Field of the Invention

This invention is to: a) a catalyst system, b) an improved catalyst system containing specified amounts of water, c) a component of that system comprising certain transition aluminas promoted with a Lewis acid (preferably BF ₃ ) , and d) a catalytic process for the alkylation of isoparaffin with olefins. The catalyst component is produced by contacting the transition alumina with the Lewis acid at relatively low temperatures. The catalyst system comprises that component and an additional amount of free Lewis acid. The process entails olefin/isoparaffin alkylation using the catalyst component and its allied catalyst system.

BACKGROUND OF THE INVENTION

The preparation of high octane blending components for motor fuels using strong acid alkylation processes (notably where the acid is hydrofluoric acid or sulfuric acid) is well-known. Alkylation is the reaction in which an alkyl group is added to an organic molecule, typically an aromatic or olefinic molecule. For production of gasoline blending stocks, the reaction is between an isoparaffin and an olefin. Alkylation processes have been in wide use since World War II when high octane gasolines were needed to

satisfy demands from high compression ratio or supercharged aircraft engines. The early alkylation units were built in conjunction with fluid catalytic cracking units to take advantage of the light end by- products of the cracking units: isoparaffins and olefins. Fluidized catalytic cracking units still constitute the major source of feedstocks for gasoline alkylation units. In spite of the mature state of strong acid alkylation technology, existing problems with the hydrofluoric and sulfuric acid technologies continue to be severe: disposal of the used acid, unintentional emission of the acids during use or storage, substantial corrosivity of the acid catalyst systems, and other environmental concerns. Although a practical alkylation process using solid acid catalysts having little or no corrosive components has long been a goal, commercially viable processes do not exist.

The open literature shows several systems used to alkylate various hydrocarbon feedstocks.

The American Oil Company obtained a series of patents in the mid-1950\'s on alkylation processes involving (preferably Gj or C ₃ ) olefins and C ₄ -C ₈ isoparaffins. The catalysts used were BF ₃ -treated solids and the catalyst system (as used in the alkylation process) also contained free BF ₃ . A summary of those patents is found in the following list:

BF ₃ -Treated Catalyst ^* Patent No. Inventor (with free BF _? ) 2,804,491 May et al. Si0 ₂ stabilized A1 ₂ 0 ₃

(10%-60% by weight BF ₃ )

2,824,146 Kelly et al. metal pyrophosphate hydrate

2,824,150 Knight et al. metal sulfate hydrate

2,824,151 Kelly et al. metal stannate hydrate 2,824,152 Knight et al. metal silicate hydrate . 2,824,153 Kelly et al. metal orthophosphate hydrate

2,824,154 Knight et al. metal tripolyphosphate hydrate

2,824,155 Knight et al. metal pyroarsenate hydrate

2,824,156 Kelly et al. Co or Mg arsenate hydrate

2,824,157 Knight et al. Co, Al, or Ni borate hydrate

2,824,158 Kelly et al. metal pyroantimonate hydrate salt

2,824,159 Kelly et al. Co or Fe olybdate hydrate

2,824,160 Knight et al. Al, Co, or Ni tungstate hydrate

2,824,161 Knight et al. borotungstic acid hydrate or Ni or Cd borotungstate hydrate

2,824,162 Knight et al. phosphomolybdic acid hydrate

2,945,907 Knight et al. solid gel alumina (5%- 100% by weight of Zn or Cu fluoborate, preferably anhydrous) May be supported on A1 ₂ 0 ₃

None of the above patents disclose a process for alkylating olefins and isoparaffins using neat alumina treated with BF ₃ .

Related catalysts have been used to oligo erize olefins. U.S. Patent No. 2,748,090 to

Watkins suggests the use of a catalyst made up of a Group VIII metal (preferably nickel) , a phosphoric acid (preferably containing phosphorus pentoxide) , all placed on an alumina adsorbent, and pretreated with BF ₃ . Alkylation of aromatics is suggested.

U.S. Patent No. 2,976,338 to Thomas suggests a polymerization catalyst comprising a complex of BF ₃ or H ₃ P0 ₄ optionally on an adsorbent (such as activated carbon) or a molecular sieve optionally containing potassium acid fluoride.

Certain references suggest the use of alumina-containing catalysts for alkylation of aromatic compounds. U.S. Patent No. 3,068,301 to Hervert et al. suggests a catalyst for alkylating aromatics using "olefin-acting compounds". The catalyst is a solid, silica-stabilized alumina containing up to 10% Si0 ₂ , all of which has been modified with up to 100% by weight of BF ₃ . None of these prior references suggest either the process or the material used in the processes as is disclosed here.

Other BF ₃ -treated aluminas are known. For instance, U.S. Patent No. 3,114,785 to Hervert et al. suggests the use of a BF ₃ -modified, substantially anhydrous alumina to shift the double bond of 1-butene to produce 2-butene. The preferred alumina is substantially anhydrous gamma-alumina, eta-alumina, or theta-alumina. The various aluminas will adsorb or complex with up to about 19% by weight fluorine depending upon the type of alumina and the temperature of treatment. The aluminas are treated with BF ₃ at elevated temperatures. Hervert et al. does not suggest using these catalysts in alkylation reactions.

In addition, U.S. Patent No. 3,131,230 to

Hervert et al. describes a process for the alkylation

of aromatic compounds which utilizes a catalyst comprising boron trifluoride and boron trifluoride modified, substantially anhydrous alumina. This reference teaches that the activity of the catalyst is maintained by introducing water in an amount up to 400 parts per million molal and boron trifluoride in an amount up to 3200 parts per million molal in the hydrocarbon feed. Although this reference teaches that the modification of the alumina with the boron trifluoride gas may be carried out in a range between room temperature and up to about 300°C, it is noted that this step is highly exothermic. For example, when the modification is carried out at room temperature, a temperature wave will travel through the alumina causing the temperature to increase up to about 150°C or more.

In U.S. Patent No. 4,407,731 to Imai, a high surface area metal oxide such as alumina (particularly gamma-alumina, eta-alumina, theta-alu ina, silica, or a silica-alumina) is used as a base or support for BF ₃ . The BF ₃ treated metal oxide is used for generic oligomerization and alkylation reactions. The metal oxide is treated in a complicated fashion prior to being treated with BF ₃ . The first step entails treating the metal oxide with an acid solution and with a basic aqueous solution. The support is washed with an aqueous decomposable salt such as ammonium nitrate. The support is washed using deionized H ₂ 0 until the wash water shows no alkali or alkaline earth metal cations in the filtrate. The support is dried and calcined. The disclosure suggests generically that BF ₃ is then introduced to the treated metal oxide support. The examples show introduction of the BF ₃ at elevated temperatures, e.g, 300°C or 350°C.

Similarly, U.S. Patent No. 4,427,791 to Miale et al. suggests the enhancement of the acid catalytic activity of inorganic oxide materials (such as alumina or gallia) by contacting the material with ammonium fluoride or boron fluoride, contacting the treated inorganic oxide with an aqueous ammonium hydroxide or salt solution, and calcining the resulting material. The inorganic oxides treated in this way are said to exhibit enhanced Brδnsted acidity and, therefore, are said to have improved acid activity towards the catalysis of numerous reactions (such as alkylation and iso erization of various hydrocarbon compounds) . A specific suggested use for the treated inorganic oxide is as a matrix or support for various zeolite materials ultimately used in acid catalyzed organic compound conversion processes.

U.S. Patent No. 4,751,341 to Rodewald shows a process for treating a ZSM-5 type zeolite with BF ₃ to reduce its pore size, enhance its shape selectivity, and increase its activity towards the reaction of oligomerizing olefins. The patent also suggests using these materials for alkylation of aromatic compounds. Certain Soviet publications suggest the use of A1 ₂ 0 ₃ catalysts for alkylation processes. Benzene alkylation using those catalysts (with 3 ppm to 5 ppm water and periodic additions of BF ₃ ) is shown in Yagubov, Kh. M. et al., Azerb. Khim. Zh. _f 1984, (5) p. 58. Similarly, Kozorezov, Yu and Levitskii, E.A. , Zh. Print. Khim. (Leningrad) . 1984, 57. (12), p. 2681, show the use of alumina which has been heated at relatively high temperatures and modified with BF ₃ at 100°C. There are no indications that BF ₃ is maintained in excess. Isobutane alkylation using A1 ₂ 0 ₃ /BF ₃ catalysts is suggested in Nef ekhimiya. 1977, 12 (3), p. 396; 1979, 19 (3), p. 385. The olefin is ethylene. There

is no indication that BF ₃ is maintained in excess during the reaction. The crystalline form of the alumina is not described.

U.S. Patent No. 4,918,255 to Chou et al. suggests a process for the alkylation of isoparaffins and olefins using a composite described as "comprising a Lewis acid and a large pore zeolite and/or a non- zeolitic inorganic oxide". The process disclosed requires isomerization of the olefin feed to reduce substantially the content of alpha-olefin and further suggests that water addition to the alkylation process improves the operation of the process. The best Research Octane Number (RON) product made using the inorganic oxides (in particular Si0 ₂ ) is shown by Table 6 therein to be 94.0.

Similarly, PCT published applications WO 90/00533 and 90/00534 (which are based in part on the U.S. patent to Chou et al. noted above) suggest the same process as does Chou et al. WO 90/00534 is specific to a process using boron trifluoride-treated inorganic oxides including "alumina, silica, boria, oxides of phosphorus, titanium oxide, zirconium oxide, chromia, zinc oxide, magnesia, calcium oxide, silica- alumina-zirconia, chromia-alumina, alumina-boria, silica-zirconia, and the various naturally occurring inorganic oxides of various states of purity such as bauxite, clay and diatomaceous earth". Of special note is the statement that the "preferred inorganic oxides are amorphous silicon dioxide and aluminum oxide". The examples show the use of amorphous silica (and BF ₃ ) to produce alkylates having an RON of no greater than 94. None of these disclosures shows crystalline transition aluminas which were promoted with Lewis acids at lower temperatures nor any effect upon the NMR spectrum because of such a treatment. Nor do these

disclosures show their use in isoparaffin/olefin alkylation. These disclosures further do not show any benefit to the alkylation of isoparaffins and olefins using these specifically treated aluminas.

BRIEF DESCRIPTION OF THE DRAWINGS

Figures 1A-1D are nuclear magnetic resonance

(NMR) plots of certain transition aluminas treated with

BF ₃ at a range of temperatures. Figure 2 is a three-dimensional graph showing octane sensitivity for the inventive process as a function of olefin feed content.

Figure 3 is an FTIR spectrum showing the effect of the thermal treatment on the free water content (1500-1650 cm ^"1 band) and hydroxyl content

(3300-3800 cm\' ¹ ) of gamma alumina as a function of temperature.

Figures 4A and 4B are graphs showing the effect of specific amounts of free water on the alumina surface on catalyst life (amount of olefin processed or aged) and formula octane number (FON or R+M/2) achieved in the alkylation of a isobutene and mixed butene feed stream.

Figure 5 is a graph showing the effect of initial water and performance in alkylation in terms of

%C ₈ of alkylate product, RON of alkylate in the mixture, catalyst life (amount of olefin processed or age) and the total amount of olefin processed prior to significant C ₈ yield decline.

SUMMARY OF THE INVENTION

This invention is variously a catalyst component comprising one or more transitional aluminas which are treated with one or more Lewis acids (preferably BF ₃ ) at a fairly low temperature desirably

so low that the component exhibits specific NMR spectra, an improved catalyst system containing transition alumina and one or more Lewis acids and a limited amount of water, a catalyst system comprising that catalyst component with free Lewis acid, and an olefin/isoparaffin alkylation process step using these catalyst systems.

Use of the catalyst systems, i.e., the catalyst component in conjunction with free Lewis acid, produces high octane alkylate from isobutane and butene at a variety of reaction temperatures between -30°C and 40°C. The catalyst\'s high activity can result in low operating costs because of its ability to operate at high space velocities as well as enhanced alkylate production.

DESCRIPTION OF THE INVENTION This invention is:

A) a catalyst component comprising certain Lewis acid treated transition aluminas,

B) a catalyst system comprising the catalyst component in combination with at least a minor amount of free Lewis acid,

C) a catalyst component (A) or system (B) containing limited amounts of water, and

D) an alkylation process for producing branched paraffinic products from olefins and isoparaffins using that catalyst system.

The Catalyst Component

The catalyst component of this invention comprises or consists essentially of a major amount of transition aluminas (preferably eta- or gamma-alumina) which has been treated with a Lewis acid, preferably BF ₃ . The catalyst component preferably does not

contain any metals (except, of course, aluminum and any metal associated with the Lewis acid such as the semi- metal boron) in catalytic amounts capable of hydrogenating the hydrocarbons present in the feeds except those impurity metals which may be present in trace amounts in the Lewis acid or the alumina.

Alumina

Aluminum oxide (alumina) occurs abundantly in nature, usually in the form of a hydroxide in the mineral bauxite, along with other oxidic impurities such as Ti0 ₂ , F& ₂ 0 ₃ , and Si0 ₂ . The Bayer process is used to produce a reasonably pure A1 ₂ 0 ₃ having a minor amount of Na ₂ 0. The Bayer process may be used to produce a variety of alumina hydroxides:

The aluminum hydroxides may then be treated by heating to produce various activated or transition aluminas. For instance, the aluminum hydroxide known as boehmite may be heated to form a sequence of 5 transition phase aluminas: gamma, delta, theta, and finally, alpha (see Wefers et al., "Oxides and Hydroxides of Alumina", Technical Paper No. 19, Aluminum Company of America, Pittsburgh, PA, 1972, pp.1-51) . 10

Transition aluminas (and their crystalline forms) include:

gamma tetragonal delta orthorhombic/tetragonal

15 eta cubic theta monoclinic chi cubic/hexagonal kappa hexagonal lambda orthorhombic

20 Activated aluminas and aluminum hydroxides are used in various chemical processes as catalyst and adsorbents.

The aluminas suitable for use in this process include the noted transition aluminas: gamma, delta, eta, theta, chi, kappa, rho, or lambda. Especially

₂₅ preferred are gamma- and eta-aluminas. Mixtures of the two are also desireable.

Since it is difficult to produce a substantially pure single phase transition alumina, mixtures of various aluminas are tolerable so long as a

_ ₀ major amount of the specified alumina, e.g., an amount greater than about 50% by weight of the alumina present in the catalyst, is present in the catalyst. For instance, in the production of eta-alumina, gamma- alumina is often concurrently present in the resulting

_ ₅ product. Indeed, x-ray diffraction analysis can only

difficultly detect the difference between the two phases. Aluminum hydroxides (boehmite, gibbsite, etc.) may be present in the predominately transition phase product in more than trivial amounts so long as they do not substantially affect the desired alkylation reaction.

While the surface area of the alumina may suitably vary over a wide range dependent on the specific type of transition alumina employed, best results in the alkylation process of the invention are associated with the use of aluminas having surface areas in excess of about 160 m ² /g. Accordingly, transition aluminas having surface areas above 160 m ² /g are preferred with aluminas having surface areas in the range of about 200 to about 400 m ² /g being most preferred.

The alumina may be produced in any appropriate form such as pellet, granules, bead, sphere, powder, or other shape to facilitate its use in fixed bed, moving bed, slurry, or fluidized bed reactors.

Lewis Acids

The catalyst component of this invention contains one or more Lewis acids in conjunction with the alumina noted above. A Lewis acid is a molecule which can form another molecule or an ion by forming a complex in which it accepts two electrons from a second molecule or ion. Typical strong Lewis acids include boron halides such as BF ₃ , BC1 ₃ , BBr ₃ , and BI ₃ ; antimony pentafluoride (SbF _s ) ; aluminum halides (A1C1 ₃ and

AlBr ₃ ) ; titanium halides such as TiBr ₄ , TiCl ₄ , and TiCl ₃ ; zirconium tetrachloride (Zrcl ₄ ) ; phosphorus pentafluoride (PF ₅ ) ; iron halides such as FeCl ₃ and

FeBr ₃ ; and the like. Weaker Lewis acids such as tin, indium, bismuth, zinc, or mercury halides are also acceptable. Preferred Lewis acids are boron containing materials (BF ₃ , Bcl ₃ , Bbr ₃ , and BI ₃ ) , SbF ₅ , and A1C1 ₃ ; most preferred is BF ₃ .

It is believed that the Lewis acid forms complexes or surface compounds with the alumina substrate. In particular, we believe that BF ₃ strongly adsorbs in the vicinity of the hydroxyl groups found on the alumina surface and additionally is physi-sorbed at the alumina surface.

The total amount of Lewis acid in the alumina surface is between 0.5% and 40% by weight of the catalyst depending in large measure on two factors: the Lewis acid chosen and the susceptibility of the alumina surface to accepting the Lewis acid by chemisorption or by physisorption. In the case of BF ₃ , we believe that 5-20% of the weight of the alumina catalyst component is attributable to BF ₃ products (e.g., the production of aluminum fluoroborate or similar compounds) and the remainder is physi-sorbed BF ₃ . Preferably, the total amount of BF ₃ (as BF ₃ products) added is in excess of 7% by weight of the alumina catalyst component and, most preferably, from about 10% to about 20% by weight of the alumina catalyst component.

To maintain the presence of sufficient Lewis acid on the catalyst composition, we have found it desirable to maintain at least a minor amount of the Lewis acid in the proximity of the alumina surface, preferably in the reaction fluid. This amount is an amount at least sufficient to maintain the concentration of the Lewis acid specified above on the alumina. At the WHSV ranges specified below with regard to the alkylation reaction, we have found that

generally an amount of at least 0.5% of Lewis acid (by weight based on the hydrocarbon) is sufficient to maintain the Lewis acid level on the alumina. For BF ₃ it is preferred to use BF ₃ concentrations of about 0.8% to about 15% (by weight based on hydrocarbon) with BF ₃ amounts in the range of about 1.5% to about 6.0% by weight being most preferred. On an alumina basis, the ratio of free Lewis acid (that is, Lewis acid in the proximity of the alumina but not associated with the alumina by chemisorption or physisorption) to alumina is in the range of 0.05 to 30 g Lewis acid/g A1 ₂ 0 ₃ . For BF ₃ , the preferred range is 0.08 to 10 g BF ₃ /g A1 ₂ 0 ₃ , and more preferably in the range of 0.10 to 8 g BF ₃ /g A1 ₂ 0 ₃ .

Catalyst Component Preparation

The catalyst component may be prepared in a variety of ways including preparation in situ in, e.g., an alkylation reactor by passing the Lewis acid in gaseous form through the vessel containing the transition alumina. Alternatively, the alumina may be contacted with the Lewis acid and later introduced into the reactor.

In any case, the alumina may be substantially dry or anhydrous prior to contact with the Lewis acid and maintained in a state of dryness, i.e., maintained at a very low free H ₂ 0 content. The alumina phase chosen in conjunction with proper treatment of the alumina to maintain the presence of hydroxyl groups (usually by maintaining the alumina at temperatures below 450°C during pretreatment) allows the presence of about 4-10 hydroxyl groups per 100 A ² of alumina surface area. Preferred is 6-10 hydroxyl groups per

100 A ² of alumina surface area. The alumina is preferably completely hydroxylated since that

hydroxyla ion, in turn, permits the formation of the maximum amount of the Al-OH-Lewis acid complex, believed to be one element of the active alkylation catalyst at the alumina surface. The alumina may be partially or substantially dehydroxylated but the catalyst is not as efficacious.

Alumina, as received from the manufacturer or exposed to the atmosphere for appreciable periods of time, picks or adsorbs substantial water. Careful heating and control of the atmosphere surrounding the alumina is consequently desirable. Suitably, the alumina is heated at a temperature below 500°C and preferably at a temperature in the range of about 50°C to about 400°C prior to treatment with the Lewis acid. Additionally, free water (in distinction to the water which may be identified as hydroxyl groups on the alumina surface) may be present in limited amounts in the alumina. The free water content in the alumina is suitably less than about 15% by weight but preferably is less than about 8.0% by weight. Most preferably the free water content of the alumina is between 0.5 and 6.0% by weight. Higher amounts of water appear both to degrade the catalyst and to impair the effectiveness of the catalyst in the practice of the alkylation reaction. Higher amounts of water also tend to form compounds, such as BF ₃ hydrates, which are corrosive and therefore undesirable.

The use of Lewis acid promoted transition aluminas having the limited free water content as set forth above in a process of alkylating lower olefin with isobutane has several surprising benefits. This catalyst component or catalyst composition improves the octane number of the resulting alkylate, the percentage of C ₈ in alkylate, and the effective life of the catalyst before regeneration. We have not observed

these advantages when water is added to the alkylation process feedstock.

Lewis acid-alumina, contact temperatures between -25°C and less than about 150°C are acceptable; a temperature between -25°C and 100°C is desirable; a temperature between -30°C and 30°C is preferred. The partial pressure of gaseous Lewis acid added to the alumina is not particularly important so long as a sufficient amount of Lewis acid is added to the alumina. We have found that treatment of the alumina with BF ₃ at the noted temperatures will result in an alumina-BF ₃ complex containing BF ₃ sufficient to carry out the alkylation. The alumina contains between 0.5% and 30% by weight of BF ₃ . We have observed that solid state boron-nuclear magnetic resonance ( ⁿ B-NMR) analysis of the catalyst component provides evidence (a pronounced peak at about -21.27 ppm relative to boric acid) of tetragonal boron in the catalyst composite produced at the lower temperatures. Aluminas treated with BF ₃ at temperatures of 150°C and higher show spectra which are indicative of the presence of trigonal symmetry about the boron. We prefer catalysts in which the relative amounts of trigonal boron:tetragonal boron (as calculated by the integration of the respective "B-NMR spectra) are in the range of 0:1 to 1:1. More preferred is the range of 0:1 to 0.25:1; most preferred is 0:1 to 0.1:1.

Obviously, the alumina may be incorporated into a binder prior to its treatment with Lewis acid. The binders may be clays (such as montmorillonite and kaolin) or silica based materials (such as gels or other gelatinous precipitates) . Other binder materials include carbon and metal oxides such as alumina, silica, titania, zirconia, and mixtures of those metal oxides. The composition of the binders is not

particularly critical but care must be taken that they not substantially interfere with the operation of the alkylation reaction.

The preferred method for incorporating the catalytic alumina into the binder is by mixing an aluminum hydroxide precursor (such as boehmite) with the binder precursor, forming the desired shape, and calcining at a temperature which both converts the aluminum hydroxide precursor into the appropriate transition phase and causes the binder precursor to bind the alumina particles. The absolute upper temperature limit for this calcination is about 1150°C. Temperatures below about 1000°C may be appropriate.

Alkylation Process

The inventive catalyst component and the allied catalyst composition are especially suitable for use in alkylation processes involving the contact of an isoparaffin with an olefin. The catalyst component should be used in conjunction with an amount of free Lewis acid.

Specifically, the catalyst system (the inventive catalyst component in combination with a free Lewis acid) is active in alkylation reactions at low temperatures (as low as -30°C) as well as at higher temperatures (nearing 50°C) . Lower temperatures (-5°C to 15°C) are preferred because of the enhanced octane of the alkylate produced and are particularly preferred if the feedstream contains more than about 1% isobutylene. Higher temperatures also tend to produce larger amounts of polymeric materials.

The pressure used in this process may be between atmospheric pressure and about 750 psig. Higher pressures within the range allow recovery of excess reactants by flashing after the product stream

leaves the alkylation reactor. The amount of catalyst used in this process depends upon a wide variety of disparate variables. Nevertheless, the Weight Hourly Space Velocity ("WHSV" = weight of olefin feed/hour ÷ weight of catalyst) may effectively be between 0.1 and 120, especially between 0.5 and 30. The overall molar ratio of isoparaffin to olefin may be between about 1:1 to 50:1. With recycle reactors the paraffin to olefin ratio could be substantially higher and could exceed 1000:1. The preferred range is between 2:1 to 25:1; the more preferred range is between 3:1 to 15:1.

The feedstreams introduced to the catalyst are desirably chiefly isopar? ffins having from four to ten carbon atoms and, most preferably, four to six carbon atoms. Isobutane is most preferred because of its ability to make high octane alkylate. The olefins desirably contain from three to twelve and preferably from three to five carbon atoms, i.e., propylene, cis- and trans-butene-2, butene-1, and amylene(s) . Preferably, the olefin stream contains little (if any) isobutylene. Similarly, for the inventive catalysts the process works better in producing high octane alkylate if the feedstream contains little or no butadiene (preferably less than 0.2% to 0.3% molar of the total olefins) and a minimal amount of isobutylene, e.g., less than about 2.5% molar based on the olefins. Although the catalyst alkylates butene-1, it is preferred to operate with a minimum of butene-1, e.g., less than about 10% by mol, since it lowers the octane values of the resulting alkylate. Of course, if it is desired to operate a process with high throughput rather than with highest octane, a higher level of butene-1 is tolerable. An excellent source of a feedstock containing a low level of isobutylene is the

raffinate from a process which produces methyl-t- butylether (MTBE) .

The water content of the feedstocks may vary within wide limits, but preferably is at a low level. The water content should be less than about 200 ppm (by weight) and most preferably less than about 50 ppm (by weight) . Higher levels of water content tend to lower the octane value of the resulting alkylate and form corrosive hydrates or reaction products with the Lewis acids. Because the sources of most alkylation unit feedstocks tend to introduce water into those feeds, we prefer to dry one or more of the feedstocks to achieve the preferred water content.

The feedstocks should contain a minimum of oxygenates such as ethers and alcohols. Oxygenates appear to lessen substantially the effectiveness of the catalyst system.

The process of this invention includes increasing the effective catalyst life by conducting the alkylation process with isobutane and lower olefins using catalyst components and catalyst compositions having the limited free water content specified above in the discussion of preparation of the transition alumina catalyst compound and catalyst composition. The products of all variations of this alkylation process typically contain a complex mixture of highly branched alkanes. For instance, when using isobutane as the alkane and n-butylene as the olefin, a mixture of 2,2,3-; 2,2,4-; 2,3,3-; and 2,3,4- trimethylpentane (TMP) will result often with minor amounts of other isomeric or polymeric products. The 2,3,4-TMP isomer is the lowest octane isomer of the noted set. The 2,2,3- and 2,2,4-TMP isomers are higher octane components. Calculated average octane values (the average of the Research Octane Number (RON) and

the Motor Octane Number (MON), as denoted by (R + M)/2) of the various C _g isomers are:

Isomer Octane (R + M) /2

2,2,3- 104.8 2,2,4- 100.0 2,3,3- 102.8 2,3,4- 99.3

The process may be carried out in the liquid, vapor, or mixed liquid and vapor phase. Liquid phase operation is preferred.

The invention has been disclosed by direct description. Below may be found a number of examples showing various aspects of the invention. The examples are only examples of the invention and are not to be used to limit the scope of the invention in any way.

EXAMPLES

Example 1 Catalyst Testing

This example shows the preparation of a number of alumina-based catalysts in situ and their subsequent use in an alkylation reaction using model feeds. It is used to evaluate catalyst activity and selectivity.

The alumina samples were dried at 110°C overnight and charged to a semi-batch reactor having an internal volume of about 500 cc. The reactor temperature was controllable over the range of -5°C to 40 ^Q C For initial catalyst treatment, the reactor containing the catalyst was purged with an inert gas and cooled to about 0°C. About 275 cc of isobutane was

added to the reactor. After a brief degassing, BF ₃ was added batchwise. After BF ₃ is added, the temperature of the reactor rises and the pressure typically drops as the alumina adsorbs or reacts with the BF ₃ . Additional infusions of BF ₃ are made until the pressure in the reactor no longer drops. The BF ₃ saturation equilibrium pressure was about 40 psig. The liquid phase concentration of BF ₃ was about 1.5%. At that point the alumina had adsorbed or reacted with all of the BF ₃ possible at that temperature and the catalyst was in its most active form.

A 4/1 molar mixture of isobutane and trans-2- butene was added to the reactor at a WHSV of 3.5 until the paraffin to olefin ratio reached 25. The product alkylate was then removed from the reactor vessel and analyzed using gas-liquid chromatography.

The results of those runs are shown in Table 1.

Table 1

It is clear from these preliminary screening data that the transition (gamma and delta) aluminas produce significantly higher percentages of C ₈ in the product alkylate than do the other aluminum hydroxide

catalysts. The result did not appear to correlate to the specific surface area of the catalyst.

Example 2 Catalyst Screening

This example compares the performance of eta- alumina (a preferred form of the inventive catalyst) with representative samples of other acidic oxides each combined with BF ₃ for the reaction of isobutane with butylenes to produce alkylate.

The eta-alumina sample was prepared by a controlled thermal treatment of bayerite (Versal B from LaRoche Chemical) for 15 hours at 250°C and 24 hours at 500°C under a N ₂ atmosphere. The comparative oxidic materials were: silica-alumina, synthetic mordenite zeolite, and fumed silica. The silica-alumina (obtained from Davison Chemical) contained 86.5% Si0 ₂ and had a surface area of 392 m ² /gm. It was used without further treatment. The mordenite was a hydrogen form zeolite and was obtained from Toyo Soda. It was prepared from Na- mordenite and subjected to ion exchange, steam treatment, and calcination to achieve a Si/Al ratio of 28:1. Each of the samples was dried at 110°C overnight and introduced into the semi-batch reactor described in Example 1. The samples were purged with a dry inert gas and cooled to 0°C. Isobutane was added to the reactor to an initial volume of 100 cc. BF ₃ was added with stirring until an equilibrium pressure of 30 psig was obtained.

A mixture of isobutane/t-2-butene was fed to the reactor. At the completion of the reaction, alkylate was removed and analyzed by gas-liquid chromatography. The RON were calculated from the gas-

liquid chromatography data using the well-known correlations in Hutson and Logan, "Estimate Alky Yield and Quality", Hydrocarbon Processing, September, 1975, pp. 107-108. The summary of the experiments and results is shown in the following Table 2:

Catalyst charge (g) Temperature (°C) i-C ₄ charge ml (initial) i-C ₄ /C ₄ =feed ratio (molar) Space velocity (WHSV) Run time (minutes) i-C ₄ /C ₄ =(final) Butene conversation (%) Produ

C ₈ saturates

C ₉ ⁺ TMP/Cg total (%) Yield (w/w) RON Octane (R + M/2)

♦estimated

Clearly, for the eta-alumina catalyst, the yield of C _g ^, s was significantly higher; the overall yield and RON were much better.

Example 3

This example shows that the addition of either water or methanol produces no appreciable improvement on the alkylation of butene-2 with isobutane using the inventive alumina catalyst. Indeed, water and methanol appear to be detrimental.

Three separate semi-batch reactors were dried and flushed with nitrogen. A sample of 2.5 gm of a gamma-alumina (LaRoche VGL) was loaded into each bottle. The alumina samples had been previously dried at 110°c overnight. An amount of 0.278 gms of deionized water was added dropwise to one reactor. An amount of 0.988 gms of methanol was added to another reactor. These amounts were calculated to be 10% of the catalyst plus water equivalent. The remaining reactor was used as a control reactor. Isobutane (246 cc) was added to each bottle; BF ₃ was added (with stirring) until the pressure reached a constant 30 psig. A feedstock of isobutane and 2-butene was continuously added in a ratio of 2:1 and at a rate of 1.6 cc/minute. The reaction continued for about 75 minutes after which samples of the reactor liquids were removed and analyzed using a gas-liquid chromatograph. The conversion of olefin was more than 99% in each case. Other reaction conditions and a summary of reaction results are shown in Table 3:

Table 3

Reaction Conditions

Reaction temperature (°C)

Pressure (psig)

WHSV

I/O (w/w)

Product

C ₅ -C ₇

C ₈ (saturated)

TMP/Cg

Yield (w/w)

RON

(R + M) /2

It is clear that neither water nor methanol created any advantage in the operation of the process in producing a gasoline alkylate. The gross amounts of C ₈ produced were smaller than for the inventive alumina; the amount of undesirable C, ₂ ⁺ were two to four times higher than for the inventive alumina. The yields were lower and, probably most importantly, the octane values of the comparative products were significantly lower.

Example 4 This example shows the suitability of the inventive catalyst (gamma-alumina, LaRoche GL) for a variety of olefin feedstocks. The following reaction conditions were used for the test series:

Temperature 0°c Total pressure 30 psig WHSV 4

A semi-batch reactor was utilized in each run.

The olefin feedstocks were mixtures which were chosen to allow us to identify desirable and undesirable combinations of feed materials. The mixtures are shown in Table 4:

Table 4

60/40 mixture of Mixture No. 1-C ₄ = i-C ₄ = C ₃ = cis/trans 2-C ₄ =

1 25 25 20 30

2 10 25 20 45

3 25 10 20 45

4 10 10 20 60

5 25 25 05 45

6 10 25 05 60

7 25 10 05 60

8 10 10 05 75

The products made were analyzed using gas- liquid chromatography and their respective octane numbers are shown in Table 5:

2 n ON σ> co O o o ω o as σ. CM ω oo oo as CO CO en

O

%

This data shows that increases in isobutene and propylene feed concentrations directionally cause the inventive alkylation process to produce lower alkylate C _g content. As shown in Figure 2, smaller amounts of either C ₃ ⁼ or i- C ₄ ⁼ cause no more harm to alkylate quality but are generally undesirable if extremely high octane alkylates are necessary.

Example 5 This example demonstrates the performance of the transition alumina/BF ₃ catalysts in reacting isobutane with butenes to form high octane product under conditions of high space velocity and low paraffin/olefin feed ratios. A sample of gamma-alumina (VGL, LaRoche) was dried overnight at 110°C and loaded into the semi- continuous reactor unit described in Example 1. The catalyst was purged with dry inert gas and cooled to 0°C. Isobutane was added to the reactor and then the system was exposed to BF ₃ under stirring conditions until an equilibrium pressure of 30 psig was achieved. A feed comprising pure trans-2-butene was then pumped into the reactor under vigorous stirring conditions over a period of 60 minutes; samples were obtained periodically during the run (at 30 and 60 minutes) . The results are summarized in Table 6 below:

Example 6 This example shows the utility of the catalyst system on a feed obtained from a refinery MTBE unit. The feed, containing minor amounts of butadiene and isobutene, was introduced into a bed of a commercial hydroisomerization catalyst (0.3% Pd on A1 ₂ 0 ₃ ) at 400 cc/hr, 80°C, and 350 psig along with 14 seem H ₂ . The molar ratio of H ₂ :butadiene was 6:1. The thusly treated feed, containing no butadiene and 0.52 % (molar) of isobutene, was mixed with an appropriate amount of isobutane. The mixture had an approximate composition as shown in Table 7 below:

This mixture had an isoparaffin to olefin ratio of 5.8:1 and an isobutane/olefin ratio of 5.7:1.

The mixture was then admitted to a pair of continuous laboratory reactors each containing 280 cc of liquid and containing 5.04 g of catalyst. The temperature was maintained at 0°F. The WHSV for the reactor was 4.3 hr ^"1 and the LHSV was 1.07 hr" ¹ . The catalyst was a gamma alumina (LaRoche VGL) and was prepared by adding the proper amount to the reactors along with a small amount of isobutane, pressuring the reactor to about 40 psig of BF _3/ and maintaining that pressure for the duration of the test. The test was run for 41 hours total time. The catalyst was regenerated four times during the run by rinsing the catalyst in 200cc of trimethyl pentane, heating to 150°C in air for 45 minutes to volatilize a portion of the reaction product on the catalyst, and heating the catalyst to 600°C in air for 60 minutes to oxidize the remaining hydrocarbonaceous materials. Small amounts of the catalyst were added as necessary with the

regenerated catalyst to restore the catalyst to its proper amount upon return to the reactor (0.41 g 8 cycle 2, 0.97 g @ cycle 3, 0.0 g @ cycle 4, and 0.47 g § cycle 5). About 4.5 liters (3.2 kg) of stripped C ₅₊ alkylate was collected having about 7.6% C ₅ _ ₇ , 81.2% C ₈ , 4.4% C^ _u , and 6.8% C (all by weight). Using the Hutson method discussed above, the octanes were calculated to be: RON = 96.6, MON = 93.3, and the (R+M)/2 = 94.95. The product was then engine-tested using API methodology and the octanes were measured to be: RON = 98.7, MON = 93.85. The resulting (R+M)/2 = 96.28. The Hutson method clearly underestimated the RON octane values for this process.

Example 7

This example shows the preparation of a number of BF ₃ /alumina-based catalyst components. One is made in accord with this invention and three are comparative samples. Each of the samples was then tested in an alkylation reaction using model feeds and isobutane and butenes.

All four gamma-alumina samples (LaRoche- Versal GL) were infused with BF ₃ in a Cahn balance. Use of a Cahn balance allowed close control of the temperature at which the alumina contacted the BF ₃ and further allowed the weight gain to be measured during the treatment. The four samples were treated with BF ₃ respectively at 25°C, 150°C, 250°C, and 350°C. A sample of each of the catalyst components was removed and analyzed using ⁿ B- MAS-NMR. The results of these analyses are shown in the figures: Figure 1A shows the treated alumina at 25°C; Figures IB, 1C, and ID show the respective data from the 150°C, 250°C, and 350°C treated aluminas. In Figure 1A, there is a pronounced

sharp peak at about -21.27 ppm (relative to boric acid) suggesting significant tetragonal boron content. The other three NMR plots do not show such a sharp peak. Instead, the data suggest the presence of substantial trigonal boron.

The four gamma-alumina samples (LaRoche- Versal GL) were then charged to a semi-batch reactor having an internal volume of about 500 cc. The reactor temperature was controllable over the range of -5°C to 40°C. For initial catalyst treatment, the reactor containing the catalyst was purged with an inert gas and cooled to about 0°C. About 275 cc of isobutane were added to the reactor. After a brief degassing, BF ₃ was added batchwise. Additional infusions of BF ₃ are made until the pressure in the reactor no longer drops. The BF ₃ saturation equilibrium pressure was about 40 psig. The liquid phase concentration of BF ₃ was about 1.5%.

A mixture of isobutane/trans-2-butene was fed to the reactor. At the completion of the reaction, alkylate was removed and analyzed by gas-liquid chromatography. The RON\'S were calculated from the gas-liquid chromatography data using the well-known correlations in Hutson and Logan, "Estimate Alky Yield and Quality", Hydrocarbon Processing, September, 1975, pp. 107-108. The summary of the experiments and results is shown in Tables 8 and 9 below:

Table 8

o

It is clear from the data that the catalyst component treated with BF ₃ at the lower temperature is superior in operation in most practical aspects (conversion, C ₈ production, alkylate yield, RON, MON, etc.) than the other materials.

Example 8 This example compares the results of a variety of catalyst components and compositions (comprising a transition alumina and BF ₃ and a variety of water contents) when used in an alkylation process. The intent was to compare the water content—whether the water was in the form of surface hydroxyl content or in the form of "free" or surface-adsorbed water—in the catalyst to the octane number and C _g content of the resulting alkylate.

A large batch of commercial alumina (LaRoche- V-GL-Versal-250- a gamma alumina apparently made by calcination of pseudo-boehmite) was transferred from the as-received can into glass evaporating dishes. The glass evaporating dishes are maintained in a drying oven at 110°C.

This alumina is believed to be fully hydroxylated as it is received from LaRoche. By "fully hydroxylated" is meant that substantially each surface octahedral aluminum is terminated by an "-OH" group. In addition, there likely is some surface bound ^O (or

"free water" as was discussed above) depending upon the temperature, relative humidity, and handling history of the material.

We determined the temperature at which the surface water was removed (without substantial dehydroxylation) by placing an alumina sample taken from the drying oven in a Fourier Transform Infra-Red (FTIR) analyzer and made a series of scans at

progressively higher temperatures (25°C, 80°C, 125 ^β C, 175°C, and 225°C) under a dry helium purge stream.

By following the progression of the H-O-H bending band of line A in Figure 3 at 1640 cm" ¹ , it may be observed that dehydration is essentially complete between 175°C and 225°C. Although some small amount of dehydroxylation likely would occur as a result of the rise to this temperature, to a good approximation, the loss in weight occurring up to 200°C is equal to the amount of surface bound or "free" water on the surface of the alumina sample.

Samples of the alumina were then subjected to treatment with water vapor so to load specific amounts of water onto the alumina surface. This was done by placing the alumina samples in a closed vessel with distilled water held at a specific temperature. The liquid water was not placed in contact with the alumina but instead was suspended in the vessel creating an atmosphere containing water in equilibrium at the chosen temperature. The samples were held in the vessel for two hours each to allow equilibration between the alumina and the water vapor. The temperatures of treatment were variously at 0°C, 18°C, and 30°C. These three samples and a sample taken directly from the drying oven were each placed in a microbalance boat and the weight loss to 200°C determined.

In addition, each of the samples was heated in the microbalance to 1075°C, a temperature at which substantially complete hydroxylation is achieved. At temperatures above about 200°C, the samples continue to lose weight. This is thought to be due to the condensation of A1-0H to form H ₂ 0, Al ⁺ , and 0 ⁼ .

The results of these runs are shown in Table ιo below:

Several catalyst samples, after treatment with the temperature and water vapor treatments specified above, were then subjected to an alkylation reaction to check the correspondence between the products produced and the respective water contents. Care was taken to prevent evaporative water from being introduced into the catalyst. The seven catalysts contained: 18% Kfl, 6% H ₂ 0, 3% H ₂ 0 (i.e., after removal from drying oven) , fully dehydrated at 2.1 meq Al-OH/gm (after 200°C pretreatment).

The alkylation reaction utilized a model feed of isobutane and mixed butenes (1/0=6:1; where the butenes were trans-2-butene=94%, l-butene=5%, and isobutene=l%) . The feed had been twice dried using freshly regenerated 3A zeolite beds to lower the water content of the feed to less than 10 ppm. Commercial QJ s purchased from vendors typically contain 20-40 ppm of H ₂ 0. The BF ₃ concentration was held at about 1.8 wt.% of the total liquid weight. The reaction pressure was 45 psig; residence time was 56 minutes; the catalyst concentration was 1.5%; and the reaction temperature was 0°C.

The results of these runs are shown in Figures 4A and 4B. The catalyst having full hydroxylation and 3% free (or surface) water clearly is superior both in %C ₈ produced and in the octane of the resulting alkylate.

Example 9 This example shows the interrelated effects of water BF ₃ /transition alumina catalysts on the age of that catalyst and C,\'s in alkylate product.

In this example, four catalysts based on the transition alumina utilized in Example 8 (gamma-phase alumina - LaRoche VGL) were treated also as discussed

in Example 8 to produce aluminas having 1 to 1.5% HjO, 3% H ₂ 0, and 7% H ₂ 0 (two batches) .

The reactor was a Hastelloy autoclave operated in CSTR mode, with continuous addition of feed and BF ₃ and with continuous withdrawal of product alkylate along with isobutane and BF ₃ . The feed was treated both with 3A and 13X molecular sieves to remove water and other impurities.

Each of the catalysts was used in an alkylation process operated in the following procedure:

The alumina was loaded into a tube and pretreated with humidified nitrogen at a temperature necessary to achieve a desired water content in the alumina. At the start-up of the CSTR reactor, an equilibrium mixture of isobutane and alkylate was cooled to 0°C and pressurized with BF ₃ to 50 pounds pressure (gauge) . The pretreated alumina was introduced into the reactor through a port using liquid isobutane to carry the alumina. The reaction was initiated by introducing the 6:1 isobutene/olefin feed derived from a hydroisomerized MTBE raffinate. The olefin composition was:

butene-2 92.0% butene-1 4.8% isobutene 3.2%

The reaction conditions were: 0°c, 3% catalyst slurry concentration and a WHSV of 8.7 to 9. Samples were taken at several times throughout the course of the reaction. The reaction was run until a clear pattern of deactivation was observed.

As is shown in Figure 5, the catalyst containing 3% H ₂ 0 maintained its ability to produce C ₈ ^, s

for a much longer period of time or catalyst age. In a qualitative side, the optimization of moisture content of the catalyst produced an alkylate having a specific C ₈ content for about 50% longer than catalysts containing either 1.5% or 7% H ₂ 0.

It should be clear that one having ordinary skill in this art would envision equivalents to the processes found in the claims that follow and that these equivalents would be within the scope and spirit of the claimed invention.