US3224961A

US3224961A - Cracking with poison-resistant catalyst

Info

Publication number: US3224961A
Application number: US226182A
Authority: US
Inventors: Erickson Henry; Howard G Russell; Robert A Sanford
Original assignee: Sinclair Research Inc
Current assignee: Sinclair Research Inc
Priority date: 1962-09-25
Filing date: 1962-09-25
Publication date: 1965-12-21
Anticipated expiration: 1982-12-21

Description

United States Patent 3,224,961 CRACKING WITH POISON-RESISTANT CATALYST Henry Erickson, Park Forest, 111., Howard G. Russell, Munster, Ind, and Robert A. Sanford, Homewood, llll., assignors to Sinclair Research, Inc. Wilmington, DeL, a corporation of Delaware No Drawing. Filed Sept. 25, 1962, Ser. No. 226,182 5 Claims. (Cl. 208-120) This invention is a method for catalytic cracking of mineral oil hydrocarbon feedstocks containing vanadium and nickel. The cracking employs a catalyst having a considerable immunity to the poisoning effects of nickel and the catalyst may be demetallized to remove vanadium. The feedstock generally contains at least a portion of highly nickel-contaminated residual petroleum hydrocarhon and the cracking method includes processing steps for controlling the level of metal poisons by discarding some portions and demetallizing other portions of the nickel and vanadium contaminated catalyst. These procedures help to improve the catalytic cracking selectivity of the catalyst by keeping catalyst metal contaminants at or below a tolerable level.

One of the most important phases of study in the improvement of catalyst performance in hydrocarbon conversion is in the area of metals poisoning. Although referred to as metals, these catalyst contaminants may be in the form of free metals or relatively non-volatile metal compounds. It is to be understood that the term metal used herein refers to either form. Various petroleum stocks have been known to contain at least traces of many metals, and in addition to those metals naturally present, including some iron, petroleum stocks have a tendency to pick up tramp iron from transportation, storage and processing equipment. In conventional processing, these metals in the stock deposit in a relatively nonvolatile form on the catalyst during the conversion processes and regeneration of the catalyst to remove coke does not remove these contaminants, which have an adverse effect on cracking results when using conventional cracking catalysts. Nickel, vanadium, iron and copper, for example, markedly alter the selectivity and activity of conventional cracking catalysts if allowed to accumulate, producing a higher yield of coke and hydrogen at the expense of desired products, such as gasoline and butanes. For instance, it has been shown that the yield of butanes, butylenes and gasoline, based on converting 60 volume percent of cracking feed to lighter materials and coke dropped from 58.5 to 49.6 volume percent when the amount of nickel on a conventional catalyst increased from 55 ppm. to 645 ppm. and the amount of vanadium increased from 145 ppm. and 1480 ppm. in a more or less conventional catalytic cracking of a normally liquid feedstock containing some metal-contaminated stocks. Since many cracking units are limited by coke burning or gas handling facilities, increased coke or gas yields require a reduction in conversion or throughput to stay within the unit capacity.

Refiners cope with the problem of metal poisoning by adopting several techniques. One technique includes selecting only feedstocks of low metal content or treating the feedstock to minimize. its metal content. Another technique requires removing from the hydrocarbon conversion system as much metal as is fed to it per unit time, in order to obtain and retain a total amount of metal in the system below a level where the conversion process is made economically unfeasible by the poisoning effect of the metal. In most conversion processes some metal-containing catalyst is continuously lost from the system in the form of fines which leave the system with effluent gases.

"ice

The replacement of this loss with fresh unpoisoned catalyst reduces the net amount of metal in the system.

In this invention the contaminating effects of nickel in a feedstock may be substantially avoided by employing a cracking catalyst more resistant to metals, especially nickel, than conventional cracking catalysts so that metals accumulation on the catalyst has less poisoning effect on the system. In operating with this catalyst the petroleum refiner can employ more highly contaminated feedstocks and may allow more metals to accumulate on the catalyst than on a conventional catalyst Without severe, economically disadvantageous effects on the product distribution. The present invention provides for procedures where, in the use of moderately high nickelcontent feedstocks, the nickel contamination of the catalyst may be controlled by catalyst discard and vanadium contamination of the catalyst may be controlled by subjecting the catalyst to demetallization procedures. Such procedures include contacting the contaminated catalyst with a molecular oxygen-containing gas after conventional regeneration procedures to remove carbonaceous deposits, and subsequently treating the catalyst with a basic aqueous Wash containing ammonium ions.

Cracking of heavier hydrocarbon feedstocks to produce hydrocarbons of preferred octane rating boiling in the gasoline range is widely practiced and conventionally uses a solid oxide catalyst to give end-products of fairly uniform composition. The catalysts which have received the widest acceptance today are usually activated or calcined predominantly silica or silica-based, e.g., silicamagnesia, silica-zirconia, silica-alumina, etc., compositions in a state of slight hydration and containing small amounts of acidic oxide promoters in many instances. The oxide or more-or-less homogeneous oxide mixture compositions sometimes may also contain small amounts of other inorganic materials, but current practice in catalytic cracking leans more toward the exclusion from the silica materials of foreign constituents such as alkaline metal salts which may cause sintering of the catalyst surface on regeneration and a drop in catalytic activity. For this reason, the use of wholly or partially synthetic gel catalysts, which are more uniform and less damaged by high temperatures in treatment and regeneration, is often preferable. Such homogeneous catalysts, however, are sensitive to the poisoning effects of nickel in the feedstock, while the catalyst used in this invention is almost insensitive to the effect of nickel at moderate levels.

The catalyst comprises a particulate material having a coating or deposit of synthetic alumina hydrate on a sulr strate. The substrate is a solid inorganic oxide mixture, generally a clay or a synthetic silica-alumina gel or a mixture of the two. The substrate usually contains at least about 40% silica and often is predominantly silica. Preferably the substrate contains alumina, generally in an amount of at least about 5 or 10 to 30% and the combined silica and alumina content is at least about or of the substrate, the remainder, if any, generally comprising other inorganic oxides, such as those found originally in the clay, or added for additional promoting effects. The substrate may be derived from one of the clays conventionally used in catalytic cracking, such as halloysite or dehydrated halloysite (kaolinite), montmorillonite or bentonite. In most cases it is desirable to treat the clay with mineral acid for purposes of activation or at least for iron removal. The substrate may also be a completely synthetic gel oxide material, which may be silica-based and ordinarily contains a substantial amount of a gel or gelatinous precipitate comprising a major portion of silica and at least one other material, such as alumina, magnesia, zirconia, etc.

The substrate is in the form of recognizable particles.

While the size range of the particles is not of the utmost significance, the particles are greater than colloidal in size, that is, they are larger than the micelles, the submicronic particles which make up a colloid. The substrate particles may be characterized by their lack of electric charge and the fact that they do not disperse to form a colloid when placed in an aqueous medium, and, even if severely agitated, do not form a true, stable, colloidal suspension but rather settle, on standing, to leave a supernatant liquid. Also, particles suitable for forming the substrate do not grow by accretion or inorganic polymerization with each other. The silica-based gel substrate is generally prepared for alumina deposition by being washed, dried, if desired, and sized.

Hydrated alumina gel can be mixed with the substrate particles to form the catalyst. Alternatively the alumina gel can be prepared in the presence of the substrate particles. Preferably the hydrated gel is formed by reacting ammonia with an aqueous solution of an aluminum salt after which the alumina hydrate or alumina hydratesubstrate slurry is washed and the hydrate concentrated as by settling and the aqueous material filtered off. The aluminum salt is generally a sulfate, such as Al (SO or NH Al(SO The solution may contain a concentration of about 5 to 20% aluminum salt and the ammonia will generally be added as ammonia water until the desired amount of alumina hydrate is precipitated. During the formation of the alumina hydrate the pH is generally controlled to produce certain characteristics in the alumina hydrate. For example, a catalyst formed from a substrate plus alumina precipitated at a pH greater than has good resistance to nickel, preferably, however, the catalyst is prepared by precipitation of hydrous alumina in the presence of the substrate, at a pH of about 5 to 9. Preferably this hydrated alumina gel is formed at a pH of about 7 to 7.5. Precipitation of alumina from an aqueous solution of an alkali aluminate by addition of an acid may also be employed. Also, the hydrous alumina may be precipitated by hydrolysis from alcohol solutions of aluminum alkoxides although the use of inorganic salts is preferred. The alumina produced in the presence of the substrate and at the pH conditions described, is, upon calcination, mostly amorphous, as distinguished from alumina precipitated at higher pI-ls, which is crystalline in its form. Generally, about 3 to 100 parts of the hydrous synthetic alumina (dry basis), preferably 10 to 25 parts of alumina, are coated on 100 parts of substrate (dry basis). Thus the finished catalyst contains about 3 to 50% of synthetic alumina on the substrate, preferably about 9 to 20% and the total alumina content of the catalyst is between about 20 and 65 percent, preferably around 25 to 50% dry basis. For example, about 10 to 20 parts synthetic alumina hydrate gel may be added to or mixed with about 100 parts of an acid-treated clay containing about 20% alumina to give a catalyst having a total alumina (natural and synthetic) content of about 27 to 33%. After precipitation the alumina hydrate-substrate slurry is washed and the hydrate concentrated as by settling, and the aqueous material is filtered off, after which the catalyst precursor is thoroughly washed to remove sulfate or other interfering anions.

The substrate particles will generally be provided in a fluidizable particle size and thus the resulting coated material will be fluidizable. Alternatively the coated substrate may be formed to macro-shape by pelleting, extrusion, etc., dried, and generally the catalyst is calcined before use. The physical form of the catalyst varies with the type of manipulative process to which it will be exposed. In fiuid processing, gases are used to convey the catalyst between reaction and regeneration zones and to keep it in the form of a dense turbulent bed which has no definite upper interface between the dense (solid) phase and the suspended (gaseous) phase mixture of catalyst and gas. This type of processing requires the catalyst to be in the form of a fine powder, generally in a size range of about 20 to 150 microns or less.

Cracking is ordinarily effected to produce gasoline as the most valuable product and is generally conducted at temperatures of about 750 to 1100 F., preferably about 850 to 950 F., at pressures up to about 200 p.s.ig., preferably about atmospheric to p.s.ig., and without substantial addition of free hydrogen to the system. In cracking, the feedstockis usually a mineral oil or petroleum hydrocarbon fraction such as straight run or recycle gas oil or other normally liquid hydrocarbon boiling above the gasoline range. For typical operations, the catalytic cracking of the hydrocarbon feed would normally result in a conversion of about 50 to 60 percent of the feedstock into a product boiling in the gasoline boiling range.

Hydrocarbon petroleum oil used as a cracking feedstock often contains traces of poisoning metals; the advantageous features of this invention are more fully exploited when the feedstock contains poisoning metal; that is, nickel and vanadium, and perhaps other metals mentioned above. More than one-third part per million of nickel (0.3 p.p.m. measured as NiO) and/or one-half part or more per million vanadium (0.5 p.p.m. measured as V 0 may be in the feedstock and may result from blending a residual feedstock component containing perhaps as much as about 500 or even 1000 p.p.m. metal with a relatively unpoisoned stock of any desired type normally utilized in catalytic conversion operations. The process of this invention is of greatest value in converting feedstocks such as residual and heavy distillate stocks, that is, atmospheric tower bottoms and materials derived therefrom. Such stocks, when blended with relatively unpoisoned stocks, may contain more than about one p.p.m. nickel and more than about two p.p.m. vanadium. The total nickel in the feed may range up to about 5 or 15 p.p.m.

The catalyst is generally used as a fluidized bed, preferably containing at least 40% of particles smaller than 200 mesh, that is, smaller than 74 microns. Preferably at least about 25% of the catalyst particles are in the 4080 micron range. Catalytic conversion systems also include regeneration procedures in which the catalyst is periodically contacted with free-oXygen-containing gas in order to restore or maintain the activity of the catalyst by removing carbon. conventionally, fluid catalyst regenerators process about 560 tons of catalyst per minute, using about 2000 to 2800 standard cubic feet of air per ton of catalyst. The average residence time for a quantum of catalyst is often about 3-10 minutes. The regeneration rate is generally designed to keep the catalyst in the reactor at a carbon level up to about 1.2% and regenerated catalyst usually has a carbon content of about 0.2 to 0.5%.

As mentioned, this invention usually employs a feedstock having appreciable amounts of vanadium and more heavily contamianted with nickel than conventional hydrocarbon feeds. The invention also employs a nickel resistant catalyst. In accordance with this invention, the use of a semi-synthetic alumina-on-halloysite or kaolin clay catalyst containing about 10 to 65% alumina to crack a feedstock containing more than about one p.p.m. nickel and two p.p.m. vanadium is operable economically with a catalyst having an equilibrium nickel content level about two or three or even more, times as high as normally causes a severe penalty in cracking activity or selectivity. The ability to operate at this higher level of metal on the catalyst has a number of advantages. By allowing a greater metal content on the catalyst this invention provides for a greater metal removal per weight of catalyst when the catalyst fines are lost inadvertently or catalyst particles are deliberately discarded, or when portions of the catalyst are subjected to demetallization procedures. In conventional cracking systems the tolerance for poisoning metal oxide is seldom greater than about 200 p.p.m.

nickel and/or about 500 ppm. vanadium, measured as the oxides. According to this invention, the equilibrium nickel content level may be, and usually is higher than in ordinary cracking systems because of the particular composition of the catalyst, as set forth in this invention. The equilibrium nickel content may be allowed to reach as much as about 1000 ppm. NiO and may be maintained by the removal of catalyst from the system either purposely or inadvertently and the introduction of fresh catalyst. As mentioned, catalyst is usually unavoidably removed from the system in the form of fines which leave with effiuent gases. In addition to this particular catalyst loss, it has been found expedient to discard purposely enough poisoned catalyst per unit time so that replacement with unpoisoned or less poisoned catalyst will usually keep the nickel content level at the desired equilibrium.

The amount of catalyst discard bears a definite relationship to the amount of nickel contaminant present in the feedstock and the equilibrium nickel content level tolerable in the processing unit. Thus, in the practice of this invention the discard requirement may be embodied in the following equation which defines the amount of catalyst which may be withdrawn to maintain the equilibrium nickel content level tolerable in a particular unit:

In this equation Y equals the amount of discard (lbs. catalyst/bbl. of feedstock); X equals the amount of nickel in the feedstock (p.p.m.); Z equals the equilibrium nickel content level tolerable in the unit; D equals the feed density, lbs./bbl.; and S equals stack loss,

lbs. catalyst bbl. of feedstock Thus for a given desired equilibrium nickel content level and a given nickel content level and a given nickel content in the feedstock, variables which may be chosen or determined by the operator, the amount of discard or loss advantageous for the practice of this invention may be ascertained.

Catalyst discard based on the nickel content of the feed may not be sufiicient to keep the vanadium level of the catalyst within tolerable limits because of the generally higher vanadium/nickel ratio in feedstocks and susceptibility of the catalyst to vanadium poisoning. Accordingly, the equilibrium vanadium content level may be maintained, in addition to that achieved concurrently with the discard procedures, by catalyst demetallizatioin which may be accomplished by the intermittent or continuous withdrawal of contaminated catalyst from the cracking system, for example, from the regenerator standpipe. A suitable amount, generally a very small portion of the catalyst, may be removed from the cracking system preferably after the oxidation regeneration which serves to remove carbonaceous deposits. With a continuously circulating catalyst stream, such as in the ordinary fluid system the portion of catalyst for treating may conveniently be an intermittently or continuously removed slipstream of catalyst from the regenerator standpipe. In a continuous operation of the commercial type a satisfactory withdrawal rate from the regenerator may be about 5 to 50% of the total catalyst inventory in the system, per twenty-four hour day of operation although other rates may be used.

After regeneration, the slip-stream of the catalyst, having usually a carbon content of about 0.2 to 0.5%, is treated with molecular oxygen-containing gas to improve the removal of vanadium from the poisoned catalyst. This treatment is preferably performed at a temperature at least about 50 F. higher than the regeneration temperature, that is, the average temperature at which the major portion of carbon is removed from the catalyst.

The temperature of treatment with molecular oxygencontaining gas will generally be in the range of about 1000 to 1800 F. but below a temperature where the catalyst undergoes any substantial deletrious change in its physical or chemical characteristics, preferably a temperature of about 1150 to 1350 F. or even as high as 1600 F. The duration of the oxygen treatment and the amount of vanadium prepared by the treatment for subsequent removal is dependent upon the temperature and the characteristics of the equipment used. If any significant amount of carbon is present in the catalyst at the start of this high-temperature treatment, the essential oxygen contact is that continued after carbon removal, which may vary from the short time necessary to produce an observable effect in the latter treatment, say a quarter of an hour, to a time just long enough not to damage the catalyst. In any event, after carbon removal, the oxygen treatment of the essentially carbon-free catalyst is at least long enough to convert a substantial amount of vanadium to a higher valence state as evidenced by a significant increase, say at least about 10% in the vanadium removal in subsequent stages of the process. This increase is over and above that which would have been obtained by other metals removal steps without the oxygen treatment. The maximum practical time of treatment will vary from about 4 to 24 hours, depending on the type of equipment used. The oxygen-containing gas used in the treatment contains molecular oxygen as the essential active ingredient and there is little significant consumption of oxygen in this treatment. The gas may be oxygen, or a mixture of oxygen with inert gas, such as air or oxygen-enriched air, containing at least 1%, preferably at least about 10% O The partial pressure of oxygen in the treating gas may range widely, for example, from about 0.1 to 30 atmospheres, but usually the total gas pressure will not exceed about 25 atmospheres.

Vanadium may be removed from the catalyst after the high temperature treatment with molecular oxygencontaining gas by Washing it with a basic aqueous solu tion. The pH is frequently greater than about 7.5 and preferably the solution contains ammonium ions which may be NH ions or organic-substituted NH ions such as methyl ammonium and quaternary hydrocarbon radical ammoniums. This aqueous wash solution can be prepared by addition of a dry reagent or a concentrated solution of the reagent to water. Ammonia or methylamine gas may be dissolved directly in water.

The amount of ammonium ion in the solution is sufficient to give the desired vanadium removal and will often be in the range of about 1 to 25 or more pounds per ton of catalyst treated. Five to fifteen pounds is the preferred ammonium range but the use of more than about 10 pounds does not appear to increase vanadium removal unless it increases pH. The temperature of the wash solution does not appear to be significant in the amount of vanadium removed, but may vary within wide limits. The solution may be at room temperature or below, or may be higher. Temperatures above 215 F. require pressurized equipment, the cost of which does not appear to be justified. The temperature, of course, should not be so high and the contact should not be so long as to seriously harm the catalyst. The time of contact also may vary within wide limits, so long as thorough contact between the catalyst and the wash solution is assured. Very short contact times, for example, about a minute, are satisfactory, while the time of washing may last 2 to 5 hours or longer.

After the ammonium Wash the catalyst slurry can be filtered to give a cake which may be reslurried with water or rinsed in other ways, such as, for example, by a water wash on the filter, and the rinsing may be repeated, if desired, several times. A repetition of the ammonium wash without other treatments seems to have little effect on vanadium removal if the first washing has been properly conducted but a repetition of the basic aqueous ammonium wash after a repeated high temperature oxygen treatment usually does serve to diminish further the vanadium content of the catalyst.

In practicing this invention at the refinery, that portion of the poisoned catalyst, aside from that lost unavoidably, which is necessary for nickel control, is removed as discard. This amount of discard (pounds of catalyst per barrel of feedstock) as stated above, may conveniently be determined from the relationship existing between it and the feedstock nickel content and the nickel equilibrium content. Where necessary for additional vanadium control, a portion of regenerated catalyst is also removed and treated with an oxygen-containing gas at a temperature and for a length of time found to be sutficient to increase vanadium removal without unduly damaging the catalyst. Then the treated catalyst is washed with the aqueous ammonia solution and returned to the cracking unit, for example, to the regenerator. The amount of metal removed in practicing the combination of steps, i.e., the discard and demetallization procedures, will control catalyst metal contaminants at or below a tolerable level. In situations where the feedstock nickel content or total metal content becomes excessively high, a nickel demetallization treatment may prove advantageous. Accordingly certain demetallization procedures may be employed such as the sulfiding, chlorination, Washing, etc. techniques described below. The chlorination procedure is also helpful for vanadium removal in addition to or substituted for, the vanadium removal tech niques described above.

Sulfiding, for instance, may be accomplished by contacting the metal-contaminated catalyst with a sulfur-containing vapor such as elemental sulfur vapors or more conveniently a volatile sulfide such as H 8, CS or a mercaptan, at a pressure from atmospheric to about 1000 p.s.i.g. and at an elevated temperature generally in the range of about 750 to 1600 F., preferably about 1000 to 1200 F. The preferred upper pressure limit is about p.s.i.g. Other treating conditions may include a sulfurcontaining vapor partial pressure of about 0.1 to p.s.i.g. or more, preferably about 0.5 to 15 p.s.i.g. Hydrogen sulfide is the preferred sulfiding agent. The sulfiding gas may contain about 10 to 100 mole percent H 5, preferably at least about 80 mole percent H 5. Pressures below atmospheric can be obtained either by using a partial vacuum or by diluting the vapor with gas such as nitrogen or hydrogen. The time of contact may vary on the basis of the temperature and pressure chosen and other factors such as the amount of nickel to be removed. The sulfiding may run, for, say, up to about 24 hours or more depending on these conditions and the severity of the poisoning. Usually about l-6 hours is a sulficient time. Temperatures of about 900 to 1200 F. and pressures approximating 1 atmosphere or less seem near optimum for sulfiding and this treatment often continues for at least 1 or 2 hours but the time, of course, can depend upon the nature of the treating system, e.g., batch or continuous, as well as the rate of diffusion within the catalyst matrix.

Subsequent to sulfidation, the metal poison containing catalyst may conveniently be chlorinated at a moderately elevated temperature, up to about 700 or even 1000 F. wherein the catalyst composition and structure is not materially harmed by the treatment and a substantial amount of the poisoning metal content is converted to chlorides. The chlorination takes place at a temperature of at least about 300 F., preferably about 550 to 650 F. with optimum results usually being obtained near 600 F. The chlorinating agent is essentially anhydrous, that is, if changed to the liquid state no separate aqueous phase would be observed in the reagent.

The chlorinating reagent is a vapor which contains chlorine or sometimes HCl, preferably in combination with a promoter, preferably a carbon or sulfur compound, for example, a chlorine substituted light hydrocarbon, such as carbon tetrachloride, which may be used as such or formed in situ by the use of, for example, a vaporous mixture of chlorine gas with low molecular weight hydrocarbons such as methane, n-pentane, etc.

The stoichiometric amount of chlorine required to convert the nickel and vanadium to their most hi hly chlorinated compounds is the minimum amount of chlorine ordinarily used and may be derived from free chlorine, combined chlorine or the mixture of chlorine with a chlorine compound promoter described above. However, since the stoichiometric amount of chlorine frequently is in a neighborhood of only 0.001 g./g. of catalyst, a much larger amount of chlorine, say about 1-10 percent active chlorinating agent based on the weight of the catalyst is generally used. The amount of chlorinating agent required is increased if any significant amount of water is present on the catalyst so that substantially anhydrous conditions preferably are maintained as regards the catalyst as well as the chlorinating agent. The promoter is generally used in the amount of about l-S or 10 percent or more, preferably about 23 percent, based on the weight of the catalyst for good metals removal; however, even if less than this amount is used, a considerable improvement in metals conversion is obtained over that which is possible at the same temperature using chlorine alone. The amount of promoter may vary depending upon the manipulative aspects of the chlorination step, for example, a batch treatment may sometimes require more promoter than in a continuous treatment for the same degree of effectiveness and results. The chlorine and promoter may be supplied individually or as a mixture to a poisoned catalyst. Such a mixture may contain about 0.1 to 50 parts chlorine per part of promoter, prefably about ll0 parts per part of promoter. A chlorinating gas comprising about l30 weight percent chlorine, based on the catalyst, together with one percent or more S CI gives good results. Preferably, such a gas provides l10 percent C1 and about 1.5 percent S Cl based on the catalyst. A saturated mixture of CCL; and C1 or HCl can be made by bubbling chlorine or hydrogen chloride gas at room temperature through a vessel con taining CCl such a mixture generally contains about 1 part C Cl :510 parts C1 or HCl.

Conveniently, a pressure of about 0-100 or more p.s.i.g., preferably about 0l5 p.s.i.g.. may be maintained in chlorination. The chlorination may take about 5 to minutes, more usually about 20 to 60 minutes, but shorter or longer reaction periods may be possible or needed, for instance, depending on the linear velocity of the chlorinating and purging vapors.

After chlorination the catalyst may be washed in a liquid aqueous medium to remove, for instance, nickel chloride, preferably after the catalyst is cooled to avoid the use of excessive pressures to maintain the liquid phase. The catalyst structure may be quite sensitive to HCI formed in the treatment, so that several precautions should be observed in the aqueous liquid washing. A great excess of water can be used, for instance, sufficient to give a slurry containing only minor amounts of solids. Also, the catalyst should not be allowed to remain in this slurry for too long a time, ordinarily not more than about 5 minutes; a residence time of 2 to 3 minutes in the original wash is generally preferred.

The water used may be distilled or deionized prior to contact with the chlorinated catalyst. However, the aqueous medium can contain extraneous ingredients in trace amounts, so long as the medium is essentially water and the extraneous ingredients do not interfere with demetallization or adversely affect the properties of the catalyst. Temperatures above 212 F. and elevated pressures may be used but the results do not seem to justify the added equipment. The aqueous liquid is preferably acid and a weakly acidic condition may be obtained by the chlorides generally present in a chlorinated catalyst which has not been purged too severely.

After the wash the slurry can be filtered to give a filter cake which may be reslurried with more water or rinsed in other ways, such as, for example, by a water wash on the filter, and the rinsing may be repeated, if desired, several times. After this or after the final treatment which may be used in the catalyst demetallization procedure, the catalyst is conducted to its conversion system, for instance to the hydrocarbon conversion reactor or catalyst regenerator, although it may be desirable first to dry the catalyst filter cake or filter cake slurry at say about 250 to 450 F., and also, prior to reusing the catalyst in the conversion operation it can be calcined in air as described above.

Alternative to the removal of nickel contaminants by procedures involving contact of the catalyst with aqueous meida after chlorination nickel poison may be removed from the catalyst as the volatile nickel carbonyl by treatment with carbon monoxide, as described in copending application Serial No. 47,598, filed August 4, 1960, incorporated herein by reference. In such a procedure the catalyst is treated with hydrogen at an elevated temperature during which nickel contaminant is reduced to the elemental state, then treated, preferably under elevated pressure and at a lower temperature with carbon monoxide, during which nickel carbonyl is formed and flushed off the catalyst surface.

Hydrogenation takes place at a temperature of about 800 to 1600 F., at a pressure from atmospheric or less up to about 1000 p.s.i.g. with a vapor containing 10 to 100% hydrogen. Preferred conditions are a pressure up to about 15 p.-s.i.g. and a temperature of about 1l00 to 1300 F. and a hydrogen content greater than about 80 mole percent. The hydrogenation is continued until surface accumulations of poisoning metals, particularly nickel, are substantially reduced to the elemental state Carbonylation takes place at a temperature substantially lower than the hydrogenation, from about ambient temperature to 300 F. maximum and at a pressure up to about 2000 p.s.i.g., with a gas containing about 50 100 mole percent CO. Preferred conditions include greater than about 90 mole percent CO, a pressure of up to about 800 p.s.i.g. and a temperature of about ll80 F. The CO treatment generally serves both to convert the elemental metals, especially nickel and iron, to volatile carbonyls and to remove the carbonyls.

The following examples of the method of this invention are to be considered illustrative only and not limiting.

Example I A crude oil containing nickel and vanadium contaminants is fractionated to produce a 650 F. boiling residual fraction. The residual fraction (atmospheric reduced crude) is solvent deasphalted to produce a gas oil containing about 1.2 p.p.m. NiO and 2.5 p.p.m. V 0

This feedstock is sent to a catalytic cracker at a temperature of about 900 to 925 F. and a pressure of about 5-15 p.-s.i.g. under fluidizing conditions. The catalyst is one derived from halloysi-te clay by acid activation and impregnation with about 23 parts synthetic alumina gel to 100 parts clay. The catalyst after calcination contains about 51 weight percent Al O and has bulk density of about 0.805 gm./ cc. The cracked products are introduced to a fractionator where approximately 60% gasoline and other low boiling components are recoverd. The residue, including gas-oil fractions, is recycled to the cracker for further processing.

A portion of the silica-alumina catalyst, about 0.2 pound per barrel of feed, is lost from the system With the effluent gases. Another portion of the silica-alumina catalyst, about 0.2 pound per barrel of feed is continuously removed from the catalyst as discard in order to maintain the equilibrium NiO content at about 1000 p.p.m. This amount was determined according to the equation establishing the relationship between the desired equilibrium nickel content level and the nickel content equilibrium nickel content level and the nickel content in the feedstock. Consequently about 0.4 pound of virgin catalyst is added to the cracking reactor for each barrel of fresh feed processed to make up for these catalyst losses The remainder of the catalyst is continuously removed from the cracking reactor and brought to a regenerator and back to the cracker. Average residence time in the re generator is about 10 minutes at a temperature of about 1100 F. to produce a catalyst having a carbon level of about 0.4%. About 20 tons of the cracking catalyst inventory is sent each day as a side stream from the regenerator to demetallization. In the demetallization process the catalyst is held in air for about 3 hours at about 1300 F. and then cooled to about 500 F. The catalyst is then slurried with water containing about 10 to 15 pounds of ammonia per ton a catalyst for about 15 minutes and continuously filtered and washed to remove vanadia. The demetallization procedure removes about 25% of the vanadium from the catalyst and serves, along with the disard and loss of poisoned catalyst, to control the equilibrium vanadium content level at about 1200 p.p.m. V 0 The demetallized catalyst is returned to the regenerator.

Example 11 A 115,000 B/D crude refining operation was modified so that the operation of its vacuum stills and solvent deasphalting unit produced an increase in fluid cracking feedstock from 35,670 B/D to 38,065 B/D with concomitant feedstock metals content increases of from about 0.35 to .60 p.p.m. NiO and about 0.55 to 0.95 p.p.m. V 0 This increased burden of metals in fluid feed is economically handled by the use of a nickel resistant catalyst similar to that described and used in Example I and a simple demetallization process to remove vanadia. The usual stack loss of 0.2 pound of catalyst per barrel of fresh feed maintained the equilibrium nickel content at or below about 1000 p.p.m. NiO with this feedstock as determined by the relationship defined hereinbefore. The vanadia content was controlled by demetallizing about a 7 tons per day slip-stream of regenerated catalyst by continuously air treating at about 1300 F. for about 2 hours. The catalyst is cooled to about 500 F. and slurried with water containing about 10 to 15 pounds of ammonia per ton of catalyst for about 15 minutes and continuously filtered and washed to remove vanadia. The demetallization removes about 15% of the vanadia content from the catalyst and serves to control the equilibrium vanadium content level at about 1200 p.p.m. V 0 The demetallized catalyst is returned to the regenerator.

It is claimed:

1. A method for producing gasoline in a hydrocarbon cracking system having a catalytic cracking zone and a catalyst regeneration zone which comprises cracking at elevated temperature in said cracking zone a hydrocarbon feedstock heavier than gasoline and containing at least about 0.3 p.p.m. nickel and at least about 0.5 p.p.m. vanadium contaminants, measured as the oxides, said cracking being conducted in the presence of a fluidized catalyst consisting essentially of about 10 to 65% total alumina prepared by the addition of about 3 to parts by weight of a synthetic alumina hydrate gel to 100 parts of a solid silica-alumina substrate and during which cracking the catalyst becomes contaminated with nickel and vanadium of said hydrocarbon feedstock, cycling the catalyst between the cracking zone and the catalyst regeneration zone in which latter zone carbon is oxidized at an elevated temperature and thereby removed from the catalyst, discarding from the system that portion of contaminated catalyst which satisfies the equation D X S where Y equals the catalyst discard, Z equals the desired equilibrium nickel content level of the catalyst, X equals the nickel content of the feedstock, D equals the feedstock density and S equals the stack loss, thereby maintaining an equilibrium amount of nickel on the catalyst no greater than about 1000 p.p.m., measured as NiO, bleeding a portion of the vanadium-contaminated catalyst from the cracking system, removing vanadium from the catalyst and returning resulting devanadized catalyst to a hydrocarbon cracking system.

2. The method of claim 1 in which the nickel level on the catalyst is maintained between about 200 and 1000 p.p.m., and in which the bled, substantially carbonfree catalyst is contacted for at least about 15 minutes with a gas containing molecular oxygen at a temperature of about 1150 to 1800 F., to increase subsequent vanadium removal from the said catalyst, said subsequent removal being accomplished by converting the vanadium to a compound selected from the group consisting of volatile vanadium salts and vanadium salts dispersible in an aqueous medium, by contact at elevated temperature with a gas reactive with the vanadium.

3. The method of claim 1 in which the liquid hydrocarbon feedstock contains more than about 0.6 p.p.m. nickel and 0.95 p.p.m. vanadium impurities measured as the oxides.

4. The method of claim 1 in which the normally liquid tydrocarbon feedstock contains more than about 1.0 p.p.m. nickel impurities.

5. The method of claim 1 in which the nickel level on the catalyst is maintained between about 200 and 1000 p.p.m.

References Cited by the Examiner UNITED STATES PATENTS 2,487,576 11/1949 Meyers 20847 2,575,258 11/1951 Corneil et al 252-417 2,935,463 5/1960 Secor et al. 208120 3,122,497 2/1964 Erickson 208120 3,150,103 9/1964 Anderson 20812O ALPHONSO D. SULLIVAN, Primary Examiner.

Claims

1. A METHOD FOR PRODUCING GASOLINE IN A HYDROCARBON CRACKING SYSTEM HAVING A CATALYTIC CRACKING ZONE AND A CATALYST REGENERATION ZONE WHICH COMRPISES CRACKING AT ELEVATED TEMPERATURE IN SAID CRACKING ZONE A HYDROCARBON FEEDSTOCK HEAVIER THAN GASOLINE AND CONTAINING AT LEAST ABOUT 0.3 P.P.M. NICKEL AND AT LEAST ABOUT 0.5 P.P.M. VANADIUM CONTAMINANTS, MEASURED AS THE OXIDES, SAID CRACKING BEING CONDUCTED IN THE PRESENCE OF A FLUIDIZED CATALYST CONSISTING ESSENTAILLY OF ABOUT 10 TO 65% TOTAL ALUMINA PREPARED BY THE ADDITION OF ABOUT 3 TO 100 PARTS BY WEIGHT OF A SYNTHETIC ALUMINA HYDRATE GEL TO 100 PARTS OF A SOLID SILICA-ALUMINA SUBSTRATE AND DURING WHICH CRACKING THE CATALYST BECOMES CONTAMINATED WITH NICKEL AND VANADIUM OF SAID HYDROCARBON FEEDSTOCK, CYCLING THE CATALYST BETWEEN THE CRACKING ZONE AND THE CATALYST REGENERATION ZONE IN WHICH LATTER ZONE CARBON IS OXIDIZED AT AN ELEVATED TEMPEATURE AND THEREBY REMOVED FROM THE CATALYST, DISCARDING FROM THE SYSTEM THAT PORTION OF CONTAMINATED CATALYST WHICH SATISFIES THE EQUATION Y=(D/Z)X-S WHERE Y EQUALS THE CATALYST DISCARD, Z EQUALS THE DESIRED EQUILIBRIUM NICKEL CONTENT LEVEL OF THE CATALYST X EQUALS THE NICKEL CONTENT OF THE FEEDSTOCK, D EQUALS THE FEEDSTOCK DENSITY AND EQUALS THE STOCK LOSS,THEREBY MAINTAINING AN EQUILIBRIUM AMOUNT OF NICKEL ON THE CATALYST NO GREATER THAN ABOUT 1000 P.P.M. MEASURED AS NIO, BLEEDING A PORTION OF THE VANADIUM-CONTAMINATED CATALYST FROM THE CRACKING SYSTEM, REMOVING VANADIUM FROM THE CATALYST AND RETURNING RESULTING DEVANADIZED CATALYST TO A HYDROCARBON CRACKING SYSTEM.