Flow distributors maximize uniform use of catalyst in the bed. Reaction-heat removal controls the reaction temperature.

For deep HDS of gasoline, staging the hydrotreating reactor into two catalyst beds with inter-bed quench takes advantage of different catalyst activities.

For example, Co-Mo catalyst is effective for hydrodesulfurizing light sulfur compounds such as mercaptans and sulfides and can be used in the first reaction bed. The second bed could use DBT reactive catalyst such as Ni-Mo catalyst.⁵ For staged hydrotreating reactors, trickle (fixed) bed reactors are probably preferred because these types of reactors involve the least complicated design.

In addition to providing more reaction zones for more than one catalyst, better efficiencies (less reactor volume) can be achieved if the hydrogen is free from H₂S and other contaminants.

Further reductions in the required catalyst volume can be accomplished by using a counter-current reactor where the hydrogen stream flows upwards, countercurrent with the hydrocarbon stream. The average partial pressure of H₂S in a counter flow system is less than that in the cocurrent flow.

Counter-flow systems have reportedly resulted in a 10-15% reactor volume reduction for an HDS target.²³

As yield and desulfurization guarantees are provided by catalyst vendors, strategic characterizations of sulfur compounds in the feed will be needed by the vendors to take advantage of the higher-efficiency staged system.

References

1. Andari, M.K., Behbehani, H., and Stanislaus, A., "Sulfur Compound Type Distribution Naphtha and Gas Oil Fractions of Kuwaiti Crude," Fuel Science & Technology International, Vol. 14, No. 7, 1996, pp. 939-61.
2. Wilson, J.W., FCC Technology and Operations, PennWell Publishing Co., 1997.
3. Gary, J.H., and Handwerk, G.E., Petroleum Refining, Technology and Economics, 2nd edition, Marcel Dekker Inc., p. 100.
4. Letzsch, W.S., and Ashton, A.G., "The Effect of Feedstock on Yields and Product Quality," Fluid Catalytic Cracking: Science and Technology, Studies in Surface Science and Catalysis, Vol. 76, Elsevier Science Publisher.
5. Landau, M.V., "Deep hydrotreating of middle distillates from crude and shale oils," Catalysis Today, Vol. 36, 1997, pp. 393-429.
6. Mudra, James, "Fluidized Catalytic Cracking of Hydrotreated Charge Stock for Naphtha Sulfur Reduction," Catalytic Hydroprocessing of Petroleum and Distillates, edited by Michael C. Oballa and Stuart S. Shih, Marcel Dekker Inc., 1994.
7. Badra, C., Perez, J.A., Salazar, J.A., Cabrera, L., and Garcia, W., "Sulfur and Octane Trade Off in FCC Naphtha Conventional Hydrotreating," Erdol Erdgas Kohle, Vol. 6, No. 113, June 1997.
8. Beuttner, H., and Schmid, B.K., Proc. 6th world Petroleum Congress, 1963, Vol. 3, p. 197.
9. Kumar, V.R., Balaraman, K.S., Rao, V.S.R., and Ananth, M.S., "Modelling of Hydrotreating Process in a Trickle Bed Reactors," Petroleum Science and Technology, Vol. 15, Nos. 1&2, 1997, pp. 283-95.
10. Raychaudhuri, T.S. Banerjee, and Ghar, R.N., "Kinetic Parameters of Hydroprocessing Reactions in a Flow Reactor," Fuel Science and Technology International, Vol. 12, No. 2, 1994, pp. 315-33.
11. Sau, M., Narasimhan, C.S.L., and Verma, R.P., "A Kinetic Model for Hydrodesulfurization," Hydrotreatment and Hydrocracking of Oil Fractions, edited by G.F. Froment, B. Delmon, and P. Grange, Elsevier Science BV, 1997.
12. Vanrysselberghe, V., and Froment, Gilbert F., "Kinetic Modeling of Hydrodesulfurization of Oil Fractions: Light Cycle oil," Ind. Eng. Chem. Res., 1998, Vol. 37, pp. 4231-40.
13. Ma, X., Sakanishi, K., and Mochida, I., "Hydrodesulfurization Reactivities of Various Sulfur Compounds in Diesel Fuels," J. Ind. Eng. Chem. 1994, 33, p. 218-22.
14. Lamure-Meille, V., Schulz, E., Lemaire, M., and Vrinat, V., Appl. Catal., Vol. 131, 1995, p.143.
15. Grange, P., and Vanhaeren, X., "Hydrotreating Catalysts, An Old Story with new Challenges," Catalysis Today, Vol. 36., 1997, pp. 375-91.
16. Upson, L.L., and Schnaith, M.W., "Low-sulfur specifications cause refiners to look at hydrotreating options," OGJ, Dec. 8, 1997.
17. Rock, K.L., Foley, R.M., Putman, H.M., Bakshi, A.S., and Som, M., "Catalytic Distillation to Enhance Gasoline Quality: Part I," Petroleum Technology Quarterly, Spring 1998.
18. Refining Processes '98, Hydrocarbon Processing, November 1998.
19. Antos, G.J., Solari, B., and Monque, R., "Hydroprocessing to Produce Reformulated Gasoline, The ISAL Process," Hydrotreating and hydrocracking of oil fractions, edited by G.F. Froment, B. Delmon, and P. Grange, Elsevier Science BV, 1997.
20. Desai, P.H., Keyworth, D.A., Asim, M.Y., Reid, T., and Pichel, A.H., "Enhance Gasoline Yield and Quality," Hydrocarbon Processing, November 1992.
21. Scherzer, J., and Gruia, A.D., Hydrocracking Science and Technology, Marcel Dekker Inc., 1996.
22. Morel, F., Kressmann, S.

    , Harle, V., and Kasztelan, S., "Processes and Catalysts for Hydrocracking of Heavy Oil and Residues," Hydrotreatment and Hydrocracking of Oil Fractions, edited by G.F. Froment, B. Delmon, and P. Grange, Elsevier Science BV, 1997.

23. Krishna, R., and Sie, S.T., Chem. Eng. Science, Vol. 49, No 24A, 1994, p. 4029.

The Author