NPRA O&A-2 CRACKING PROCESSES DEPEND ON GOOD CATALYST CHARACTERISTICS

April 2, 1990
Good catalyst characteristics are necessary to maintain efficient operation of two key conversion processes: fluid catalytic cracking (FCC) and hydrocracking. How catalyst types and loadings affect conversion process operation, and how certain process operating conditions affect the catalysts in them, were important topics at the most recent National Petroleum Refiners Association annual question and answer session, Oct. 46, 1989, in New Orleans.

Good catalyst characteristics are necessary to maintain efficient operation of two key conversion processes: fluid catalytic cracking (FCC) and hydrocracking. How catalyst types and loadings affect conversion process operation, and how certain process operating conditions affect the catalysts in them, were important topics at the most recent National Petroleum Refiners Association annual question and answer session, Oct. 46, 1989, in New Orleans.

This excerpt from the transcript of the meeting covers criteria for the selection of FCC combustion promoters, the effect of FCC catalyst zeolite content on catalyst attrition and selectivity, and the effects of magnesium buildup on catalyst. Operating variables that affect catalyst life in hydrocracking operations are also examined, along with some refineries' experiences with the disposal of hydrocracking catalysts.

For more information on this important meeting and its format, see OGJ, Feb. 26, 1990, p. 83.

FLUID CATALYTIC CRACKING

In a high efficiency complete combustion regenerator, is it economical to use a fully factory-promoted catalyst? With a non-promoted FCC catalyst, we are experiencing a high rate of CO promoter addition for complete combustion. What are the major criteria being used in selecting CO promoters, apart from their platinum content?

FISCHER: The economics of factory-promoted catalyst are variable due to the relative effectiveness of promoters and the different prices vendors charge for their products. In a recent UOP survey of high efficiency units, it was found that nearly two thirds of the respondents added promoter separately. It was also found that factory-promoted catalysts use more platinum to achieve the same degree of afterburn control. We have found that it is generally more economical to add promoter on an as-needed basis.

Apart from platinum content, the main criterion we have used in the past in selecting promoters is cost effectiveness. This is determined by unit trials. However, we do look at other promoter properties, including density, hardness, and particle size distribution. We try to match those to the equilibrium catalyst analysis so that the promoter stays in the unit. Another caution is that if you are withdrawing catalyst for activity or metals control, you will need additional platinum makeup.

MOTT: We have some industry surveys that agree with what was just mentioned. Approximately 70%, according to our survey, of all of the FCC units in North America are using combustion promoter. Virtually all of that combustion promoter usage is the separately added variety. When making a selection of combustion promoters, they really should be selected on a cost per performance basis, and not on a cost per noble metal or cost per platinum basis. This is because there are very big differences in the effectiveness of the support technologies and the impregnation technologies that are used in manufacturing promoters.

Some of the things to look for in an effective promoter are: good attrition resistance, so that it is not rapidly lost from the unit; good particle size distribution with a low fines content; and good activity retention. The relative stability of the various promoters is more difficult for the refiner to gauge for himself.

While we have laboratory tests that indicate that there are large differences in activity and stability between various brands of promoters, I would suggest to the refiners that they use a "trial by fire" kind of approach to selecting promoters. Here you set up a particular set of operating guidelines for your operator such as adding 10 lb of promoter only when the spread between dense and dilute phase thermocouples reaches 30 F. or whatever kind of specifications you want to set up.

I would give them specific operating specifications, and then keep track of promoter usage. Let the cost effectiveness of each promoter speak for itself. As a rule, without specific operating guidelines, operators tend to use much more promoter than is actually required (under the "if a little is good, then more must be better" theory of operation). Without these guidelines, money that the refiner should save by selecting a more cost effective promoter, ends up going up the stack due to excess usage.

Another part of the question seemed to be specific to a combustor style unit, so the following answer is taken from the perspective that this refiner may have a UOP high efficiency combustor style regenerator. Generally it is not necessary for the fully factory-promoted catalyst to be used in the UOP high efficiency regenerator design since this type of unit is usually capable of sustaining complete CO burn even without CO promoter.

There are two criteria that the refiner should meet in order to keep the UOP high efficiency regenerator in complete combustion. First the temperature at the bottom of the combustor vessel should be monitored. Afterburning can usually be eliminated by keeping this bottom combustor temperature at or above 1,250 to 1,275 F. It the unit is experiencing afterburning, then the quantity of recirculated catalyst that is delivered back to the combustor should be increased in order to raise the bottom combustor temperature some more. Second, the operators must make sure that there is excess oxygen in the flue gas available to complete the combustion process. Most UOP high efficiency regenerators operate with 2-3 mole % excess oxygen present in the flue gas.

While the high efficiency design does not require CO promoter, in actual practice, most refiners who have this design do use a separate CO promoter or at least keep some on hand for an emergency, since the ability to add promoter as required adds a great deal of operating flexibility if the unit runs into some problems.

DAVID B. BARTHOLIC (Intercat Inc.): As a general rule, it is normally not economical to use a factory-added additive, with your desired catalyst when compared to adding the additive separately and continuously at the unit in response to the variable of choice. For the catalyst supplier to add the oxidation promoter during manufacture requires that he over promote so that he is sure that the catalyst, which is added for activity, will give the desired oxidation level in all units under all operating conditions. The last thing the catalyst supplier wants is a problem with the additive, as his main business is selling catalyst. When selecting an oxidation promoter, besides platinum content, one needs to be concerned with the surface properties of the promoter. Low surface area materials with platinum on the surface tend to deactivate rapidly because the platinum is selectivity attrited off the surface. This can be readily determined by the platinum in the precipitator versus the platinum in the unit inventory. Likewise, one needs to be concerned with the particle size of the promoter and its attrition resistance. If NO, is a concern, the lower the platinum concentration the lower the NOx.

The most economical method of addition of any additive is a continuous system which will add on process control. These types of systems are available for additive addition from some of the suppliers.

KEN LOUDER (Unocal Corp): CO promoter will not work well in our Los Angeles FCCU. We theorize that the poor air and catalyst distribution in this regenerator results in a CO breakthrough on the side of the regenerator where the spent catalyst is introduced; whereas, there will be oxygen breakthrough on the opposite side of the regenerator. The oxygen and CO come together in the cyclones, resulting in a very severe afterburn. Promoter does not work, and we do not expect it to work because we do not have oxygen and CO together in the bed or in the lower part of the dilute phase.

JOHN GAUGHAN (CCSI/Metaloy): To confirm the comments of the last speaker, it has been our experience from the consulting we have done on units with just such problems, that each unit seems to have a unique threshold limit, above which CO promotion will take place very efficiently. However, this is so highly dependent upon the efficiency of fluidization within each regenerator configuration that each unit has to figure out where its own efficiency and threshold limits are. Once this baseline is established, it can become a useful tool to follow changes in regenerator flow patterns. A change in CO promoter level necessary to maintain a desired "Delta T" between the dense and dilute phases is often a first signal that channeling or other air/catalyst distribution problems are developing.

JAMES WILLIAMS (Akzo Chemicals Inc.): We have noticed that most combustion promoters are very soft, and up to 80% of that combustion promoter ends up in precipitator fines bins where it does not do a good job of promoting CO combustion.

As the zeolite content of an FCC catalyst is increased, what happens to the catalyst's attrition resistance and selectivity?

MOTT: As was mentioned earlier, some of the modern advances in catalyst binding technologies now allow some of us catalyst manufacturers to incorporate very high levels of zeolite in the catalyst without significant deterioration of catalyst's physical properties. Many of the selectivity characteristics that are exhibited by an equilibrium FCC catalyst are the result of the balance that is achieved between the zeolite and the matrix activities of the catalyst.

A useful way for the refiner to get a rough approximation of this balance is to measure the zeolite-to-matrix surface area ratio of the equilibrium catalyst. Catalysts with low zeolite-to-matrix surface area ratios tend to be poor in coke and gas selectivity in general. As the Z-to-M ratio of the equilibrium catalyst increases, say from 0.5 to 3 or above, the selectivity of the catalyst improves dramatically. Generally, catalysts that equilibrate with Z-to-M ratio below about 1 commercially, tend to be out of balance with too much bottoms cracking activity for most situations. Typically, these units that are equilibrating at very low Z-to-M ratio also exhibit very poor coke and very poor gas selectivity.

Catalysts with higher Z-to-M ratios produce much better selectivity with respect to the lower dry gas yields, lower coke making tendencies, and increased gasoline selectivity. To put this into perspective, the last industry wide survey that we conducted shows that the Z-to-M ratio of the refiners' equilibrium catalysts range from a low of 0.37 to over 5. If you want to know the actual Z-to-M ratio of your equilibrium catalyst, talk to your catalyst manufacturer, and we can measure it for you. There is a very thorough discussion of this subject in a paper that was presented in the 1988 NPRA Annual Meeting, "FCC Catalyst Design and Selection for Optimum Performance."

DONALD A. KEYWORTH (Akzo Chemicals Inc.): Catalyst manufacturers have now learned to produce commercial catalysts with very high levels of zeolite while still retaining good attrition indexes. As you increase any active ingredient to get more conversion, you run into a region where you start to get nonlinear response for fuel gas and coke make. However, when you increase the zeolite, it tends to retain octane performance and better selectivities for coke and gas make, and for LPG olefinicity over that of other ingredients that you might add.

Has anyone noticed a magnesium buildup phenomenon on equilibrium catalyst analyses with various FCC catalysts? Has anyone had deposit buildup in cyclone diplegs or on regenerator internals? Has analysis of these deposits shown the presence of large amounts of magnesium? What was the composition?

PARKINSON: At our refinery in Oakville, Ont., one of our FCC units experienced a problem when magnesium-based deposits built up in the regenerator secondary cyclone diplegs, Excessive catalyst loss resulted in a unit shutdown. By that time, we had already switched to a magnesium free catalyst. Our analysis showed the following: at the time the problem was detected, the magnesium level on the unit equilibrium catalyst was 50% of the fresh catalyst. When we analyzed the dipleg deposits, we found 30% magnesium oxide.

MILLER: We have not observed any buildup related to FCC catalyst or additives. We did, however, experience magnesium salt fouling some years back in the regenerator. At the time, water sprays were used to quench the dilute phase. A problem with boiler carryover had developed which resulted in significant solids carryover into the steam-water sprays. This caused large magnesium and calcium salt buildup on the cyclones. The salts most likely also built up on the catalyst.

FISCHER: We normally do not analyze our equilibrium catalysts for magnesium. We did, however, analyze 6 months of precipitator fines for magnesium when an expander deposit indicated that we had some magnesium. The deposit was obtained in September of 1988, and the fine samples were from November 1988, through April 1989. We found that there was no direct correlation; i.e., the magnesium content of the fines was much less than in the deposit. The actual expander deposit was mostly catalyst with trace metals about equal to the precipitator fines. These analyses are given below:

Expander Typical

deposit fines

analysis, analysis,

wt% wt%

Fe 0.54 0.5

Ni 0.088 0.022

V 0.067 0.06

Mg 0.31 N/A-0.08 for next 6 mo.

Ti 1.38 N/A-1.35 for E cat

Na 0.48 0.48

P 0.25 N/A-0.26 for next 6 mo.

COMEAU: Our refinery cracks 100% atmospheric tower bottoms. Some crudes with large concentrations of alkali metals and irons have caused dipleg plugging problems. Analysis of the deposit showed barium oxide, 14.3%, zinc, 0.7%; silica oxide, 40.7%; alumina oxide, 12.1%; calcium oxide, 4.9%; potassium oxide, 0.3%; iron, 24.6%; and other components, 2.4%. Our theory is that at high temperatures and high iron and alkali metals loading, the catalyst can form sticky, low melting point eutectics, causing deposit formation, and eventually plugging the diplegs and catalyst loss problems.

BIGGS: Several years ago we did away with our pipe distributor in the reactor stripper and went to a stream chest type, lower in the stripper. After some months we noticed higher than usual pressure drop in the flue gas line. We injected boiler feed water just below the flue gas valve. We blew out all kinds of ceramic material that was high in magnesium. I do not recall the exact analysis of the material. At the next turnaround we changed back to the pipe distributor in the stripper and we had no more problems in that area.

ROB GILMAN (Akzo Chemicals Inc.): We assisted a customer in analyzing deposits from his secondary regenerator cyclone dipleg. These cyclone deposits primarily consisted of magnesium oxide, about 23%. This deposit also contained 5% silica and negligible amounts of bulk alumina. This suggests that the deposits originated from a source other than the catalyst.

WARREN S. LETZSCH (Katalistiks-UOP): The form of the magnesium or alkali metal is important. Years ago, W. R. Grace used to sell silica magnesium catalysts that contained 30% magnesium. It was used in many units throughout the United States in the early 1960s and there was never any problem with buildup of magnesium on the cyclone diplegs or anywhere else. As some of you know, Katalistiks makes an SO), additive. It is a very stable compound, and does not decompose at any known regenerator conditions,

We have recently looked at some regenerator deposits, and we have found a magnesium compound, cordierite, which does seem to decompose and form deposits in various parts of the regenerator. Desox was never used in this unit.

Just recently another refiner sent us a sample of a deposit that was 30% calcium. That either came from the refractory, or it was in the feed. Free alkali metals can form deposits. I would agree with the comments of Mr. Comeau that eutectics are formed, the buildup occurs, and it eventually leads to malfunction of the regenerator.

JAMES A. TAYLOR (Caltex Services Corp.): We have only just opened a unit after a 3-1/2 year run. From the very beginning of the run, we experienced catalyst losses of 1 to 1-1/2 tons per day more than usual. These losses were a function of feed rate. This led us to believe that we had cyclone damage at startup. However, having gone in there in the last week or so, we have discovered the cyclones are okay, but five of the secondary diplegs are plugged. They are plugged with a mixture of catalyst and a binder which is very high in calcium. They have not finished the analysis yet. We suspect on first guess that we leached off some refractory repair at initial startup.

R.L. FLANDERS (Consultant): I would just like to remind Mr. Letzsch that the early Davison catalysts containing magnesia also contained 3 to 4% fluorine. This left as fluoro-silicic acid and digested all of the cyclones.

GEOF R. WILSON (Wilson & Associates): It is probably not that well known, but well into the middle of the 1970s, Filtrol produced a number of cracking catalysts that contained magnesium as part of the zeolite formulation. As far as I recall, there were no problems of this nature attributable to the catalyst. These catalysts did not contain fluorine.

HYDROCRACKING

How does the concentration of sulfates on regenerated catalyst affect catalyst activity? Is there a relationship between iron content of the catalyst and sulfate formulation?

SHIFLETT: The presence of sulfates on regenerated catalysts is usually a function of the type of catalyst as well as the service it has been in prior to regeneration. High nickel tungsten contents and also high vanadium contents typically will yield more sulfates.

It has been stated in previous NPRA O&A Session transcripts, that sulfur levels above 1% tend to be accompanied by a lower surface area, and my own observations would tend to support that concept.

I also note that there is a school of thought that is very well convinced that the sulfates will be converted to sulfides when you put the regenerated catalyst into application. Relationships between iron content and sulfate formation seems to be somewhat tenuous at best.

MILLER: Our regenerated hydrotreating catalysts normally contain about 0.5 wt % sulfate. At this level we do not see or anticipate any significant impact on catalyst activities.

We expect that other qualities, such as surface area and metals contamination, are more significant to variation in activity than sulfate at this level. The catalysts from our FCC feed hydrotreater contains very large amounts of iron sulfide upon dumping from the reactor. Most of the iron is deposited on the surface of the catalyst. We have not seen any relationship between iron content and catalyst sulfate content. Iron content on regenerated catalyst is normally under 1%.

Our experience is that sulfate content of the catalyst is more a function of the regeneration operation. It is well known that high oxygen and high temperatures during the initial burn will lead to significant conversion of metal sulfides to sulfates. To help insure this problem does not develop, we often send onsite inspectors to witness the regeneration of our catalyst during ex situ regeneration operations. One of the regenerators indicates that some catalysts with high vanadium, for some reason, end up with higher levels of sulfate, normally in the 2% range.

Cracked and pretreated stocks often contain very refractory nitrogen compounds. What effect on operation and catalyst performance do these feeds have? Is the ammonia effect on second stage catalyst totally reversible?

COMEAU: Much higher first stage temperatures, 20 to 35 F. on the inlets, are required to remove the nitrogen contained in the aromatic feedstock, such as LCO. Lower catalyst life due to higher deactivation from coking at higher temperatures is a result.

Assuming all nitrogen is converted to ammonia, ammonia effect on a second stage non-noble metal catalyst is reversible, to the extent that once the ammonia is decreased, catalyst activity improves.

The higher temperatures required to maintain constant conversion during high levels will result in irreversible coke formation. As for hydrocracker reactor operation, higher feed ammonia will increase reactor temperature, decrease isomerization reactions, increase C5 + yields, and decrease LPG yields.

SHIFLETT: It is well-documented that feeds containing refractory nitrogen, whether they are thermally or catalytically cracked, or pretreated at high severity, are significantly more difficult to process and generally result in lower activities and higher decline rates on the catalyst. While it is not unusual to find that the nitrogen and coke in FCC stocks are more difficult, some people, I think, are a bit surprised to discover that nitrogen compounds from residue hydrocrackers, whether they are fixed bed geometry or they are moving bed, also are real tough nuts to crack. If you think about it for a minute, that is not overly difficult to rationalize if you realize that the easy nitrogen is already gone, and what is left over has been pressure cooked in aromatic chicken wire for some time. Our evidence and experience indicates that the ammonia poisoning of the second stage is totally reversible, in that once it is removed, the problem will go away.

RICK BERTRAM (Unocal Corp.): Both treating and cracking catalysts will be affected by these types of feedstocks, as has already been mentioned.

In the treating reactor, the temperatures will have to be increased to meet effluent nitrogen target levels. This can increase production of polycyclic aromatics (PCA).

The unconverted nitrogen compounds that remain are extremely refractory and are difficult to convert. They are mostly converted at the top bed of the cracking reactor, but there is a poisoning effect. The top bed of cracking catalyst will typically lose activity relative to the lower beds. The net loss of cracking activity will require increased reactor temperatures to maintain conversion and prevent PCA buildup. If the ammonia concentration in the recycle gas is reduced or eliminated, the second stage activity will improve, but not always to pre-ammonia levels. If the level of refractory nitrogen compounds in the pretreated effluent is reduced, the activity of the cracking catalyst will recover. However, the improvement will most likely be considerably less than 100% expected activity.

What is the experience with ex situ presulfided non-noble metal hydrocracking catalyst in commercial service?

SHIFLETT: I do not know of any commercial experience in which ex situ presulfided, non-noble metal hydrocracking catalyst has been employed, but I am aware of some promising pilot plant data on such catalysts. This does include nickel tungsten formulations.

STEPHEN R. MURFF (Eurecat U.S. Inc.): Eurecat U.S. has recently applied Sulficat ex situ presulfiding process combined with the new 4A-CAT process on a non-noble metal hydrocracking catalyst for a U.S. refiner. The hydrocracking catalyst has been regenerated by Eurecat prior to this process being applied. This unit startup was said to be the smoothest in the experience of the 20 + year old unit. The unit temperatures and performance, both conversion and yield pattern, are very close to the performance seen in this unit with a fresh catalyst.

JAMES D. SEAMANS (Catalyst Recovery Inc.): CRI has had very positive results from laboratory and pilot plant studies regarding ex situ presulfiding of non-noble metal hydrocracking catalyst. We expect to be offering this service for commercial applications in the very near future.

What are the most important operating variables (i.e., LHSV, hydrogen partial pressure, reactor temperature, feedstock qualities) which affect catalyst deactivation? Is 90 + % sulfur removal achievable on 650-1,000 F. vacuum gas oil (VGO) at 450-500 psi hydrogen partial pressure? If so, what is the expected catalyst deactivation rate?

MILLER: The most important operating variable affecting catalyst deactivation is probably reactor temperature. Temperature is basically a dependent variable and a function of all the other variables. Feed quality is probably the next most important. A very significant factor beyond bulk feed properties is contamination. This can especially be a problem with purchased or poorly handled stocks. Even air contamination can significantly foul and deactivate catalyst.

The second part to this question is a tough one. The answer may be yes, provided you have enough reactor activity and temperature, but length might be very short. We tried running VGO through our 600 psi distillate treater a few years back. At the time I think we were looking for about one half this desulfurization level. The unit was unable to run due to a severe emulsion problem in the product separator. The wash water turned the product into mayonnaise.

COMEAU: All of the operating variables mentioned are important and will effect run length and deactivation rate. Another important parameter, although not an operating variable, is extremely good liquid distribution for the two-phase reactor. Dense loading should help the liquid distribution. One plant runs a VGO unit which operates as follows: The feedstock is a 27.5 API material; 5% point, 600 F.; 90% point, 1,145 F.; UOP K factor of 12:23; liquid hourly space velocity of the reactor 2.5; hydrogen partial pressure 660 psi; weight average bed temperature 680 F.; 1 wt % sulfur in the feed; desulfurization is 85 to 90%. Cycle lengths are 2 years at a deactivation rate of 0.35 F./bbl/lb. If pressures were lowered to reach the 500 psi hydrogen partial pressures as mentioned in the question, the deactivation rate is estimated to increase by 25%.

PARKINSON: In our experience, the three factors which have the most dramatic negative impact on catalyst life are: high severity, low partial pressure, or poor quality feedstock.

SHIFLETT: I would only say the two most important factors are hydrogen partial pressure and hydrogen partial pressure. In a feed that does not contain significant metals, the role of hydrogen pressure is paramount in slowing down coke deactivation on catalyst. Of course, feedstock heaviness comes to play as well, but those who like to have flexibility, I think, would prefer to have the coke deactivation slowed down no matter what.

In terms of carbon residue and endpoint, you can choose feeds that will really cut into your activity. These coupled with high severity operations, either in terms of high space velocity or high temperature, also are going to work together to limit your cycle life. Raising hydrogen partial pressure by either increasing recycle gas purity or even makeup gas purity, if your makeup is not overly pure, can have some profound effects on what I would call marginal units, or those units running on the ragged edge for whatever processing goal they are trying to achieve.

For the specific case here, there was not a space velocity given and cycle life estimation becomes a little bit difficult. This is one of those cases that I would lump in with being on the ragged edge with respect to low hydrogen partial pressure. None the less, at the risk of sticking out my neck, I would go ahead and make a rough estimate which, in part, I think would go along with what Mr. Comeau has already said.

If you have a 100% straight run material with an endpoint of no higher than 1,000 F. and let's say sulfur, no more than about 2% or so, then it's possible to get a run length of about 1 year provided your reactor temperature, at end of run, could get up into the mid 700 F. range. The inclusion of any coker or cracked stocks is probably going to cut into that significantly and you wouldn't be able to make it.

Of course, being a vendor, we would recommend a high activity cobalt molybdenum catalyst for this situation, assuming nitrogen is not a factor and you are only looking for sulfur removal.

TOM PUSTY (ICI Katalco): We agree with the comments already made concerning the variables affecting catalyst deactivation rate. Adding to the second part of this question we had the following feed:

Heavy gas oil

23.4 API

590 F. (IBP) 896 F. (50%) 1,112 F. (EP)

2.4% sulfur

670 F. WABT

2,300 scf/bbl hydrogen to oil

325 scf/bbl hydrogen consumption

1.04 hr-1

740 psig

85-90% desulfurization

With this particular feed, a 0.32 F. bbl/lb deactivation rate was observed. Depending on changes in operating conditions and feed quality we have seen deactivation rates ranging from 0.2-2.0 F./bbl/lb.

What methods are used to reduce or eliminate catalyst agglomeration in hydrotreating/conversion units? This is particularly prevalent for resid hydrodemetallization catalyst. What mechanical techniques are used for catalyst removal?

FISCHER: We do not take any unusual precautions in our gas oil Unibon unit, and we do not have any particular problems unloading catalyst. We do use a liquid nitrogen cooldown procedure to help speed up the cooldown process. Our catalyst is normally dry and relatively free flowing, but sometimes we do encourage it with hydroblasting lances. Coke on spent catalyst normally runs less than 10%.

SHIFLETT: Naturally, you will want to avoid any agglomeration that could be caused by inorganic materials, in particular, salt deposition and rust scale deposition. This calls for effective and consistent desalter operation as has been discussed earlier. In addition, after shutdowns and changeouts, one needs to take care to purge the upstream components of the system to make sure that any trash, scale, and particulate that was knocked loose in the various maintenance activities is not going to end up on top of the catalyst bed. In terms of operation, agglomeration by coke deposition is minimized, again at the risk of harping on it, by the highest hydrogen partial pressure practical.

Feedstock selection can also factor into this in the case of resid hydrotreaters. Asphaltene flocculation can occur if a feed blend is created in which the solvating ability of the oil is inadequate to keep the asphaltenes in solution. This can be tested for by traditional fuel oil stability tests, such as the hot filtration test, the spot test, or the Martin-Bailey test.

Catalyst system tailoring is also helpful, both in terms of size grading, particularly in guard beds or guard reactors. The use of high void volume, void fraction geometries on bed tops, coupled with extensive size grading, can help, and certainly increases run lengths that are shortened by pressure drop in guard reactors or guard beds.

As a rough guideline, any two adjacent nominal catalyst sizes should not vary more than a factor of two. There will probably be some disagreement on that. That is a relatively conservative recommendation. As for unloading, the use of hydroblasting water lances is useful. I have also seen a preflush and soak with water with appropriate concerns for corrosion being taken care of by inclusion of soda ash. The most dramatic mechanical techniques I know about involve the use of explosives.

COMEAU: We have a hydrotreater that cokes up and "cements" catalyst together. Use of nitrogen powered jack hammers and inert entry is the mechanical means we have used.

DELBERT F. TOLEN (Akzo Chemicals Inc.): Has anyone tried this? I talked a customer into it, that is why I want to know. I do not know what the result was. Has anyone tried loading a 4 to 6 in. layer of large balls near the bottom of the bed to help break the bed loose when it is agglomerated like that? This customer would have to get in and almost mill it out. It would take him three or four weeks, and so I suggested that they put a 4 to 6 in. bed of large balls a couple of feet from the bottom of the reactor and then up about 5 ft, and see if it would not break free readily. I do not know whether it worked or not. He was overseas, and I have no idea.

SHIFLETT: I have seen some reactors where they do use large ceramic balls near the dump nozzle and in that region. Fortunately, the ones that I have seen that do it dumped rather well. Now, whether that is the cause or just serendipity, I do not know. It certainly sounds reasonable.

FISCHER: We use large inert balls on the bottom support plate in our Unibon reactors, and we do not have any particular plugging problems. We do not have any intermediate layers of balls higher up in the middle of the bed.

Where do refineries dispose of hydrocracking catalyst that has metals content too low for most North American reclaimers to reclaim economically?

SHIFLETT: North American reclaimers can process spent hydrocracking catalysts with low metal contents. The metals recovery processes in general are somewhat complex such that in many places the value of the low level of metals recovered does not really cover the processing cost. To cite some rough values, again that I have obtained through that legendary grape vine, the low metals levels increase the processing fee somewhere in the vicinity of maybe 20% to 40 cents per lb. This is however only a small percentage of original cost basis, and if one puts it on a percentage of original cost basis, it is less than the processing fees for most hydrotreating catalysts.

There is an active commercialization effort for processing tungsten-containing catalyst. One of the vendors that handles recycling of metals, CRI-MET, is due to announce relatively shortly that they will have that capability.

PARKINSON: At our Edmonton refinery, our hydrocracking catalyst is nickel-based which is currently being sold for metal recovery. Past disposal has entailed shipment to the Alberta Special Waste Treatment Center for landfill with double synthetic liners, leachate collection, and ground water monitoring. The catalyst is classified as hazardous waste because of the potential to release nickel carbonyl vapors and to react exothermically with air.

Copyright 1990 Oil & Gas Journal. All Rights Reserved.