Revamping HDS Units to Meet High Quality Diesel Specifications

Revamping HDS Units to eet High Quality Diesel Specifications by F. Emmett Bingham Haldor Topsoe, Inc., California, USA and Preben Christensen Haldor Topsøe A/S, Lyngby, Denmark Presented at the Asian Pacific Refining Technology Conference, 8-10 arch, 2000, Kuala Lumpur, alaysia

2 Abstract Environmental legislation continues to favor more stringent global diesel quality specifications. The U.S. and the European Union have adopted low sulfur diesel specifications, and specifications with ultra low sulfur contents are either mandated or under serious consideration after year 2005. Historically in Asia, air quality has tended to deteriorate, perhaps as a consequence of very rapid industrial growth. However as Asia becomes richer, the demand for future cleaner burning fuels is expected to become a significant issue. ore stringent diesel quality criteria for properties other than sulfur content are also expected in the future. The automobile industry, in their World Wide Fuel Charter issued in early 1999, has recommended rather stringent limits on, aromatics, poly-aromatics, cetane number/index, and back-end distillation temperatures. The European Union is expected to finalize their year 2005 fuel specifications during year 2000 and the specifications in the World Wide Fuel Charter are anticipated to have a strong influence on the final criteria. Depending on which of the above diesel quality specifications are adopted, it is likely that meeting proposed diesel aromatics and burning quality criteria will be more challenging than meeting the sulfur specifications. ore sophisticated process technologies and catalysts will be needed to meet these specifications. Topsoe has been a pioneer in the design of units for diesel desulfurization and aromatics saturation to produce cleaner burning fuels. This paper presents examples of Topsoe commercial experience and provides a detailed illustration of the approach one refiner in the United States will take to economically revamp an existing HDS unit to produce ultra low sulfur, low aromatics diesel.

3 Changes in Diesel Specifications and Demand In recent years, the development and use of environmentally friendly fuels has been of high priority throughout the world. In Europe, the specifications adopted by EU for year 2000 concerns sulfur content, density, Poly-Aromatic Hydrocarbons (PAH) content, cetane number and 95% AST distillation point. Year 2005 concerns further reduction in the diesel sulfur specification. The remaining specifications for the year 2005 are expected to be settled during year 2000 and these will most likely be more stringent than those adopted for year 2000. This expectation is based on the fact that, in January 1999, the automobile and engine manufacturing industry in USA, Japan and Europe issued a World Wide Fuel Charter, that specified stringent requirements for a large number of diesel fuel parameters. The year 2000 diesel specifications in European Union and the World Wide Fuel Charter are given in Table 1. Table 1: Future Diesel Specifications Specification EU Year 2000 Fuel Charter Cetane Number 51 (min) 55 (min) Cetane Index NA 52 (min) Density @ 15 C, g/cm 3 0.845 (max) 0.840 (max) Distillation 90% Boiling Point, C NA 320 (max) 95% Boiling Point, C 360 (max) 340 (max) Final Boiling Point, C NA 350 (max) Poly-aromatic Hydrocarbons, wt% 11 (max) 2.0 (max) Total Aromatics Content, wt% NA 15 (max) Sulfur Content, wppm 350 * (max) 30 (max) *From Year 2005 the European Union has adopted a sulfur content of 50 wppm The United States is currently debating diesel specifications with most recent speculation that both EPA and CARB will slash allowable diesel fuel sulfur levels by year 2007 (EPA target: 5-40 ppm sulfur). Additionally, specifications for lower aromatics content and increased diesel cetane number are expected. Current diesel sulfur specifications in Asia vary over a wide range from 0.5 wt% to less than 0.05 wt%. The average sulfur level, estimated to be 0.19 wt% for year 2000, will continue to drop due to the definite movement away from 0.5% grades toward 0.05% grades. Japan and Korea, currently producing diesel at or below 0.05 wt% sulfur, are expected to tighten their specifications to closely follow the specifications set by the EU and the United States. Thailand, alaysia, Singapore, and the Philippines will have 0.05% diesel sulfur specifications by 2000 2004, and will most likely consider further future reduction. India has recently introduced a sulfur specification of 0.2 wt% and is already now considering further reductions to 0.05 wt% sulfur. Japan and Korea will also most likely follow the

4 specifications for reduced aromatics and T95 promulgated in the EU and the United States. However, there is less consensus among other Asian countries concerning the adoption of these specifications. At the same time diesel fuel specifications are tightening, the demand for diesel is growing in Asia as well as in Europe. Although the growth rate for middle distillate in Asia was relatively stagnant from 1997-1999, the forecast future demand, according to Trans-Energy Research Associates, Inc. will be equivalent to the pre 1997 rate of approximately 5-8% per year. In Europe, the demand for home heating oil and fuel oil cutter stocks, two of the other common uses of diesel boiling-range material, is decreasing. Consequently, more diesel boiling-range material is becoming available for blending into the diesel pool. Unfortunately, these lower quality stocks will require severe hydrotreatment to meet the future specifications for diesel fuel. The refining industry is therefore facing a double-edged challenge: to meet new, more stringent specifications for diesel product, while simultaneously producing more diesel product from lower quality feedstocks. The combination of these factors places a heavy burden on the refiner s hydroprocessing capabilities. As a result, increased new hydrotreating capacity and revamp of existing facilities will be needed to meet the future diesel specifications. eeting Future Specifications Depending on the diesel specifications adopted, the issues facing the Asian refiner in the future could be very similar to those faced by the EU refiner today. Properties being considered for future diesel specifications in the EU can serve as basis of discussion for this paper. Sulfur Reduction The current diesel sulfur specification in the EU is 350 wppm, and for year 2005 a further reduction to 50 wppm will be mandatory. Germany is planning to offer tax incentives to phase in diesel with 50 wppm sulfur from November 2001 and 10 wppm sulfur by November 2003. In Asia, the estimated average, year 2000, diesel sulfur content is 0.19 wt%. However, the average diesel sulfur content is expected to rapidly approach 0.05 wt%, with lots of pressure for continued reduction. The impact of reducing diesel sulfur content from 2500 wppm to 10 wppm for the same feed is illustrated in Table 2. The data, summarized in the table, was calculated using Topsøe s kinetic model, for a typical diesel hydrotreater using a Coo type of catalyst. The left column shows the target diesel sulfur content in wppm. The right-hand column shows how much additional catalyst, having the same currently available activity, is required to reach the lower sulfur targets while maintaining a constant cycle length. All other operating conditions remain unchanged.

5 Table 2: Impact of Diesel Sulfur Reduction Diesel Sulfur wppm Relative Catalyst Volume to aintain Constant Cycle 2500 1.0 500 1.5 350 1.8 50 3.8 10 5.3 Table 2 can be used to predict the impact of future Asian regulations for reduced diesel sulfur. New diesel HDS catalysts are being developed to provide sufficient additional activity to allow the refiner to reduce the diesel sulfur content from 2500 wppm to 500 wppm with only moderate impact on the cycle length. However it is very unlikely that catalyst HDS activities will be increased sufficiently in the near future to meet the 10 50 wppm sulfur specifications for Germany or the EU. easures other than a simple catalyst replacement must therefore be considered to meet these specifications. Process parameters that can be modified to assist in meeting deeper desulfurization are discussed in more detail in the process modification section. Aromatics Reduction The new EU diesel specifications for year 2000 and most likely also for year 2005 will include a limit on the content of poly-aromatic hydrocarbons (PAH s). The maximum content of PAH s for the year 2000 diesel has been set at 11 wt%. This can be met by most existing hydrotreaters unless the feed being processed contains a very high amount of catalytically cracked material. The PAH content increases significantly if LCO is blended in the feedstock. The conversion of PAH s is moderate in hydrotreaters having relatively low hydrogen partial pressures. Therefore for units with hydrogen partial pressures in the range of 30 bar, the feed blend should contain no more than approx. 30 wt% PAH s in order to meet the year 2000 specification of 11% PAH. No limit for PAH content for year 2005 diesel, has as yet been established by the EU. However, it is likely that PAH content will be reduced considerably compared to the year 2000 specification and values as low as 1 wt% have been mentioned. Although PAH s react quite readily, their conversion is thermodynamically limited and can only be improved by conversion of mono-aromatics compounds. As a result, it will be difficult for hydrotreaters operating at typical design pressures with conventional base-metal catalysts, to meet the 2 wt% specification for PAH s proposed by the automobile and engine manufacturers. To demonstrate the thermodynamic limitations on PAH conversion, pilot plant tests at operating conditions typical of many existing hydrotreaters, were conducted on a straight-run gas oil containing 32 wt% total aromatics and 16 wt% PAH s. The hydrogen partial pressure was increased from 30 bars to 60 bars and the product PAH content was measured. The results of these tests are shown in Figure 1.

6 Figure 1: Saturation of PAH s as Function of Temperature and Pressure 7 6 Increasing Hydrogen Pressure 5 Wt%, PAH 4 3 2 1 0 310 320 330 340 350 360 370 380 390 400 410 Temperature, C As seen in Figure 1, it is barely possible to reach 2 wt% PAH content in the diesel product, even at the highest hydrogen partial pressure. Lowering space velocity lowers the required reaction temperature and the thermodynamic equilibrium becomes more favorable for aromatic saturation. Even so, as temperature is increased to compensate for catalyst deactivation it becomes increasingly difficult to maintain the low PAH content. Achieving a diesel product with 2 wt% PAH content will be even more demanding for the refiner processing feeds containing cracked stocks. In general, diesel hydrotreaters, operating at hydrogen partial pressures of 35-50 bars and 1.5-2.0 hr -1 space velocity, with base-metal catalysts, can produce diesel with a PAH content of 2-6 wt% when processing 100% straight-run feed. However, at the same operating conditions, the product PAH content will increase to 4-8 wt%, if the feed contains 20-30% LCO. any existing hydrotreaters may be capable of achieving a future PAH specification of 6 wt% or higher. However, if a future PAH specification of 4 wt% or lower must be met, the addition of a second stage hydrodearomatization reactor to the existing diesel hydrotreater or the installation of a high pressure single stage unit will likely be required. When the final specification for PAH content is settled, refiners will have to carefully evaluate their existing hydrotreaters and determine how best to integrate new equipment with the existing unit. Topsøe has the engineering experience to complete feasibility studies to help the refiner make these evaluations. In addition to reduction of the PAH content, the reduction in total aromatics content in diesel may also become an issue in order to meet the anticipated new diesel specifications for density, cetane number or cetane index. Eventually a specification on the total aromatics content as suggested in the World Wide Fuel Charter may be introduced. Reduction of the total aromatics content is much more difficult than reduction of the PAH content, because saturation of mono-aromatics to naphthenes is much more difficult than saturation of poly-aromatics to mono-aromatics. Processing a typical straight-run gas oil in a single-stage hydrotreater, to reduce the total aromatics content to 15 wt% as required by the World Wide Fuel Charter, will require rather severe reactor conditions including a hydrogen

7 partial pressure greater than 70 bars. Processing feed blends containing cracked material, to meet the 15 wt% total aromatics specification, will require even more severe conditions making the single-stage approach less economically attractive. For an existing moderate pressure diesel hydrotreater using base-metal catalyst (Nio or Coo), the reduction in total aromatics content is very limited, due to the relatively low hydrogenation activity of the base metal catalyst. Adding a second stage with a high activity noble metal catalyst offers the best solution for achieving the required aromatics reduction at moderate hydrogen partial pressures. A separate second stage is necessary because nitrogen and sulfur containing compounds must be removed in the first stage, as they are temporary poisons to the noble metal catalyst. Improvement of Cetane Number/Cetane Index Cetane number and cetane index are measures of the ignition quality of diesel fuel. Cetane number is determined by running the fuel in a test engine and cetane index is calculated based on measured values of the density and AST distillation. Both the cetane number and the cetane index are strongly dependent on the amount and type of aromatic components in the feedstock. Saturation of aromatics will therefore improve the cetane number/index considerably. However the improvement attainable is also strongly dependent on the amount of sulfur and nitrogen species in the feed and the nature of the feedstock. Deep hydrodesulfurization will in itself improve the cetane number/index, and further increase in cetane number can be achieved through the use of additives. However, it is possible that both cetane number and cetane index specifications will be adopted to limit the use of additives to boost the cetane number. If this happens, the refiner will be forced to reduce the total aromatics content of their diesel. Again a two-stage HDS/HDA process will be an effective way of achieving this. Reduction of 95% Distillation Point Future diesel specifications may call for a reduction of up to 20 C below the year EU 2000 T95 specification of 360 C. Similarly future reductions in tail end distillation temperature reductions are being considered in some parts of Asia. The simplest method of reducing the 95% distillation point of the diesel product is to adjust the draws of the crude distillation tower to reduce the diesel endpoint. Reduction of the 95% distillation point will have the following positive effects: Density will be reduced ost refractive sulfur components will be removed from the feedstock Concentration of heavy aromatics components will be reduced Cold flow properties of the diesel will be improved However, this will leave the refiner with a heavy gas oil fraction boiling between say 340-360 C representing 10 15 vol% of the diesel previously produced. One solution to bring this material back in the future diesel pool would be to crack this narrow boiling fraction to lighter material. The chemical nature of the hydrocarbons in this narrow boiling fraction, however, limits what can be achieved by cracking this material. The molecules present in this narrow boiling fraction will typically have between 20 and 23 carbon atoms. If these molecules are

8 cracked in the middle, the product fractions will have between 10 and 12 carbon atoms, which boil in the heavy naphtha/light kerosene range. If the molecules in the narrow boiling fraction are cracked unsymmetrically, one fraction will boil in the diesel range and the other will be either light naphtha or gas. Cracking the narrow boiling fraction will therefore result in a major naphtha production as well as formation of significant amounts of hydrocarbon gasses and only a limited portion will remain in the diesel range. Topsøe therefore believes that means other than cracking of the narrow boiling fraction must be found. Density Reduction The extent to which the density is reduced during deep hydrodesulfurization depends on the feedstock that is processed and on the operating conditions. For a typical EU refinery, the density is reduced by 0.01-0.02 g/cm 3 as a result of sulfur removal and saturation of di- and tri-aromatic compounds. Together with the reduction in density of 0.003-0.005 g/cm 3 obtained by the anticipated reduction in 95% distillation temperature, meeting the year 2005 density specification recommended by the automobile industry, will be possible for many existing diesel hydrotreaters operating to meet 50 wppm of sulfur. If additional density reduction is required, saturation of mono-aromatics compounds will be necessary but difficult to achieve at the pressures normally employed in diesel hydrotreaters. In this case a two-stage HDS/HDA unit is a good approach for achieving further density reduction. Figure 2 shows the relationship between density reduction and aromatics saturation obtained using Topsøe s aromatic saturation catalysts TK-907 and TK-908. There is a spread in the data, but roughly speaking a reduction in total aromatic content by 10% (absolute) will result in approximately 0.007 g/cm 3 reduction in density. Figure 2: Density Improvement vs. Aromatics Removal Density Improvement 0-0.005-0.01-0.015-0.02-0.025-0.03 TK-907, TK-908 0 10 20 30 40 50 Aromatics Removed, vol%

9 Process Changes for eeting Future Stringent Regulations The primary process changes for meeting future sulfur regulations in an existing single stage unit can be categorized into either reactor modifications, or recycle gas loop modifications. ost units revamped to meet future product sulfur targets will employ one or both of these modifications. Reactor odifications odifications associated with the reactor can include: Use of more active catalyst in the existing reactor Increasing the catalyst quantity by dense loading the existing reactor Increasing the catalyst quantity by addition of a new reactor Improving the performance of the existing reactor by installing state of the art reactor internals. Catalyst Options Since the first introduction of the hydrotreating process in the 1950 s, catalyst manufacturers have made significant improvements in catalyst activity. Today s high activity hydrotreating catalyst is about eight times more active than the first generation hydroprocessing catalyst. Topsøe first introduced a full range of hydrotreating catalysts for refining applications in the early 1980 s. Table 3 lists our Cobalt oly catalyst typically used in hydrodesulfurization applications and the relative improvement over the years. Topsøe s most active Coo catalyst, TK-574, has about twice the activity of our first generation catalyst, TK-550, introduced in early 1980 s. By applying TK-574 instead of older types of Coo catalysts will enable a typical Asian refiner to reduce the sulfur level from say 2500 wppm to 350-500 wppm with only minor influence on the cycle length. Table 3: TK-Catalysts 500 Series: Coo Name TK-550 HDS Activity Base TK-554 Base * 1.5 TK-554+ Base * 1.7 TK-574 Base * 2.0 Catalyst Loading Catalyst quantity can be increased by 15% by dense loading an existing reactor. This corresponds to a 3 C to 4 C reduction in start of run temperatures. Hardware limitations such as reactor internals design and compressor pressure drop limitations may prevent some refiners from dense loading their catalyst. LHSV Adding a reactor or replacing an existing reactor is the normal approach taken when revamping a unit to meet more stringent product requirements. Assuming all other operating

10 conditions remain unchanged, doubling the catalyst volume results in a 20 C reduction in average temperature. This reduction can be used to offset the increased severity required for lower product sulfur. Increasing the catalyst volume has a double effect on performance: The start of run temperature is lower resulting in an increase in available temperature span from start to end of run. The deactivation rate is lower due to the lower start of run temperature and larger quantity of catalyst. If an additional reactor is to be needed, the refiner has several choices as to the location of the new reactor. The new reactor may be located in series, either up stream or down stream of the existing reactor, or in parallel to the existing reactor. The decision will largely depend on the limitations of the existing equipment. Reactor efficiencies based on mass velocity as well as reactor pressure drop will play an important role in the choice of series or parallel configuration. The recycle compressor capability should also be considered when making this decision. Other factors that should be considered include quantity of catalyst to be added, reactor temperature control requirements, heater requirements, etc. Topsøe has revamped several hydrotreaters requiring the modification of, addition to, or replacement of the existing reactor, and therefore has the experience to assist the refiner in making the best decisions for modifying his particular unit. Reactor Internals Reactor internals play a key role in facilitating the contact of reactants with the catalyst. Poor distribution of the reactants over the catalyst can contribute to channeling through the catalyst bed resulting in inefficient utilization of the catalyst, development of hot spots, and catalyst deactivation due to coke formation. Figure 3 is a simple representation of the significant effect poor distribution can have on deep diesel desulfurization. Figure 3: Effect of liquid distribution on product sulfur FEED Sulfur = 1.5 wt% Poor Flow Sulfur Contribution from Unreated Liquid (wt ppm) 350 300 250 200 150 100 50 0 Bypassing 0.5 1 2 % Untreated Liquid

11 The schematic on the left shows a poorly performing distributor causing catalyst bypassing. The bar chart on the right shows the contribution of sulfur in the total product caused by increasing levels of bypassing of the feed. As demonstrated by this figure, improved distribution to eliminate bypassing will be mandatory as the product sulfur target gets lower and lower. Reactor internals in most hydroprocessing units are out-dated and may not be able to provide the high performance efficiency needed to fully utilize today s high performance catalysts. Topsøe realized this dilemma and undertook and extensive development program in 1990 using several different pilot plants and test apparatus. From this test work computer models were developed and used to design new state of the art reactor internals. These models can also be used to evaluate the performance of existing reactor internals. The most critical piece of hardware for evenly distributing reactants across the catalyst is the liquid distribution tray. Topsøe s Vapor-Lift distributor tray is a state-of-the-art design for use at the top of the reactor or below a quench section. This design was first tested in our pilot plants in 1996-1997 and the first commercial installation was completed in 1998. To date we sold more than 40 trays. The commercial performance of the Vapor Lift Tray has been excellent. The replacement of the impingement mixers and bubble cap trays at the Syncrude Heavy Coker Gas Oil Hydrotreater with Topsøe s Vortex mixing chambers and Vapor Lift trays demonstrates typical commercial results with Topsøe reactor internals. Reactor radial temperature gradients are indicative of the performance of the reactor internals. Figure 4 shows reactor temperature profiles for the existing reactor internals. Figure 5 shows temperature profiles after the impingement mixers and bubble cap trays were replaced with the Topsøe Vortex mixers and Vapor Lift trays. These figures show the reactor temperatures at a variety of radial locations from the top of the first bed (percent catalyst = 0) to the bottom of the third bed (percent catalyst = 100). The radial temperature gradients were reduced from over 40 C to an average radial temperature gradient of 2 C with the Topsøe internals, demonstrating the superiority of the Vortex mixers and Vapor Lift Trays. It should be noted that at the time the trays were replaced, the reactor thermometry was also changed to include Gayesco Flexible-type thermocouples. This resulted in an increase in the thermocouple coverage at any given elevation. The data from the dense pattern of flexible thermocouples further establishes the temperature uniformity attained with the Topsøe tray.

12 Reactor Temperature, C Figure 4: Temperature profile in existing 3 beds with bubble cap trays +90 +80 +70 +60 +50 +40 +30 +20 +10 Figure 5: Temperature profile in the 3 beds with Topsøe Vapor Lift trays Base 0 10 20 30 40 50 60 70 80 90 100 Percent of Catalyst +80 Reactor Temperature, C +70 +60 +50 +40 +30 +20 +10 Base 0 10 20 30 40 50 60 70 80 90 100 Percent of Catalyst

13 Recycle Gas Loop odifications odifications to the recycle gas loop for meeting future regulations include: Increasing the hydrogen concentration of the gas Reducing the hydrogen sulfide concentration of the gas Increasing the recycle gas to oil ratio. Recycle Gas H 2 Concentration Increasing the recycle gas H 2 concentration will increase the reactor hydrogen partial pressure. This is accomplished by purging recycle gas or by increasing the hydrogen concentration in the make-up gas. Increasing the hydrogen partial pressure reduces the reactor start of run temperature and also reduces the rate of catalyst deactivation. If purge of recycle gas is used, the purge gas can be sent to a membrane separation unit or PSA unit in order to recover the hydrogen, which can then be recycled. In new units, higher hydrogen partial pressure can be achieved by designing the unit for higher operating pressure. However, the effect of increasing hydrogen partial pressure by increasing total pressure is not as pronounced as that achieved by increasing hydrogen purity. The reason for the difference in response is that H 2 S partial pressure is also increased when total pressure is increased where as increasing the recycle gas purity does not affect the H 2 S partial pressure. Figure 6: Impact of Increasing H 2 Partial Pressure Reduction in in WABT, WABT, F o C 00-2-4-8 -4-12 -8-16 -10-20 -12-24 -14-28 Gas purity 100 110 120 130 140 150 160 Hydrogen Partial Pressure (% Base) Recycle Gas H 2 S Concentration Reactor Pressure If the hydroprocessing unit does not have a recycle gas scrubber, H 2 S in the high-pressure loop will build up to a high concentration and inhibit the desulfurization reaction. The reactor temperature must then be increased to offset the hydrogen sulfide inhibition. This effect is greater at higher total reactor pressure and more pronounced for Coo catalysts than for Nio catalysts.

14 Figure 7 shows the effect of recycle gas scrubbing on the average bed temperature at two pressure levels. Figure 7: Impact of Recycle Gas Scrubbing I n c r e a s e I n W A B T,ºF Increase in WABT, o C 25 50 20 40 15 30 10 20 5 10 0 P r e s s u r e = 3 0 B a r P r e s s u r e = 6 0 B a r 0 2 4 6 % H 2 S a t t h e R e a c t o r I n l e t The effect on catalyst activity illustrated by this plot takes into account only the effect of the hydrogen sulfide. Since the recycle gas hydrogen concentration also decreases when the hydrogen sulfide concentration increases, there is an additional debit on the catalyst deactivation rate due to the lower hydrogen partial pressure. An increase from zero to five percent hydrogen sulfide in the recycle gas is equivalent to a required increase in average reactor temperature of about 17 C at 30 bar total pressure and 22 C at 60 bar total pressure. Recycle Gas/Oil Ratio Increasing the recycle gas to oil ratio also decreases the average hydrogen sulfide partial pressure in the reactor, and in turn increases the apparent catalyst activity for Coo type catalysts. However, a relatively large increase in the gas rate is required to have the same effect as scrubbing the recycle gas. Even if the recycle gas is scrubbed, increasing the recycle gas to oil ratio decreases the reactor hydrogen sulfide partial pressure and therefore reduces the required reactor temperature. The effect is greater at higher pressure as shown in Figure 8.

15 Figure 8: Impact of Increasing the Gas/Oil Ratio Increase in WABT, C 15 10 5 0-5 -10-15 Pressure = 30 Bar Pressure = 60 Bar 50 100 150 200 Gas/Oil Ratio (% Base)

16 Topsøe Two-Stage HDS/HDA Process Process Topsøe was one of the pioneers to design units for hydrodearomatization of diesel using noble metal catalysts. The Topsøe two-stage HDS/HDA process is illustrated below in Figure 9. Figure 9: Diesel Upgrading Topsøe s Two-stage Process ake-up Hydrogen Recycle Gas Compressor Diesel Feed Wash Water Amine Scrubber Overhead Vapor First Stage Reactor Charge Heater HDS Stripper Water HDS Reactor HDS Separator Wild Naphtha Sour Water Product Diesel Stripper Steam Diesel Product Second Stage HDA Reactor Diesel Cooler HDA Separator The first stage is a conventional hydrotreating step performed over a base-metal catalyst to reduce the feed sulfur and nitrogen contents to sufficiently low levels to allow the secondstage noble-metal catalyst to perform the required degree of dearomatization at a high LHSV. Following this first stage hydrotreating, the diesel leaving the separator contains a significant quantity of dissolved hydrogen sulfide and ammonia, which is removed in an intermediate stripper column, using recycle hydrogen as the stripping medium. The hydrogen sulfide and ammonia containing off-gas is then purified in an amine scrubber. The stripped diesel and scrubbed hydrogen are then fed to the second-stage, hydrodearomatization reactor. The catalyst used for dearomatization is one of Topsøe s noble metal catalysts, TK-907, TK-908, or TK-915. After the second dearomatization stage, the diesel is steam stripped to remove small amounts of H 2 S present in the oil and to adjust the front-end distillation temperature to meet flash point requirements for the product diesel. The Topsøe two-stage diesel HDS/HDA process uses conventional hydrotreater technology at moderate pressure which enables the potential maximum reuse of equipment during the revamp of existing units.

17 Commercial Experience, Topsøe s HDS/HDA Process Kuwait Petroleum (Denmark) Refining One of the first units in Western Europe to produce Swedish Class I diesel (with less than 5 vol% aromatics and less than 10 wppm of sulfur) was the Kuwait Petroleum (Denmark) Refining (KPDR) unit in Stigsnæs, Denmark, licensed from Topsøe in 1992. The KPDR unit proved it was more practical to increase the severity in the HDS stage to produce a better feed for the HDA stage than to increase the severity in the HDA stage to compensate for insufficient pretreatment of the feedstock. KPDR therefore aimed at having a feed to the HDA stage that only contained a few wt ppm of sulfur and nitrogen throughout the life of the catalyst. Figures 10 shows the actual of feed and product aromatics content and reactor WABT during operation of this unit. Subsequent to the initial start-up in 1993, the unit was modified once in 1995 to increase the plant capacity. However, the original charge of TK-908 from early 1992 was still in operation, without any regeneration, when the KPDR refinery was closed in ay 1997. The analysis of operating data showed that the deactivation of the HDA activity was 15 C to 20ºC from SOR until the unit was shut down.

18 Figure 10: Performance of KPDR Diesel HDS/HDA Plant (After Revamp) San Joaquin Refining, California In September 1997, a Topsøe two-stage HDS/HDA unit was licensed to San Joaquin Refining in California. The unit was designed to convert light vacuum gas oil to CARB diesel, i.e. diesel with a total aromatics content of less than 10 wt%. This Topsøe two-stage unit, having a hydrogen partial pressure of approximately 70 bar and a HDA reactor average bed temperature of 287 C, produces diesel containing less than 4 wt% aromatics. The specific gravity of the diesel is also reduced by about 5%, and the cetane index of the diesel product is increased by 10 numbers. The properties of the feed, intermediate first-stage product, and the final product from the HDA reactor are given in Table 4. As can be seen, the severe pretreatment of the feed in the HDS stage allows a much higher LHSV in the HDA stage resulting in a major saving in catalyst cost.

19 Table 4: Performance of Topsøe Two-stage HDS/HDA Unit Density @ 15 C, g/cm 3 Sulfur, wt ppm Nitrogen, wt ppm Cetane Index (AST D976) Total Aromatics, wt% TBP Distillation, C 10% 50% 95% Conditions H 2 Pressure, kg/cm 2 Temperature, C LHSV, hr -1 Feed (1st Stage) TK-555 0.91 6515 775 39 33 255 315 370 71 367 Base Feed (2nd Stage) TK-907 0.88 9 3 46 24 - - - 70 287 Base x 3 Product 0.87 1 <1 49 3.5 - - - Revamping Existing HDS Units to Topsøe HDS/HDA Technology Ultramar/Diamond Shamrock Wilmington, California A good example of retrofitting Topsøe HDS/HDA technology is the revamp design of Unit 80 at the Ultramar/Diamond Shamrock (UDS) refinery in Wilmington, California. The existing unit was originally designed as an FCC feed pretreater but was downgraded to diesel hydrotreating service when UDS commissioned a new higher severity FCC pretreater. UDS realized that Unit 80 had the potential for more than simple HDS service and thus after evaluating currently available HDS/HDA technologies, selected Topsøe to convert this unit to produce diesel meeting CARB specifications. Pilot Plant Confirmation The feed to the unit was a blend of straight-run light gas oil, with approximately 35% coker light gas oil and FCC light cycle oil. Because the feed blend was very aromatic, UDS required assurance that the proposed revamp modifications could accomplish the desired 10% aromatic diesel product required by the CARB specifications. Therefore a pilot plant test was undertaken to process the exact feed blend provided by UDS at the operating conditions proposed for the revamped unit. The pilot plant test results in Table 5 summarize the properties of the feedstock, product from the first HDS stage and final product from the second HDA stage.

20 Table 5: Ultramar/Diamond Shamrock Unit 80 Revamp Pilot Plant Results Feed 1st Stage Product 2nd Stage Product Catalyst TK-555 TK-907 Temperature, C Base Base 75 LHSV, h -1 Base Base x 2 Sulfur, wppm 8220 11 2 Density @ 15 C, g/cm 3 0.871 0.859 0.842 Distillation, C IBP/50%/95% 90/269/353 211/264/352 201/258/348 Aromatics Content, wt% Total ono PAH 39.6 19.1 20.5 37.9 33.0 4.9 6.0 6.0 <0.4 Cetane Index 41.4 43.9 48.0 Cetane Number 36.2 38.8 47.4 The data in Table 5 shows that although approximately 99.9% desulfurization was achieved in the first stage with high activity TK-555 Nio catalyst, the overall reduction in aromatic content at the 56 bar operating hydrogen partial pressure was quite small. Likewise, there was only a moderate increase in cetane number and decrease in density due to the removal of sulfur and nitrogen. The PAH content remained close to 5% because the removal of PAH s was constrained by thermodynamic equilibrium at the required 360 o C operating temperature of the first stage. In the second HDA stage, using TK-907 aromatic saturation catalyst, more than 84% aromatic saturation was obtained confirming a total product aromatics content of less than 6% was achievable from the proposed unit. The cetane number was increased by more than 8 numbers compared to the first stage product and by more than 11 numbers compared to the feed. Despite the much lower reaction temperature and higher LHSV used in the second stage, the concentration of PAH was below the detection limit, confirming that the removal of PAH was equilibrium controlled. In addition to meeting CARB diesel specifications, UDS wanted to maximize the yield of diesel while retaining the ability to extract 10 to 30% jet fuel boiling range material present in the feed. The estimated product yields predicted by the pilot plant test work, based on the 218.6 m 3 /hr (33,000 bpd) design capacity, are shown in Table 6. Table 6: Product Yields Yields at 6% Product Aromatics Rundown Product Start-of-Run End-of-Run Fuel Gas, Nm 3 /hr 2.3 2.8 Total Naphtha, m 3 /hr 11.1 13.4 Jet Fuel, m 3 /hr 33.1 33.1 CARB Diesel, m 3 /hr 186.9 185.2

21 The fractionation section was designed so that UDS could consistently draw over 33 m 3 /hr of jet fuel throughout the operating cycle and if desired they could draw up to 73 m 3 /hr. Since the jet fuel fraction would be treated in the aromatic saturation section, the quality was very good with a smoke point around 25 mm throughout the cycle. Furthermore, if jet fuel demand decreased, UDS could reduce the jet draw and increase the production of CARB diesel. Revamp Procedure Existing Configuration Having proven that the proposed revamp design was capable of producing CARB quality diesel; the next hurdle was to complete the revamp design while maximizing the use of the existing equipment. The high pressure section of the existing Unit 80, as illustrated in Figure 11, is a conventional HDS single stage configuration with two single bed reactors in series, an amine scrubber to remove H 2 S from the recycle gas, but no hot high pressure separator. Figure 11: Unit 80 Existing Reaction Section Configuration Reactor Charge Heater HDS Reactors ake-up Hydrogen Recycle Gas Compressor Amine Scrubber E1A-D PC LSRGO LCGO V1 P4 Water P1 Charge Pump Wash Water Surge Drum E2 Product Separator Sour Water Effluent to Fractionation Section The process flow of the existing fractionation section is illustrated in Figure 12. The liquid from the high-pressure separator is preheated by exchange with fractionator product streams and light ends are removed in a high-pressure stripper. The stripped product then enters the main fractionator which produces naphtha overhead and has side cut strippers to separate jet and light diesel. Heavy diesel is produced as fractionator bottoms, and polished through a salt dryer.

22 Figure 12: Unit 80 Existing Fractionation Section Configuration Sour Fuel Gas Wild Naphtha S W Fractionator Naphtha Wash Water to Rx Section Stripper A Jet Stripper From Reactor Section A Fract Feed Heater Diesel Stripper Jet Fuel Diesel Diesel Salt Dryer Process Feasibility Studies One of UDS primary project objectives was to minimize the capital investment for the revamp by reusing as much existing equipment as possible. Furthermore, the available plot space was very tight, so the addition of new equipment had to be minimized. To pursue this objective, the initial steps in the revamp approach included a series of feasibility studies to determine the best options for utilizing the existing equipment. Figure 13: Process Optimization Studies Evaluate Suitability of Existing Internals in HDS Reactors Evaluate Using Surplus Reactor for HDA Service Evaluate Liquid Quench versus Recycle Gas Quench Evaluate Options for Revamp of Recycle Gas Scrubber Optimize Conditions for HDS Stripper

23 The optimization studies undertaken, summarized in Figure 13, included the following: 1. The top liquid distribution trays in the existing HDS reactors, originally designed for vacuum gas oil service, were evaluated to determine their suitability for use under the revamp conditions. The flow distribution model predicted the performance of the existing trays would be adequate for the new service and they were not changed. 2. UDS had a new idle reactor. The use of this reactor was evaluated for second stage aromatics saturation service as a possible means of minimizing the capital expenditure. Although the catalyst volume of this reactor was acceptable, the reactor did not have sufficient catalyst beds to handle the large heat release anticipated for HDA service, and the diameter was too large to meet the mass flux criteria. 3. Various alternatives of quenching the large exotherm in the new HDA reactor were studied. These included conventional gas quench options such as increasing the recycle gas compressor capacity and adjusting the temperature profile through the reactor. However, the most efficient and cost effective alternative proved to be the use of hydrotreated diesel from the high-pressure separator bottoms as liquid quench. The highly efficient Topsoe reactor internals in the new reactor made this choice a viable option. 4. The existing amine scrubber was re-piped to treat the overhead gas from the new high pressure inter-stage stripper and had to be evaluated to insure it s suitability for the new service. The existing stripper was modified and the trays were replaced with packing to enable reduction of the H 2 S content in the recycle gas below 5 ppm. 5. All of the existing exchangers in both the reaction and fractionation sections were evaluated and reused. Some re-piping and minor re-tubing of some of the exchanger bundles was required and some new exchangers were needed to complete the new exchange scheme. However, the final layout ultimately provided an additional source of heat to the fractionation section. This significantly reduced the fired heater duty of the revamped unit by about 70% compared to SOR conditions and by about 50% compared to the EOR conditions of the existing unit. As illustrated in Figure 14, the fuel savings associated with the revamped heat integration further helped defer some of the incremental capital expenditure.

24 Figure 14: Fuel Savings Fuel Gas (Absorbed Duty), BTU/hr Before Revamp 86 (Test Run) After Revamp 27 (SOR) 42 (EOR) Revamp Configuration Upon completion of the process studies discussed above, the process configuration of the revamped unit was finalized. The new process flow of the integrated HDS/HDA unit is illustrated in Figure 15.

25 Figure 15: Revamp Configuration of High Pressure Loop Reactor Charge Heater HDS Reactors ake-up Hydrogen Amine Scrubber Recycle Gas Compressor HDS Strip OH Drum LSRGO LCGO LCO V1 E101 E1B-D E1A PC HDS Stripper E102 P4 Water Wash Water Surge Drum Sour Water Effluent to Fractionation Section P1 Charge Pump HDA Reactor E2 Product Separator P101 New Equipment odified Equipment E103 E104 HP Stripper Preheat E1A As seen, all existing equipment in the reaction section was reused, and only a few pieces of equipment required modification. A new inter-stage high-pressure stripper, which used makeup hydrogen to remove hydrogen sulfide and ammonia from the HDS stage reactor effluent, was installed. A new second-stage HDA reactor and associated heat exchange bundles to recover heat from the aromatic saturation reaction were also added. The existing highpressure separator was relocated to the second stage, and replaced by a new first stage separator. Finally, a new low head pump was added to recycle treated diesel product for use as liquid quench to the aromatic saturation reactor. The combined liquid effluent from both first and second stage separators proceeds to the fractionation section. Again, modifications to the existing process flow were minimized as shown in Figure 16.

26 Figure 16: Revamp Configuration of the Fractionation Section New Equipment odified Equipment Disused Equipment SW Fractionator Sour Fuel Gas Wild Naphtha Naphtha Wash Water to Rx Section Stripper A Jet Stripper From Reactor Section A Fract Feed Heater Diesel Stripper Salt Dryer Jet Fuel Diesel Diesel Salt Dryer Of all the existing equipment, only the diesel side-cut stripper was not reused. The existing product and pump-around pumps were reused as spares in jet fuel service. The diesel product heat exchangers were also all reused, by splitting the fractionator bottoms product into two streams through the existing parallel run down lines formally used for the light and heavy diesel products. Revamp Summary In summary, we believe that this was a very successful revamp project. Figure 17: Revamp Summary All Product Quality Objectives were met or exceeded 24 Weeks for Engineering Design Package 9 Weeks for Prerelease of Reactor Shell Design The Total Installed Revamp Unit Cost was < $550/barrel

27 All of the following initial objectives and requirements by UDS were met or exceeded: 1. All of the product quality objectives met or exceeded the specifications required for CARB diesel. 2. UDS s objectives of a fast-track project were met, in as much as the early release of the HDA reactor shell design was completed with 9 weeks and the final Engineering Design Package was completed within 24 weeks from the kick-off date for the start-of-work (including time for data collection and reconciliation for the existing Unit 80). 3. UDS s objective to minimize capital expenditure was realized, in as much as the estimated total incremental installed cost of the revamped unit was less than $550 US per barrel of throughput capacity. Conclusions This is excellent Bang for the Buck. New product specifications for diesel will be introduced in the European Union in year 2005. Over time, the United States, and other nations of the world community, including Asia, are expected to adopt similar specifications, concerning the sulfur content and diesel burning quality. It is expected that the future specifications in Asia will also become more stringent and the Asian refiner will be faced with perhaps even more difficult problems than the EU refiner is faced with today. Reducing the sulfur content from 2500 wppm to 350-500 wppm may be achieved in many existing diesel hydrotreaters by a simple catalyst replacement or by a combination of catalyst replacement and installation of low cost modification to the reactor internals. However, meeting sulfur specifications of 10 wppm to 50 wppm is going to require more than these simple changes. Reduction of the aromatics content and diesel density, improvement of Cetane Number and Cetane Index can be accomplished by hydrotreatment of the feed. The severity of the treatment and the processing configuration will greatly depend on the specifications that are mandated. A two stage unit utilizing base metal catalyst to clean the feed in the HDS stage, followed by an HDA section using a noble metal catalyst offers a flexible cost effective way to meet the most stringent specifications in these areas. Topsoe has commercially proven HDS/HDA technology, and has demonstrated our engineering capability and experience to efficiently apply this technology to either new units or the cost-effective revamp of existing HDS units.