Arthur J. Suchanek
Criterion Catalyst Co. L.P.
Houston
Producing low-aromatics diesel fuel may be accomplished catalytically rather than through large capital investment.
Pilot-plant test results show that some catalyst types and catalyst systems such as various combinations of nickel-molybdenum (NiMo) and nickel-tungsten (NiW) on alumina in the presence of sulfur; noble metal on alumina for low-sulfur feeds; and noble metal on zeolite for moderate sulfur feeds can achieve the proposed tougher specifications on diesel aromatics at minimum capital investment.
For the past several years, we have been hearing that highway diesel sulfur levels may be reduced to as low as 0.05 wt % and aromatics to as low as 10 vol % across the U.S. In fact, some areas of California already market diesel fuels that meet these very strict regulations.
For the refining industry, the reduction of sulfur and aromatics will be expensive. Numerous studies have already been done or are being done. Results of those studies indicate that very high levels of capital expenditures will be necessary to reduce sulfur and aromatics in diesel fuel.
The potential for increased costs of diesel production make it necessary to determine both the technical and economical means of reducing sulfur and aromatics in diesel fuels.
Criterion concluded earlier that use of a high activity aromatics catalyst system under a particularly severe test case assessed (i.e., feed, operating conditions, product goals) that a great deal of capital would most likely be needed to achieve the proposed aromatics specifications.1
This article carries that analysis a step further in light of some novel catalyst systems. There are many approaches that will allow refiners to solve the aromatics reduction problem catalytically, rather than being forced to spend major capital.
TECHNICAL OVERVIEW
Technically, at a given pressure level and adequate hydrogen treat-gas rate, sulfur removal is directly related to the amount of catalyst in the reactor and temperature.
High activity cobalt molybdenum (CoMo) catalysts achieve required sulfur levels with total pressure as low as 500 psig.
Aromatic reduction, however, is on the other side of the toughness curve compared to sulfur reduction. Whether by saturation or hydrocracking, aromatic reduction is strongly dependent on some tough rules of nature.
Pressure, catalyst type, and an understanding of the interaction of these variables on chemistry and thermodynamic equilibria are necessary to determine how best to handle the aromatics present in the refiner's available diesel-blending streams. Then, by knowing the product specifications, stream properties, limits of existing equipment, and available technical approaches, a sound economic solution to aromatic reduction can be selected.
We will discuss the chemistry of the diesel boiling range materials, identify some problem streams to the refiner, and cite ways that we, as a catalyst company, can economically help reduce aromatics and improve the quality of future diesel fuels.
However, the future aromatic specifications have not yet been pinpointed and, in fact, are still being discussed.
THE CHEMISTRY
The boiling range of diesel-blending components generally falls between 400 and 680 F., with plus or minus some front or back end. The molecular structure is generally in the C12-C25 or so range and is comprised of aromatics, naphthenes, iso and n-paraffins, olefins, or molecular combinations of them all. But because of the very wide array of streams available for blending into the ultimate fuel, the proportion of the various molecules varies widely.
In a discussion of aromatics reduction one also needs to associate aromatic content and cetane number because cetane number is the accepted measure of diesel quality.
Cetane quality is highly dependent on the paraffinicity of molecular structures, whether they are straight-chain or alkyl attachments to rings. Therefore, a stream which is mostly aromatic rings with few or no alkyl side chains would be a low cetane quality material, and a highly paraffinic stream would be high in cetane quality.
An FCC cycle oil fits the former description quite well and, because it is highly aromatic, it is also very hydrogen deficient. Naphthene molecules, which are saturated aromatics, have a somewhat higher cetane potential, but still not overwhelming.
For years, the author has used a "molecular thought pattern" to look at diesel boiling range or any other boiling range streams (Fig. 1).
By studying Fig. 1, much can be seen about the chemical nature of diesel fuels. Note the highly condensed aromatic structures with the very low cetane quality.
If the refiner simply saturates the aromatic rings, then the resulting product reaches a 40-41 cetane number level and one immediately wonders it aromatics saturation of cycle oils is advisable. Past work has shown that it would take high pressure, and possibly two-stage processing, to saturate 60% or more of the aromatics.1
At this point in the decision analysis, the refiner would then be forced to seriously consider a hydrocracker with the same or marginally higher investment. A cycle oil hydrocracker, for instance, would produce not only good quality diesel, but also high quality jet fuel.
Another consideration for highly aromatic cycle oils would be blending the cycle oil with virgin straight run, thus improving the cetane potential via dilution with better molecules. Subsequent processing of the stream to low sulfur with accompanying marginal aromatic conversion would then be achieved at lower severity. Thus, understanding the chemistry can hopefully lead the refiner to solid economic choices for diesel component upgrading.
CETANE/AROMATICS
There is a somewhat confusing relationship between cetane number, cetane index, fluorescent indicator adsorption (FIA) aromatics, and aromatics measured by many other techniques. Cetane quality measurements, resulting in either cetane number or cetane index, and aromatics analysis by any means, should be complementary, but they are not. Therefore, we manage with what we have while understanding the drawbacks of each method.
While technically studying finite incremental changes in aromatic hydroprocessing, it is important to watch the trail of aromatics removal or saturation as it progresses from multi-ring molecular structures to mono-ring structures, and finally to naphthene. It is equally important to also watch the alkyl or naphthenic appendages as they directly affect the cetane number of the product.
To help follow this trail, we must spend a little more time discussing aromatics analysis. In an NPRA paper last year, Richard Nash discussed aromatics analysis and cited the following derived equation for relating FIA aromatics analysis and UV mono-aromatics analysis for the particular feed in his study:1
FIA aromatics =
4.7(mono-aromatics) - 11.6
While we use various sophisticated analytical determinations in research and development work, the FIA trail best fits the logic flow of this article. This is necessary because the FIA analysis, even with all of its problems, is still solidly entrenched as the accepted test for diesel aromatics and, by correlation, cetane quality.
Fig. 2 illustrates a generalized correlation of cetane number vs. FIA aromatics content for a refinery mixture.2 Our thanks to George Unzelman for this relation.
While not meant to be precise, this relation aids in the overall thought pattern for understanding the means to reduce aromatics in diesel fuels.
Another interesting technical paper has been presented by Syncrude Canada Ltd., relating cetane quality to its fluid coker-based operation on bitumen feed.3
TODAY'S DIESEL
Looking at the worldwide quality of diesel today, quite a large variation in its molecular makeup is evident. Most of the U.S. diesel meets a specification minimum cetane number of 40.
However, in the European Economic Community (ECC), the cetane number is in the high forties and refiners there meet this without any problems. The reason for this discrepancy can be correlated to U.S. conversion capacity.
It has been stated many times that U.S. cetane quality is directly related to consumption of gasoline and, thus, the high dependence on FCC and coking units.4 As a result, FCC cycle oils and coker gas oils constitute a high proportion of the streams available for ultimate blending into diesel fuels.
The technical problems have been compounded even further by making diesel molecules even tougher and more difficult to treat by processing at high contact times and high temperatures in U.S. heavy-oil upgrading processes. Fundamentally then, the streams available for hydroprocessing and ultimate blending into diesel can be characterized as low in hydrogen and high in sulfur, nitrogen, and aromatics.
REFINERY STREAMS
Table 1 illustrates six streams that a refiner will consider for ultimate hydrotreatment and blending into the diesel pool.
The molecular thought pattern can be used to determine how to best bend, reshape, or break the molecules to meet the overall objective of improving diesel quality. And then, with interpolation and extrapolation, the correct catalytic and/or process route can, it is hoped, become evident.
A point to reiterate here is that a major aim of the technical approach is to determine catalysts or technologies which can help the refiner fit existing refinery equipment, with added investment recommended only as absolutely needed. This subject has been covered in a previous publication.5
CATALYTIC APPROACHES
For the streams of interest, there are various catalysts that can be used in varying ways to achieve overall reduction in diesel fuel aromatics. Each catalyst has, however, restrictions on where and how it can be used. Table 2 shows Criterion Catalyst Co. catalysts and their areas of application.
Generally, NiMo catalysts, such as C-411 and C-424, are used for contaminant heteroatom (N,S,O) removal, but they also saturate some aromatics. The effectiveness of NiMo catalysts is highly dependent upon the system hydrogen pressure.
While it is true that hydrocracking is the ultimate sledgehammer for aromatics reduction of diesel boiling range streams, the reality is that these units carry a high investment cost. For most streams, a catalyst system that would achieve moderate aromatics saturation should suffice.
The catalyst used should be strong, not only for aromatic saturation, but also for hydrodesulfurization (HDS) and hydrodenitrification (HDN). An NiW on alumina catalyst can fit this job description quite well.
Still, in many situations hydrocracking is desirable and must be considered for refining systems laden with aromatics (particularly for FCC cycle oil). There are many operating hydrocrackers achieving high yields of jet and diesel from FCC cycle oils. Several NiMo first-stage catalysts, as well as second-stage zeolite-based hydrocracking catalysts, are available for this type of processing operation. Table 2 cites some information on these hydrocracking catalysts.
Finally, noble metal catalysts are well known to be very high in activity for saturation of aromatics, but conventional wisdom teaches us they are easily poisoned by sulfur and nitrogen. For example, noble-metal on silicaalumina catalyst is excellent for diesel range streams of very low sulfur-less than 10 PPM.
More exciting, however, is a new noble metal on zeolite catalyst (704A). This catalyst represents a significant improvement in sulfur and nitrogen tolerance to about 1,200 ppm and 200 ppm, respectively.
PROCESSING LOGIC
Straight-run streams or blends of straight-run with coker and FCC cycle oils are presently treated to sulfur specifications of 0.05-0.2 wt % in units with at least 500-600 psig total pressures and liquid hourly space velocities (LHSV's) of at least 1.0 hr 1, but most likely 3.0 hr-1 or higher. It would be nice if the product from this normal hydrotreater could be further hydrogenated to reduce its aromatic content.
It would be even nicer if it could be done in the same hydrogen circuit. Better yet, the ultimate would be single-stage processing in the same hydrogen loop, which certainly should be an economically attractive route.
The problem with this scenario is that high pressures (approximately 1,500 psi H2) are needed to achieve significant aromatic reduction via saturation to greater than 60-65%. Last year's Criterion paper showed that a 60% aromatic stream could be reduced to the 10-20% aromatics range with two-stage processing.
However, units such as this are not usually available and would carry a high investment cost. The capital cost for the saturation unit would probably be similar to the investment for a hydrocracker, yet a hydrocracker would net higher value products.
Thus, as mentioned, high-level aromatic saturation of high aromatic content streams is probably not the optimum means of cetane improvement of diesel fuel. The FCC cycle oil and hydrotreated blend cited in Table 1 are examples that fit this situation. Overall, the refiner would like to determine what can be done in an existing refinery reaction system or hydrogen loop before considering high investment approaches.
AROMATICS SATURATION TESTS
Pilot studies were conducted on three feedstocks to define aromatics reduction via saturation:
0.16% sulfur, 58% aromatics
1% sulfur, 58% aromatics
0.5% sulfur, 40% aromatics
These studies used varying LHSV and pressures ranging from 700-1,500 psi hydrogen. This approach views reducing LHSV at lower constant pressure as a refiner would achieve by adding parallel or series reactors to an existing reactor system.
The catalyst system used in our studies is a stacked-bed system with the proportion of various catalysts somewhat optimized to achieve the highest saturation activity with minimal reactor fill cost. The term "somewhat optimized" means that pilot studies and commercial evaluations are not yet complete.
In processing a feedstock defined in Table 1 as a hydrotreated blend of FCC cycle oil, coker gas oil, and straight-run distillate, this diesel boiling range material was doped by adding sulfur back to the 0.16 wt % level, nearly typical for units operating today.
The studies were done at 700 F. reactor temperature; 700, 1,100, and 1,500 psi partial pressure, and at LHSV's that varied between 0.4 and 1.0 hr-1 at the lowest (700 psi) pressure. Fig. 3 shows, as expected, that at constant throughput (LHSV), the pressure increase from 700 to 1,500 psi increases aromatic relative conversion from the 0-5% range to 50-55%.
Fig. 4, for the same feedstock, shows that at the lower hydrogen pressure of 700 psi, increased catalyst volume via LHSV reduction to 0.3-0.4 hr-1 increases the FIA aromatic conversion to the 15% range. The key observation here is that low-pressure processing can achieve reasonable aromatics reduction by using increased catalyst volume within the existing hydrotreater reactor system.
One might envision this as first-stage processing to a normal sulfur level product, and then a slip stream processed in the same hydrogen loop at low throughput for incremental aromatics reduction. Hydrogen circulation could be separate or cascaded.
The preceding scheme covers processing of an already hydrotreated refinery blend, thus the scheme is two-stage overall. Single-stage processing with both heteroatom contaminant removal and aromatic saturation at the same time would obviously be more desirable cost-wise. Therefore, the feedstock was further doped to a higher 1.0 wt % sulfur level.
This is where the results get interesting. The study pattern was the same, initially scouting hydrogen pressure between 700 and 1,500 psi. Fig. 5 shows similar performance at the higher pressure levels for the higher sulfur feed.
However, about 10% FIA relative reduction was found at the low 700 psi pressure, compared to less than 5% for the low-sulfur stock. As shown in Fig. 6, as LHSV was decreased at 700 psi, the performance improved even further, and 31% relative FIA reduction was obtained at 0.36 hr-1 LHSV, compared to 15% with the low-sulfur feed.
The difference in the reactivity of the two feeds is highlighted in Fig. 7. Surprisingly, the catalyst system performance improved with higher feed sulfur at low pressure and low throughput.
For further verification of the applicability of single-stage processing, we conducted an identical study of an actual refinery blend, also defined in Table 1. This feed is a California refinery blend of virgin and coker distillates containing 0.5 wt % sulfur and 40 vol % aromatics.
Figs. 8 and 9 show even more dramatic results because greater than 80% FIA aromatics relative conversion was achieved at 1,500 psi and 1.0 hr-1 LHSV, and more than 40% FIA aromatics reduction was obtained at 700 psi and 0.3 LHSV,
Ongoing temperature optimization studies at the same pressure and space velocity conditions are showing increases in FIA aromatics conversion to the 55% range. In addition, we have further confirmed the FIA aromatics conversion with results from other analysis techniques.
Therefore, for this California blend the operation at 700 psi hydrogen pressure and 0.3 hr-1 LHSV will have reduced the FIA aromatics to less than 20% and increased its cetane number to the high 40's. These are interesting data that point to a high probability of low investment cost solution to diesel aromatics reduction.
The studies are continuing with the aging of the catalyst system and the gathering of more data on the effects of operating variables.
HYDROCRACKED AND MILD HYDROCRACKED PRODUCTS
There are two streams produced in some plants which contain low levels of contaminants or virtually no contaminants. Hydrocracking yields diesel boiling range products which contain virtually no significant heteroatom contaminants.
On the other hand, mild hydrocracking of vacuum gas oils or mild hydrocracking of good quality atmospheric residues is typically achieved by operating normal HDS/HDN catalysts at higher temperatures. These operations simply increase hydrogen consumption by severe heteroatom removal and polynuclear aromatics conversion.
This is usually accomplished with long contact times at high temperatures, resulting in thermal degradation of the product molecules, particularly those in the diesel boiling range. As mentioned, the alkyl side chains on ring structures are important to cetane quality.
These side-chains are grossly affected by severe conditions, resulting in tough molecules. Mild hydrocracked products are a lot like this, varying in aromatic makeup because of crude source and upstream processing severity.
However, they nearly always contain enough sulfur and nitrogen to make the aromatics difficult to saturate, especially with low operating pressure and noble-metal catalysts. Table 1 also lists the properties of typical HC and MHC distillate products.
A noble-metal on silica-alumina catalyst, such as C-614, fits aromatic saturation requirements for the hydrocracker product and, in fact, is used in numerous commercial units for smoke point or cetane improvement, or related applications. The operating conditions required are mild, generally low pressure, in the 600 psig range, and LHSV's in the 2-3 hr range.
The key is that the feed sulfur and nitrogen must be less than 10 ppm, although further studies are planned to determine if higher contaminant levels can be tolerated.
For the diesel boiling range, thorough testing is in progress for the noble metal on zeolite catalyst 704A used alone and in combination with other catalysts to zero in on its range of applicability. Studies so far have determined that these acceptable feed levels are 200 ppm nitrogen and 1,200 ppm sulfur in the diesel boiling range.
Table 3 highlights the extremes of the catalyst testing on diesel streams. While the effect of contaminants is overcome with increased operating pressure, the catalyst activity is such that a high degree of aromatics saturation is attained with relatively low temperatures.
These data seem to point to a processing window which may be two-stage, but possibly at markedly lower pressures than obtainable with other aromatic saturation catalysts.
Work is continuing to better define this processing window.
Understanding the interactions between feedstock properties, catalyst capabilities, and equipment requirements is critical. Many applications can avoid big capital expenditures by judicious choice of catalysts and systems of catalysts used in low to moderate-pressure hydrotreating units.
ACKNOWLEDGMENT
The author thanks Opinder Bhan for his contribution to this work.
REFERENCES
- Nash, Richard M., "Meeting the Challenge of Low Aromatics Diesel," NPRA annual meeting, San Francisco, March 1989.
- Unzelman, George H., "Higher diesel quality would constrict refining," OGJ, June 29, 1987, p. 55.
- Yui, S. M., and Sanford, E. C., "Kinetics of Aromatics Hydrogenation and Prediction of Cetane Number of Synthetic Distillates," API 50th Midyear Refining Meeting, Kansas City, May 1985.
- McPherson, L. J., and Bourgeais, P., "Des Produits de Qualite? Le Craquage Catalytique, Characterization and Utilization of FCC Light Cycle Oil," Association Francaise de Techniciens du Petrole, Paris, April 1989.
- Suchanek, A. J., "Refiners must fit chemistry to the pipes," OGJ, Dec. 17, 1984, p. 115.
- Johnson, Alan D., "The Effects of Hydrotreating on Diesel Fuel Quality," API 48th Mid-Year Refining Meeting, Los Angeles, May 1983.
Copyright 1990 Oil & Gas Journal. All Rights Reserved.
