Increased demand for higher-quality fuels and tougher air-quality regulations will substantially alter the specifications for gasoline and diesel fuels in the 1990s.
Higher-octane gasoline with lower vapor pressure, and diesel fuel with lower sulfur and aromatics, will require process changes to produce.
These altered fuel specifications prompted several questions at the most recent National Petroleum Refiners Association annual question and answer session on refining and petrochemical technology held in New Orleans, Oct. 4-6, 1989.
At this important meeting, moderated by Herbert W. Bruch, NPRA's technical director, engineers and managers from refineries and petrochemical plants around the world gather to discuss their operating experiences in the form of questions and answers that cover a wide range of topics.
A panel of selected technical representatives of operating companies and suppliers first answer previously submitted questions, then the audience is invited to add its contributions. This excerpt from the transcript of the meeting examines operating conditions and catalysts that increase motor octane number in FCC units, the effects on processes of lower vapor pressure gasoline specifications, and processing routes to low sulfur and aromatics diesel fuels.
HIGHER GASOLINE MON
Although refiners have achieved FCC gasoline RON levels as high as 94-97, the MON levels have lagged at 81-82. Please comment and quantify, where possible, the effect of the use of specific catalysts, additives, feed selection and operating severity on FCCU gasoline MON. Has anyone attained MON clear levels greater than 83? If so, what is the importance of feedstock quality and boiling range of the gasoline fraction?
MOTT: In terms of catalyst selection, the general strategy involving increasing the motor octane number of the FCC gasoline involves increasing the light C5/C6 iso-paraffin content of the gasoline at the expense of certain undesirable low motor octane normal paraffins and olefins. Generally these are the high carbon number, either unbranched or only singly branched varieties.
The proper catalyst strategy to use depends very strongly on the circumstances specific to the FCC unit in question. Unless these circumstances are taken into account, the motor octane is likely to move in the wrong direction.
In FCC units that are producing relatively low research octane, decreasing the catalyst unit cell size increases both the research and the motor octane of the gasoline. This is the classic response with respect to low unit cell size and low hydrogen transfer. However, close inspection of the two octanes shows that while the motor octane and the research octane both go up, the sensitivity between them actually increases because the motor octane goes up more slowly than the research. This is because the olefins that are produced generally have good research but relatively poor motor octanes. In some particular olefin species there is as much as a 20 number octane difference.
If, on the other hand, you have an FCC unit that is initially achieving high research octane by means of high operating severity, then the octane response due to the same change in catalyst is going to be different. The catalyst, with decreasing hydrogen transfer activity, may not be able to raise the research octane of the system enough to keep the increased octane sensitivity from depressing the motor octane number. In fact, some high severity FCC units that initially had high research octane have lost 1.5 motor octane making this catalyst selection error. Instead of generating olefins, a refiner that is already producing high research octane in the FCC unit needs to switch to a catalyst that can enhance the C5 and C6 iso-paraffins if he wants to increase the motor octane. Small, highly branched iso-structures have both good motor and good research octane numbers.
Other catalyst properties can also influence the motor octane number. High matrix activity increases gasoline olefinicity and increases the octane sensitivity. Therefore, very low A/M catalyst, i.e., zeolite to matrix ratio catalyst, should usually be avoided if the object is to increase motor octane number. When bottoms cracking requirements dictate the need for an active matrix, a catalyst with a balanced Z/M ratio is preferred.
There are also new zeolites like the G series of zeolites that can be used to direct the gasoline composition toward iso-paraffins and away from the lower motor octane number olefins.
Overall, what I am emphasizing is that there is no one catalyst or type of zeolite that will universally raise the motor octane number in every situation. Instead, the specific circumstances of the operation need to be examined so that the proper type of motor octane number increasing technology can be prescribed. Otherwise the refiner may actually lose motor octane. There is a really thorough paper on the subject that was given last year at the NPRA Annual Meeting, "Strategies for Reducing the Gasoline Sensitivity." It was number AM-89-13 by Joanne Deady.
Now for some other parts of the question. In terms of what people have actually seen in the field, I have not seen motor octane numbers higher than 83 on the full range gasoline. We have recently conducted some surveys of North American refineries that revealed the following industry octanes for full range gasoline. Research had a high of 96, a low of 89, and an average of 92. Motor had a high of 83, a low of 78.2, and an average of 80.5.
However, if you fractionate the gasoline into light and heavy naphtha, you can sometimes achieve some very spectacular motor octane numbers depending on the feedstock and operating severity. For example, there is one refiner that is operating with a 965 F. reactor with a very high catalyst to oil ratio. He is achieving an 85 motor octane number and a 96.5 research octane number on the heavy FCC naphtha. This would be at 88 liquid-volume % conversion. The feed used in this example was an Alaskan North Slope blend, and the catalyst being used was 24.23A unit cell size octane catalyst. So feedstock and conversion level both play an important role in determining the composition of the FCC gasoline and its octane.
PARKINSON: Full range FCCU gasoline motor octane numbers greater than 83 have been attained, but we do not consistently achieve these levels. We generally run in the 81 to 82 motor octane number range noted in the question.
MILLER: Our FCC gasoline RON normally averages in the mid 90's and MON in the low 80's.
Rarely do we obtain MON's in the 83 range. I had some engineers look back over the last six months, and I found on one occasion that we had exceeded 83. As far as the feed impacts and operating variables, there are a couple of articles that were published over the last few years. 1.) "Four main FCC factors affect octane," Oil & Gas Journal, Aug. 5, 1985, pp. 91-96. 2.) "Gasoline octane controlled by catalyst selection" Oil & Gas Journal, Aug. 22, 1985, pp. 51-55.
Just below an 83 MON, our experience on factors which impact octane are reactor temperature and feed quality. Octane, of course, goes up with temperature. When we switch our refinery crude slate from a naphthenic to a more paraffinic, the FCC feed follows suit, and the gasoline octanes drop. We treat most of our feed to the FCCU in a high pressure hydrotreater. Increased hydrotreater severity increases aromatic saturation. This leads to reduction in FCC gasoline octanes. We have done some fractional distillation and octane studies of our FCC full range gasolines. The component octanes are high on the front end, high on the tail end, and low in the middle. The change in cut points is not in our favor to improve octanes. This will not necessarily be the case for all refiners.
I will give some brief comments for those who are considering reactor modifications or improvements. These are rules of thumb: When you drop reactor riser time, the octanes tend to increase. In the opposite direction, when you improve the contact of catalyst with the oil, the octanes tend to decrease.
COMEAU: An article published in Oil & Gas Journal, Oct. 31, 1988, entitled, "Advanced zeolites used in FCC catalysts boost motor octane number" contains information from literature searches about the effect of operating severity and feedstock quality on FCC gasoline motor octane numbers. Very severe operating conditions on a very aromatic low concarbon feed could elevate RON's high enough to obtain 83 MON'S. As a rule of thumb, olefins have a sensitivity of 13, where aromatic sensitivity is around 10. We have operated with RON's of 96 and motor octane numbers of 82.5. Fractional analysis of our FCC gasoline would indicate that the front 20% and the last 10% of the FCC gasoline would achieve the 83 motor octane on a continuous basis.
DONALD A. KEYWORTH(Akzo Chemicals Inc.): We had reported to us by one of our customers an 83.6 which is one of the highest motor octane numbers we have seen for commercial FCC production. The feed was highly naphthenic, and the operating severity was normal. The gasoline contained more aromatics, and we attributed the higher motor octane number to that. It was a full range gasoline with an endpoint of 430 F. The catalyst was of the low nonframework alumina zeolite type that was rare earth promoted. It had a moderate amount of active matrix.
LOWER SUMMER Rvp EFFECTS
What processing changes have you seen and do you foresee as a result of the legislation lowering gasoline Rvp? How much Rvp giveaway will be necessary to stay in compliance?
BIGGS: Additional octane production was required at our plant, in particular an FCCU revamp, to make up roughly a 0.2 octane number loss due to the change in Rvp. This is primarily because we were octane short during the summer. We have seen additions of isomerization units at other plants for the same reason and the shutdown of one gas plant in our area because of excess butane supply. Over the summer months we have averaged about a 0.7 octane giveaway on the Rvp of our gasoline.
COMEAU: Our refineries have reduced light condensate charge stocks, improved fractionation of the LPG's in their gas plants, implemented reformate Rvp control, exported and stored butanes, sold butanes to steam crackers, defluorinated alkylation unit butane for sales, alkylated amylenes, and sold more refinery grade propylene to allow more low Rvp, high-octane alkylate production. The use of a Herzog and Setavap Rvp analyzer has allowed blending of the gasoline to within 0.3 psi Rvp of the pipeline or EPA specifications.
FISCHER: The initial phase of EPA's Rvp reduction program has had little impact on Champlin processing. Reduction to the 9.0 psi Rvp summer specification did not reduce butane blending at Champlin to the same degree as it did at many other refineries. Due to the high fraction of reformate in our gasoline pool, we have normally been limited by the V/L volatility specification rather than Rvp. Since the V/L specification was not lowered commensurately with Rvp, the impact of the change in regulation at Champlin was less than the indicated change in Rvp. As a result, Champlin has remained a small net buyer of butane, while many other referies were producing butane in excess of blending requirements.
The proposed reduction to 7.8 psi Rvp in 1992 will substantially impact both blending and processing options. At the 7.8 psi specification, very little makeup butane will be required, making Rvp control difficult, and greatly constraining blending options. Better fractionation will be important to maintain blending flexibility.
A number of processing options may become economical in the 1992 context. Depending on an area's ethylene plant capacity, the summer price of normal butane will decline to either its steam cracking or fuel value. Sales of normal butane to ethylene plants will increase. Normal butane conversion to either isobutane or various types of butylenes is possible. Other methods of reducing gasoline pool vapor pressure include alkylation of amylenes and FCC catalyst selection.
Precision of the ASTM Rvp measurement method is 0.5 psi Rvp. Since EPA has stated that any analysis in excess of specification will be regarded in violation, the blending target must be 0.5 lower than the specification.
MILLER: We have planned or plan to upgrade our debutanizers and stabilizer to reduce the Rvp of major gasoline blending components streams. This results in the obvious loss of octane barrels and increased production of butanes. The octane loss will be made up by increased severity at reformers. The extra butane will be fueled or sold in the short term. In the long term, we plan on utilizing it as hydrogen plant feed. We do not anticipate any change in Rvp giveaway over present summer blending operations.
WELCH: This past year our biggest problem has been to keep the refineries in butane balance. It was resolved at one refinery by simply adjusting operation to the butane splitter tower to minimize the normal butane content of the tower bottoms, a stream which normally goes to gasoline blending. Another refinery had to change to a more gasoline-selective catalyst in the FCC unit to reduce the amount of LPG and olefins they were producing. Currently we target for 0.5 psi lower than the maximum allowable Rvp, and have averaged 0.11 psi lower than the target.
Which methods are available to adjust FCC unit operations to accommodate reduced gasoline Rvp limits? Please discuss catalyst and operating variables.
WELCH: The FCCU gas concentration unit debutanizer is used to control the Rvp of FCCU gasoline. If the removal of all of the butanes and butylenes from the gasoline is restricted by alkylation unit limitations, then a switch to a more gasoline-selective catalyst should be considered.
MOTT: I am thinking basically along the same lines. Except that I can add some facts about the response of the normal C4 yield to operating variables. If you try and disassociate normal C4 yield from conversion by using tricks with reactor temperature or tricks with feed preheat or tricks with catalyst-to-oil ratio, what you find is basically that you are not able to disassociate that normal C4 yield versus conversion relationship. So what that means is that if you are forced into the position of having to reduce your normal C4 yield, then you are going to have to go back and look at the catalyst design and the catalyst composition because there are some very dramatic improvements that you can make by modifying the catalyst to shift the yields back into the gasoline and away from LPG.
MILLER: Our experience on low Rvp is that the catalyst or the yields are not a major concern. With our normal yields, we are able to control the Rvp down to 5.5 psi. This, of course, takes a debutanizer that does a decent job of fractionation. We get about 0.2% volume butanes in our full range gasoline. Lowering it further would mean removal of pentanes which could be economically unattractive. At that point some catalyst modifications may be warranted.
LIOLIOS: Economic studies have shown that deeply debutanizing the FCC gasoline and taking approximately 2 to 5% of the C5's to the alkylation unit can be beneficial assuming you have alkylation capacity.
FISCHER: We basically agree with the previous comments. We also have used amylene alkylation as a way to help manage reduced Rvp requirements in gasoline.
L.J. McPHERSON (BP Oil International Ltd.): C,5 alkylation has been mentioned as a way of reducing the Rvp of the catalytic naphtha. A lower cost option to alkylation is to convert the light catalytic naphtha iso-olefins to ethers. Etherification is more selective to the iso-amylene, which is one of the more volatile components in the catalytic naphtha.
WILLIAM A. KELLY (Katalistiks-UOP): We agree, higher gasoline selectivity is the answer. Higher gasoline selectivity can be obtained by optimizing catalyst zeolite-to-matrix ratio, increasing catalyst rare earth content, or lowering reactor temperature. A typical optimum FCCU operating target is to maximize alkylation plant rate with zero excess C4's. Katalistiks' 1 7210 zeolites in Alpha and Beta catalysts have produced higher percent isobutane yields and lower percent normal butanes yields which keeps alkylation plant rate high while minimizing normal C4's to gasoline blending, Without trying to sound like a catalyst vendor, increasing catalyst MAT while reducing reactor temperature should increase gasoline selectivity at constant conversion, although at an octane penalty. This requires slightly higher fresh catalyst additions.
ROUTES TO LOW-SULFUR, LOW-AROMATICS DIESEL
What HDT processing conditions (i.e., LHS Y, total pressure, hydrogen circulation and furnace temperature range) are required for treating 1.0 wt % S light cycle oil down to the 0.05 wt % S level over an extended period of time, such as one year?
SHIFLETT: I think the question here is making an assumption that all light cycle oils (LCO) are the same, and that is strictly not the case at all.
The nature of the sulfur compounds and their difficulty in removal really depend on, in part, the severity of the FCCU operation. Typical FCC's feeding mainly VGO's yield LCO's having ratios of two different kinds of sulfur compounds: benzothiophenes and dibenzothiophenes, with ratios somewhere in the range of 3-1 to 5-1.
The reactivity, or the ease in which the sulfur can be removed from these compounds, is dramatically different, with the dibenzothiophene being much more difficult. In the case of heavy oil crackers, or FCC's feeding significant resid, this ratio changes. You will have something more like 2 to 1 benzothiophene type compounds to dibenzothiophene type. That changes your HDS reactivity, and thus you have really a different feed to deal with.
Having muddied around with this, I will go ahead and give a ballpark range for processing a West Coast LCO down to 0.05% sulfur. It is based on data given in an Oil & Gas Journal article from April 13, 1987. At a pressure of 750 psi, with a maximum reactor temperature of around 750 F., circulations of 500-1,000 scf/bbl, space velocity of 1.6 to 1.8; you should be able to do it easily. I might comment that another rule of thumb, if your reactor temperature is limited to less than 750 F., then drop the space velocity by 20% for every 10 F. you have to go down. That will give you a rough idea of what you might be able to do.
FISCHER: Generally, the operating conditions to do this job would be 500 to 900 psig total reactor pressure, reactor average bed temperature of 650 to 700 F. from start-of-run to end-of-run, a liquid hourly space velocity of 3.0 to 4.0, and hydrogen-to-oil ratio at the reactor inlet of at least three times the maximum hydrogen consumption. Hydrogen consumption would be about 300 to 250 scf/bbl over a high activity nickel/moly catalyst. The recycle gas would need to be amine treated for removal of hydrogen sulfide.
A word of caution. Too high of a reactor temperature, i.e., above 7150 F. might cause color and stability degradation of the No. 2 fuel oil or diesel. Also, the gravity improvement of the light cycle oil would not be enough to meet the 30 API gravity specification if API gravity of the untreated catalytic cycle oil is below 20. Some blending with straight-run distillate would be required. A low cetane number product significantly below 40 would also be produced after hydrotreatment at these conditions.
EUBANKS: The only thing that I want to add on this is that there will be very little aromatic saturation that will occur at these conditions.
ROBERT S. HENDERSON (Unocal Science and Technology Division): For this simple desulfurization operation, Unocal would use a cobalt moly catalyst, set the pressure level at about 600 psi; for a 1 year cycle length. The space velocity would be determined to give an average reactor temperature In the 650 to 660 F. range. For a typical LCO, from a gas oil operation, these space velocities might be in the 2 to 3 range. At this conditions, we agree, there would be very little decrease in aromatic content.
MEHMET Y. ASIM (Akzo Chemicals Inc.): I just want to add that about 2 years ago, we performed a series of pilot plant studies for 0.05% S in diesel. One feed we processed was an LCO stream which also contained 1 wt % S. This LCO was hydrotreated to 0.05 wt % S at 400, 800, and 1,200 psig hydrogen partial pressures. The results of this test program were presented at the 1987 NPRA Annual Meeting (Paper Am-87-59). We found that at about 400 psig hydrogen partial pressure, at 685 F. bed temperature, and at 2.0 space velocity with 1,500 scf/bbl hydrogen rate, 0.05% S level could be achieved for
What routes are feasible to reduce aromatics in diesel fuel? What methods can be used to measure the aromatics level?
NEGIN: Hydrogenation is one route to reduce aromatics. The difficulties of saturating aromatics were covered in detail during the 1987 (See Question III-21) and 1988 (See Questions III-22 and 23) NPRA O&A Session transcripts. Basically, the consensus is that existing diesel hydrotreaters are not capable of aromatics saturation. In order to saturate the aromatics, you must go to most severe conditions of higher pressure and more active, possibly noble metal catalysts.
Extraction is another route. This is discussed under aromatics extraction of this year's Q&A Session. It was also covered during the 1987 NPRA Q&A Session (refer to Question IV-42). In that session, the experience of one refiner was discussed. They extracted an LCO and achieved a 70% yield of diesel with a 10 cetane number increase,
SHIFLETT: There are a number of hydroprocessing routes that you can use to get to where you have to be on diesel aromatics. However, all of them are going to require that you have sufficiently high pressure to do the aromatics removal. In the case of a high sulfur environment, a high powered nickel moly catalyst is appropriate and economical for doing a fair amount of that, but there are limits to where you can go.
Beyond this case, in needing to get to deeper degrees of saturation, one could conceivably look at a two rector (or more) system in which H2S is removed between the reactors. The approach here would be to beat the heck out of it with a high activity cobalt moly catalyst first. An example would be the 448 that we offer, which we have nicknamed around the office as "diesel cat." This would do a good job of getting that sulfur down so that you could use a nickel tungsten hydrogenation catalyst at a higher temperature and pressure to finish the job.
For even deeper saturation, there is a zeolite base material that is tolerant of both ammonia and H2 that also looks quite promising. We are doing some work in that area which we will be reporting at a later time. The ultimate, of course, would be to go to a noble metal formulation where you have to remove virtually all of the sulfur, but you can do a fantastic job on getting aromatics down. Indeed, these catalysts have been around for a while for specialty saturation of things such as specialty solvents and dry cleaner fluid.
MILLER: We have looked at a number of routes for producing 10% aromatics in diesel fuel. The hydrotreating routes have been discussed. I have a brief comment on extraction. You have to find a home for the aromatics extracted. Another option is to just go ahead and hydrocrack the diesel. This may mean going out of the diesel business, and making gasoline out of it.
Some of the lower cost options are as follows. We have evaluated some 200 crudes around the world, and I think there are about four of them that can make diesel with acceptable aromatic levels. We suspect the demand will be high for those crudes. Distillate cutpoint optimization is also something else that needs to be looked at. The trend to make the diesel will probably be to cut the endpoints, and put in as much naphtha as you can up to the maximum flashpoint.
In California, low aromatic diesel regulations were recently adopted. We are already under a mandate to go to 10% aromatics in diesel fuel. The regulations do provide an exemption; wherefore, if you can come up with an equivalent fuel blend, or in combination with additives to achieve equivalent emissions of a standard test fuel, then you can certify that blend. The specifications for the standard test fuel are essentially a premium grade of jet fuel with 10% aromatics. One hope is that some of the chemical manufacturers could come up with a combustion promoter additive. This would be much cheaper than the mechanical options.
As far as methods of measurement, a number of methods of measuring aromatics in diesel have been described in the literature, all of which have better accuracy and reproducibility than the FIA test method, ASTM 1319, which has been mandated by California Air Board. Ultraviolet, mass spectrometry, high pressure liquid chromatograph (HPLC) have all shown to give better results than FIA.
FISCHER: Reduction of aromatics in diesel fuel is feasible by either liquid solvent extraction or severe hydroprocessing. The solvent extraction process is generally less attractive due to the lower diesel yield and disposal of the aromatic byproduct. Severe hydroprocessing can be done in either a one stage or two stage catalytic process at reactor hydrogen pressures above 1,500 psig and reactor average bed temperatures of 650 to 700 F. Liquid hourly space velocities of 0.5 to 1.5 are required in the saturation stage which could be either nickel moly or tungsten moly on alumina, depending on the sulfur level of the feed and the depth of aromatic saturation required.
The hydrogen to oil ratio in the reactor inlet should be about three times maximum hydrogen consumption. Hydrogen consumption can be as high as 1,200 scf/bbl. Aromatic saturation requires 26 scf/bbl per every 1 vol % removal of aromatics.
This leads to the second question on measurement. The two methods we are aware of are FIA and UV adsorption.
LOUIS J. SCOTTI (Lummus Crest Inc.): The approach mentioned by several of the panelists, namely two stage hydrogenation, has been used in our commercial Arosat unit at Neste Oy (Finland). This unit was designed to process solvents and produce a very low aromatic specification. The unit uses a noble metal catalyst in the second hydrogenation stage. The refiner has processed diesel in this unit and achieved very low aromatics content.
DAVID Y. LAW (Chevron Research Co.): Chevron Research suggests two routes for diesel aromatics reduction: (1) high pressure hydrotreating in one or two stages using noble or base metal catalyst, and (2) once-through Iso-cracking which can yield diesel with very low aromatics content (i.e., < 10%) throughout the cycle.
As for the second part of the question, aromatics levels can be measured using mass spectrometry, extended FIA, and liquid chromatography.
Is anyone currently solvent extracting FCC and/or coker light gas oils for aromatic reduction and cetane improvement for inclusion of the raffinate stream in the refinery diesel pool? How is the extract utilized?
NEGIN: SO2 extraction of virgin and cracked light gas oils for cetane number and smoke point improvement was fairly common practice 30-40 years ago, before the advent of modern hydrotreating technology. The aromatic extract was frequently sold as tractor fuel, low octane high boiling gasoline, or blended off into various stove oil and low quality distillate fuels that did not require high smoke points.
Some SO2 extraction may still be practiced to produce very high smoke point, illuminating kerosine, or minimum 50 cetane diesel. Such high smoke points and cetane numbers are frequently not achievable with some feed stocks via hydrotreating, since hydrotreating saturates aromatics to naphthenes which have smoke points and cetane numbers intermediate between the very high paraffins and the very low aromatics. The rings would have to be broken via hydrocracking to make a significant improvement in cetane number.
Copyright 1990 Oil & Gas Journal. All Rights Reserved.