More than thirty years ago, many household laundry detergents were made of branched alkylbenzene sulfonates (BABS). BABS are manufactured from a type of alkylbenzenes called branched alkylbenzenes (BAB).

Alkylbenzenes (phenyl-alkanes) refers to a general category of compounds having an aliphatic alkyl group bound to a phenyl group and having the general formula of (mj-alkylj) i-n-phenyl-alkane. The aliphatic alkyl group may also consist of one or more alkyl group branches designated by the corresponding" (m ;- alkyl The standard process used by the petrochemical industry for producing BAB consists of oligomerizing light olefins, particularly propylene, to branched olefins having 10 to 14 carbon atoms and then alkylating benzene with the branched olefins in the presence of a catalyst such as HF. Although the product BAB comprises a large number of alkyl-phenyl-alkanes having the general formula (mj-alkylj) i-n-phenyl-alkane, two examples of BAB are m-alkyl-m-alkyl-n- phenyl-alkanes where m W n, and m-alkyl-m-phenyl-alkanes where m ! 2.

The most prominent common characteristic of BAB is that, for a large proportion of BAB, there is attached to the aliphatic alkyl chain of BAB generally at least one alkyl group branch, and more commonly three or more alkyl group branches. If any alkyl group branch itself is branched, then the aliphatic alkyl group in BAB has even more primary carbon atoms. Thus the aliphatic alkyl group in BAB usually has three, four, or more primary carbon atoms. Each alkyl group branch is usually a methyl group branch, although ethyl, propyl, or higher alkyl group branches are possible.

Another typical characteristic of BAB is that the attachment of the phenyl group to any non-primary carbon atom of the aliphatic alkyl chain. Except for 1- phenyl-alkanes whose formation is known to be disfavored due to the relative instability of the primary carbenium ion and neglecting the relatively minor effect of the branches of the branched paraffins, the oligomerization step produces a carbon-carbon double bond that is randomly distributed along the length of the aliphatic alkyl chain, and the alkylation step nearly randomly attaches the phenyl group to a carbon along the aliphatic alkyl chain. Thus, for example, a BAB that has an aliphatic alkyl chain having 10 carbon atoms would be expected to be an approximately random distribution of 2-, 3-, 4-, and 5-phenyl-alkanes, and the selectivity to 2-phenyl alkane would be 25 if the distribution was perfectly random, but is typically between 10 and 40.

BAB's commonly have one of the quaternary carbon as one of the carbon atoms of the aliphatic alkyl group. The quaternary carbon may bind by a carbon- carbon bond to a carbon in the phenyl group. When it does the molecule is referred to as a"quaternary alkyl-phenyl-alkane"or simply hereinafter a"quat." and comprise alkyl-phenyl-alkanes having the general formula m-alkyl-m-phenyl- alkane."End quats"constitute 2-alkyl-2-phenyl-alkanes where the second carbon atom from an end of the alkyl side chain is a quat. Quats containing the quaternary carbon in other locations result in an alkyl-phenyl-alkane referred to as an"internal quat. "Known processes for producing BAB, create a relatively high proportion of internal quats, typically greater than 10 mol-%.

Household laundry detergents made of BABS were gradually polluting rivers and lakes. It was found that BABS were slow to biodegrade. The use linear alkylbenzene sulfonates (LABS), which biodegrade more rapidly than BABS reduced the problem. LABS are manufactured from linear alkylbenzenes (LAB). The petrochemical industry produces LAB by dehydrogenating linear paraffins to linear olefins which are then alkylated with benzene in the presence of HF or a solid catalyst. LABs comprise a linear aliphatic alkyl group and a phenyl group and have the general formula n-phenyl-alkane. LAB has no alkyl group branches and normally two primary carbon atoms (i. e. , n > 2). The standard LAB process attaches the phenyl group to any secondary carbon atom

of the linear aliphatic alkyl group. HF catalyst alkylation is slightly more likely to attach the phenyl group to a secondary carbon near the center of the linear aliphatic alkyl group. LAB produced by the DetalTM process contains approximately 25-35 mol-% of n-phenyl-alkanes as 2-phenyl-alkanes.

Recent research has identified modified alkylbenzene sulfonates"MABS", which differ from all alkylbenzene sulfonates used currently in commerce, and from all alkylbenzene sulfonates produced by prior alkylbenzene processes, including those catalyzed by HF, aluminum chloride, silica-alumina, fluorided silica-alumina, zeolites, and fluoride zeolites. MABS relative to LABS have improved laundry cleaning performance, hard surface cleaning performance, and excellent efficiency in hard and/or cold water, while also having biodegradability comparable to that of LABS.

MABS can be produced by sulfonating a third type of alkylbenzenes called modified alkylbenzenes (MAB), and the desired characteristics of MAB are determined by the desired solubility, surfactancy, and biodegradability properties of MABS. MAB comprises a large number of phenyl-alkanes, some of which may be found in LAB and BAB, but the matching phenyl-alkanes are not desirable phenyl-alkanes for MAB. The phenyl-alkanes in MAB are phenyl- alkanes comprising a lightly branched aliphatic alkyl group and a phenyl group and has the general formula (mj-alkylj) i-n-phenyl-alkane. Phenyl-alkanes in MAB usually have only one alkyl group branch and where n w 1, the MAB has three primary carbons. A preferred MAB phenyl-alkane is a monomethyl-phenyl- alkane. However, the MAB may have two primary carbon atoms if there is only one alkyl group branch and n = 1, or, if there are two alkyl group branches and n 1, four primary carbons. Thus, the first characteristic of MAB is an average number of primary carbons in the aliphatic alkyl groups of the phenyl-alkanes intermediate that in BAB and in LAB. A high proportion of 2-phenyl-alkanes also characterizes MAB, namely that from 40 to 100% of phenyl groups selectively attach to the second carbon atom of the alkyl side chain.

As a final characteristic MAB alkylate has a relatively low proportion of internal quats, typically less than 10 mol-%. Some internal quats such as 5-

methyl-5-phenyl-undecane produce MABS with slower biodegradation. MABS with end quats such as 2-methyl-2-phenyl-undecane show biodegradation similar to LABS. See the article entitled"Biodegradation of Coproducts of Commercial Linear Alkylbenzene Sulfonate,"by A. M. Nielsen et al., in Environmental Science and Technology, Vol. 31, No. 12,3397-3404 (1997).

PCT International Publication Nos. WO 99/05082, WO 99/05084, 99/05241, and WO 99/05243, all four of which were published on February 4, 1999, disclose alkylation processes for uniquely lightly branched or delinearized alkylbenzenes. PCT International Publication No. W099/07656, published on February 18,1999, discloses processes for such alkylbenzenes using adsorptive separation.

Because of the advantages of MABS over other alkylbenzene sulfonates, catalysts and processes are sought that produce MAB with a selectivity to 2-phenyl-alkanes and selectivity away from internal quaternary phenyl-alkanes.

SUMMARY OF THE INVENTION In one aspect, this invention is a process for the production of phenyl- alkanes, in particular modified alkylbenzenes (MAB), by adsorptive separation, dehydrogenation, and alkylation. The process is characterized by the composition of the adsorbent and desorbent pair used in the process. The adsorbent used is silicalite and the desorbent comprises a C5-C8 linear paraffin, a C5-C8 cycloparaffin, and/or preferably a branched paraffin such as isooctane.

In a process embodiment a paraffinic feed stream comprising a first concentration of a first acyclic paraffin having a carbon number range of C8- C28 and 2 or 3 primary carbon atoms and a second acyclic paraffin pass to an adsorption zone. The adsorption zone comprises a bed of an adsorbent comprising silicalite at adsorption promoting conditions to selectively adsorb at least portion of the acyclic paraffin having 2 or 3 primary carbon atoms. A desorbent stream comprising at least one of a C5-C8 cycloparaffin, a C5-C8 normal paraffin, and a C5-C8 branched paraffin contacts the bed of adsorbent.

An adsorption extract having a second concentration of the first acyclic

hydrocarbon that is greater than the first concentration is recovered from the adsorption zone. At least a portion of the adsorption extract passes to a dehydrogenation zone that operates at dehydrogenation conditions sufficient to dehydrogenate the first acyclic paraffin. A dehydrogenated product stream comprising a C8-C28 acyclic monoolefin having 2 or 3 primary carbon atoms is recovered from the dehydrogenation zone. An aromatic feedstock comprising a phenyl compound and at least a portion of the dehydrogenated product stream pass to an alkylation zone that operates at alkylation conditions sufficient to alkylate the phenyl compound with the acyclic monoolefin in the presence of an alkylation catalyst. The alkylation zone allows recovery of a phenyl-alkane comprising a molecule having one phenyl portion and one C8-C28 aliphatic alkyl portion with 2 or 3 primary carbon atoms and no quaternary carbon atoms except for quats. The alkylation has a selectivity to 2-phenyl-alkanes of from 40 to 100 and a selectivity to internal quaternary phenyl-alkanes of less than 10. In a preferred embodiment, the alkylation has a selectivity to non-quats of less than 10, and more preferably less than 1.

In a preferred process embodiment this invention produces an MAB composition comprising phenyl-alkanes having one phenyl group and one aliphatic alkyl group. The the phenyl-alkanes futher have an average weight of the aliphatic alkyl groups of the phenyl-alkanes of between the weight of a Clo aliphatic alkyl group and a C13 aliphatic alkyl group; a content of phenyl-alkanes having the phenyl group attached to the 2-and/or 3-position of the aliphatic alkyl group of greater than 55 wt-% of the phenyl-alkanes ; and an average level of branching of the aliphatic alkyl groups of the phenyl-alkanes of from 0.25 to 1.3 alkyl group branches per phenyl-alkane molecule when the sum of the contents of 2-phenyl-alkanes and 3-phenyl-alkanes is more than 55 wt-% and less than or equal to 85 wt-% of the phenyl-alkanes, or an average level of branching of the aliphatic alkyl groups of the phenyl-alkanes of from 0.4 to 1.3 alkyl group branches per phenyl-alkane molecule when the sum of the concentrations of 2- phenyl-alkanes and the 3-phenyl-alkanes is greater than 85 wt-% of the phenyl- alkanes. In addition the aliphatic alkyl groups of the phenyl-alkanes comprise primarily linear aliphatic alkyl groups and mono-branched aliphatic alkyl groups,

and wherein the alkyl group branches on the aliphatic alkyl chain of the aliphatic alkyl groups comprise primarily small substituents, such as methyl group branches, ethyl group branches, or propyl group branches, and wherein the alkyl group branches attach to any position on the aliphatic alkyl chain of the aliphatic alkyl groups provided that phenyl-alkanes having at least one quaternary carbon atom on the aliphatic alkyl group comprise less than 20% of the phenyl-alkanes.

This invention, when used for detergent alkylation, produces detergents that meet the increasingly stringent requirements of 2-phenyl-alkane selectivity and internal quaternary phenyl-alkane selectivity for the production of MAB that can be sulfonate to produce MABS with improved cleaning effectiveness in hard and/or cold water and biodegradability comparable LAS.

In another aspect the process of this invention produces particular MAB and MABS product compositions with specially tailored carbon branching that differs from prior art processes. In another aspects these MAB and MABS produced may be used as a lubricant or a lubricant additive, respectively.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows an embodiment of the invention.

Figure 2 is a concentration profile for a pulse test separation.

DETAILED DESCRIPTION OF THE INVENTION The process uses a feed mixture comprising a paraffin and a feedstock comprising a phenyl compound are consumed in the subject process. The feed mixture comprises acyclic paraffins having 8 to 28 carbon atoms. The acyclic paraffin is preferably a"lightly branched paraffin, "which as used herein, refers to a paraffin having three or four primary carbon atoms no quaternary carbon atoms. Normally, the lightly branched paraffin has a total number of from 9 to 16 carbon atoms, preferably from 10 to 14 carbon atoms, and highly preferably from 10 to 13 carbon atoms. The lightly branched paraffin generally comprises an aliphatic alkane having the general formula of (pj-alkylj) i-alkane.

The lightly branched paraffins comprise generally more than 30 mol-% and preferably more than 70 mol-%, of the feed mixture. Branching alkyl groups generally comprise methyl, ethyl, and propyl groups with shorter branches being preferred. Preferably, the lightly branched paraffin has only one alkyl group branch and comprise preferably more than 85 mol-% of the total lightly branched paraffins. Lightly branched paraffins having either two alkyl group branches or four primary carbon atoms comprise generally less than 30 mol-%, and preferably less than 15 mol-%, of the total lightly branched paraffins.

The feed mixture may also contain one or more nonbranched (linear) or normal paraffin molecules having a total number of carbon atoms per paraffin molecule of generally from 8 to 28 carbon atoms, and highly preferably from 10 to 13 carbon atoms. The concentration of nonbranched paraffins in the feed mixture is often above 0.3 mol-%.

In addition to lightly branched and nonbranched paraffins, other more highly branched acyclic compounds may be in the feed mixture. However, on dehydrogenation such highly branched paraffins tend to form highly branched monoolefins which on alkylation tend to form BAB. For example, paraffin molecules consisting of at least one quaternary carbon atom tend on dehydrogenation followed by alkylation to form phenyl-alkanes that have in the aliphatic alkyl portion a quaternary carbon atom that is not a quat. Therefore, the quantity of these highly branched paraffins charged to the process is preferably minimized. Paraffin molecules consisting of at least one quaternary carbon atom generally comprise less than 10 mol-%, preferably less than 5 mol-%, more preferably less than 2 mol-%, and most preferably less than 1 mol-% of the feed mixture.

The production of the feed mixture is not an essential element of this invention, and any suitable method for producing the feed mixture may be used.

Since the carbon number range of the feed mixture desired for the production of MAB is normally between 9 and 16. This range corresponds to paraffins boiling in the kerosene boiling point range, therefore kerosene fractions form suitable feed mixture precursors. Kerosene preparation methods are inherently imprecise and produce a mixture of compounds. The feed mixtures to the process may contain

quantities of paraffins having multiple branches and paraffins having multiple carbon atoms in the branches, cycloparaffins, branched cycloparaffins, or other compounds having boiling points relatively close to the desired compound isomer.

Kerosene fractions contain a very large number of different hydrocarbons and the feed mixture to the subject process can therefore contain 200 or more different compounds including sizable quantities of aromatic hydrocarbons. Fractions recovered from crude oil by fractionation will typically require hydrotreating for removal of sulfur and/or nitrogen prior to being fed to the subject process.

It is expected, however, that separation rather than oligomerization or other forms of synthesis will provide a lower cost adequate feed mixture and will therefore be the predominate source of the feed mixture. A preferred method for the production of the feed mixture is the separation of nonbranched (linear) hydrocarbons or lightly branched hydrocarbons from a kerosene boiling range petroleum fraction. Several known processes that accomplish such a separation are known. One process, the UOP Mole process, is an established, commercially proven method for the liquid-phase adsorption separation of normal paraffins from isoparaffins, cycloparaffins, and aromatics using the UOP Sorbex separation technology. Another suitable, established, and proven process is the UOP Kerosene IsosivTM Process, which employs vapor-phase adsorption for separating normal paraffins from nonnormal paraffins using molecular sieves in an adsorber vessel. See Chapters 10.3, 10.6 and 10.7 in the book entitled Handbook of Petroleum Refining Process, Second Edition, edited by Robert A. Meyers, published by McGraw-Hill, New York, 1997.

The raffinate stream of an adsorptive separation process, such as the UOP Mole process which selectively recovers the nonbranched (linear) paraffins in an extract stream, is an especially preferred feed mixture for the subject process.

The raffinate stream from such a process will be free of contaminants such as sulfur or nitrogen containing compounds, and will also have a suitably low concentration of nonbranched paraffins and olefins. The use of such a raffinate stream as the feed mixture allows integration of the subject process into an existing LAB facility, with the two adsorptive separation steps being performed in series. The separately recovered normal paraffin stream and feed mixture can

then be processed in a variety of ways. For instance, each of the nonbranched paraffin stream and the feed mixture could be processed independently via dehydrogenation and aromatic alkylation to produce two separate products.

Alternatively, the nonbranched paraffin stream and the feed mixture could be used to form a desired paraffin blend. That is, the stream charged to the dehydrogenation zone of the subject process can comprise the product of the separation zone of the subject process plus from 10 to 50vol-% nonbranched paraffins.

The composition of a mixture of linear, lightly branched, and branched paraffins as well as olefins can be determined by analytical methods that are well-know to a person of ordinary skill in the art of gas chromatography. The article written by H. Schulz, et al. and published starting at page 315 of the Chromatographia 1,1968, describes a temperature-programmed gas chromatograph apparatus and method that is suitable for identifying components in complex mixtures of paraffins or olefins. A person of ordinary skill in the art can separate and identify the components in a mixture of paraffins using essentially the apparatus and method described in the article by Schulz et al.

The aromatic-containing feedstock to the subject process comprises a phenyl compound, which is benzene when the process is detergent alkylation.

In a more general case, the phenyl compound of the aromatic feedstock may be alkylated or otherwise substituted derivatives or of a higher molecular weight than benzene, including toluene, ethylbenzene, xylene, phenol, naphthalene, etc., The adsorptive separation section recovers acyclic, lightly branched paraffins from the feed mixture. This separation can be performed in a batch or continuous mode including the use of two or more adsorbent beds in cyclic operation. In this mode one or more beds are used for the separation while another bed is being regenerated. Significant operational and economic advantages accrue to performing the separation on a continuous basis. Simulated moving bed technology (SMB) is the preferred method of achieving continuous operation and uniform products. The separation preferably recovers monomethyl paraffins from the feed mixture.

SMB adsorptive separation units for simulating movement of the adsorbent relative to the feed stream are well known. This simulation is performed using established commercial technology wherein the adsorbent is held fixed in place as a number of subbeds retained in one or more cylindrical adsorbent chambers.

The positions at which the streams involved in the process enter and leave the chambers are slowly shifted from subbed to subbed along the length of the adsorbent chambers so that the streams enter or leave different subbeds as the operational cycle progresses. Normally there are at least four streams (feed, desorbent, extract and raffinate) employed in this procedure, and the location at which the feed and desorbent streams enter the chamber and the extract and raffinate streams leave the chamber are simultaneously shifted in the same direction at set intervals. Only one line is normally employed for each subbed, and each bed line carries one of the four process streams at some point in the cycle.

Cyclic advancement of the input and output streams of this simulation can be accomplished by a manifolding system or by rotary disc valves. This simulation usually includes the use of a variable flow rate pump which pushes liquid leaving one end of the adsorbent vessel (s) to the other end in a single continuous loop.

Simulated moving bed processes typically include at least three or four separate steps which are performed sequentially in separate zones within a mass of adsorbent. Each of these zones normally is formed from a plurality of subbeds with the number of beds per zone ranging from 2 or 3 up to 8 to 10. The most widely practiced commercial process units typically contain 24 beds. All of the beds are contained in one or more vertical vessels referred to herein collectively as the adsorbent chamber. The general technique employed in the performance of a simulated moving bed adsorptive separation is well described at page 70 of the September 1970 edition of Chemical Engineering Progress (VoL 66, No 9). as shown in US-A-3,040, 777 and US-A-3,422, 848.

During the adsorption step of the process a feed mixture containing a mixture of compounds is contacted with the adsorbent at adsorption conditions and one or more compound (s) or a class of compounds is selectively adsorbed and retained by the adsorbent while the other compounds of the feed mixture are relatively unabsorbed. Normally the desired compound is adsorbed. The feed

mixture may contain a large variety of compounds including isomers of the desired compound. For instance, a mixed xylene feed stream may contain ethylbenzene and/or Cg aromatics and can be processed to recover a specific isomer by a suitable adsorbent/desorbent pair operated at suitable conditions. Differing sieve/desorbent combinations are used for different separations. For instance, X zeolites, specifically X zeolites exchanged with barium or barium and potassium ions at their exchangeable sites, are the preferred adsorbents for p-xylene recovery from xylene mixtures.

In the next step of the process, the unabsorbed (raffinate) components of the feed mixture are then removed from the void spaces between adsorbent particles and from the surface of the adsorbent as a raffinate stream. The adsorbed compound is then recovered from the adsorbent by contact a stream comprising the desorbent material at desorption conditions in a desorption step.

The desorbent displaces the desired compound to form an extract stream, which is normally transferred to a fractionation zone for recovery of the desired compound from desorbent of the extract stream. In some instances the desired product of the process can be in the raffinate stream rather than the extract stream.

For purposes of this description, various terms used herein are defined as follows. A"feed mixture"is a mixture containing one or more extract components and one or more raffinate components to be separated by the adsorption section of the subject process. The term"feed stream"indicates a stream of a feed mixture passed into contact with the adsorbent. An"extract component"is a compound or class of compounds that is more selectively adsorbed by the adsorbent. A"raffinate component"is a compound or class of compound that is less selectively adsorbed. The term"desorbent material"means a material capable of desorbing an extract component. The term"raffinate stream"means a stream for removed from the adsorbent bed after the adsorption of extract compounds that can vary from essentially 100% desorbent material to essentially 100% raffinate components. The term"extract stream"means a stream desorbed by a desorbent material and removed from the adsorbent bed and varying from essentially 100% desorbent material to essentially 100% extract components.

Separation means, typically fractional distillation columns, recover all or portions of the extract stream and the raffinate stream. The stream containing the undesired compound may be recycled to isomerization. The extract stream may be rich in the desired compound or may only contain an increased concentration.

When used relative to a process stream the term"rich"is intended to indicate a concentration of the indicated compound or class of compounds greater than 50 mole %.

It has become customary in the art to group the numerous beds in the adsorption chambers into a number of zones. Zone I, the adsorption zone, makes contact between the feed stream and the adsorbent. Zone II, the purification zone, removes undesired isomers as raffinate. In Zone III, the desorption zone, desorbent releases the desired isomers from the adsorbent for recovery in the extract stream. Zone IV contains quantity of adsorbent located between Zone I and Zone III that segregates Zones I and III and partially removes desorbent from the adsorbent. The liquid flow through Zone IV prevents contamination of Zone III by Zone I liquid by flow cocurrent to the simulated motion of the adsorbent from Zone III toward Zone I. A more thorough explanation of simulated moving bed processes is given in the Adsorption, Liquid Separation section of the Kirk-Othmer Encyclopedia of Chemical Technoloav.

The objectives of this invention are achieved by employing a novel adsorbent-desorbent pair comprising a silicalite adsorbent and a desorbent containing a branched paraffin; a linear paraffin and/or cycloparaffin ; a linear paraffin and a branched paraffin; or a linear paraffin, a cycloparaffin, and a branched paraffin. The preferred desorbent is a Ce to C8 branched paraffin.

The preferred branched paraffin for the desorbent is isooctane.

The preferred adsorbent comprises silicalite. Silicalite is well described in the article,"Silicalite, A New Hydrophobic Crystalline Silica Molecular Sieve, " Nature, Vol. 271, Feb. 9,1978 which is incorporated herein by reference for its description and characterization of silicalite. Silicalite is a hydrophobic crystalline silica molecular sieve having an MFI type structure of intersecting bent-orthogonal channels formed with two cross-sectional geometries, 6A circular and 5.1-5. 7 A elliptical on the major axis. This gives silicalite great selectivity as a size selective

molecular sieve. Due to its aluminum free structure composed of silicon dioxide silicalite does not show ion-exchange behavior. Thus silicalite is not a zeolite.

The practice of the subject invention requires no significant variation in operating conditions, adsorbent or desorbent composition within the adsorbent chambers or during different process steps. That is, the adsorbent preferably remains at the same temperature and pressure throughout the process.

The active component of the adsorbent is normally used in the form of small agglomerates having high physical strength and attrition resistance. The agglomerates contain the active adsorptive material dispersed in an amorphous, inorganic matrix referred to as the binder and having channels and cavities therein which enable fluid access to the adsorptive material. Methods for forming the crystalline powders into such agglomerates include the addition of an inorganic binder, generally a clay comprising a silicon dioxide and aluminum oxide, to a high purity adsorbent powder in a wet mixture. Silica is a suitable binder. The adsorbent particles may be in the form of extrudates, tablets, macrospheres or granules having a desired particle range, preferably from 16 to 60 mesh (Standard U. S. Mesh) (1.9 mm to 250 microns). Clays of the kaolin type, water permeable organic polymers or silica are generally used as binders.

Those skilled in the art will appreciate that the performance of a particular adsorbent is often greatly influenced by a number of variables not related to its composition such as operating conditions, feed stream composition, and the water content of the adsorbent. One such variable is the water content of the adsorbent which is expressed herein in terms of the recognized Loss on Ignition (LOI) test.

In the LOI test the volatile matter content of the zeolitic adsorbent is determined by the weight difference obtained before and after drying a sample of the adsorbent at 500°C under an inert gas purge such as nitrogen for a period of time sufficient to achieve a constant weight. For the subject process it is preferred that the water content of the adsorbent results in an LOI at 900°C of less than 7.0 wt-% and preferably within the range of from 0 to 4.0 wt-%.

A silicalite or other microporous active component of the adsorbent will ordinarily be in the form of small crystals present in the adsorbent particles in amounts ranging from 75 to 98 wt-% of the particle based on volatile-free

composition. Volatile-free compositions are generally determined after the adsorbent has been calcined at 900°C in order to drive off all volatile matter. The remainder of the adsorbent will generally be the inorganic matrix of the binder.

The present invention, passes a feed mixture comprising one or more monomethyl branched hydrocarbons and at least one nonnormal hydrocarbon of like carbon number but different structure through one or more beds of adsorbent that selectively adsorbing the monomethyl branched hydrocarbon and passes other components through the adsorption zone. At some point in time the flow of the feed stream through the adsorbent bed stops and the adsorption zone is flushed to remove nonadsorbed materials surrounding the adsorbent. Thereafter passing a desorbent through the adsorbent bed recovers the desired isomer.

The selectivity, (B), of an adsorbent/desorbent pair is defined as the ratio of the two components in the adsorbed phase divided by the ratio of the same two components in the unabsorbed phase at equilibrium conditions. Relative selectivity is given by the equation: Selectivity = wt. percent C/wt. percent DA wt. percent C/wt. percent Du where C and D are two components of the feed stream represented in weight percent and the subscripts A and U represent the adsorbed and unabsorbed phases, respectively. The equilibrium conditions are determined when the feed stream passing over a bed of adsorbent does not change composition, in other words, when there is no net transfer of material occurring between the unabsorbed and adsorbed phases. Relative selectivity can be expressed not only for one feed stream compound as compared to another but can also be expressed between any feed mixture component and the desorbent material.

An important characteristic of an adsorbent is the rate of exchange of the desorbent for the extract component of the feed mixture or, in other words, the relative rate of desorption of the extract component. Faster rates of exchange reduce the amount of desorbent material needed to remove the extract component and consequently the operating cost of the process. Exchange rates are often temperature dependent. Ideally, desorbent materials should have a

selectivity for extract components equal to or slightly less than 1 to desorb extract components as a class with reasonable flow rates of desorbent material, and so that extract components can displace desorbent material in subsequent adsorption steps.

In continuous liquid phase adsorptive separation processes the desorbent material must be judiciously selected to satisfy many criteria. Expressed in terms of the selectivity, the adsorbent should have more selectivity for all extract components with respect to a raffinate component than for the desorbent material with respect to a raffinate component. Desorbent materials must also be compatible with the particular adsorbent and the particular feed mixture and have a reasonable cost.

Adsorption conditions in general include a temperature range of from 20°C to 250°C, with from 40°C to 150°C being more preferred. Temperatures from 80°C to 140°C are highly preferred. Adsorption conditions also preferably include a pressure sufficient to maintain the process fluids in liquid phase; which may be from atmospheric to 600 psi (g). Desorption conditions generally include the same temperatures and pressure as used for adsorption conditions.

The preferred desorbent comprises a mixture of normal paraffin, a cycloparaffin (naphthene) and/or a branched paraffin. Preferred cycloparaffins are cyclopentane, cyclohexane and methyl cyclohexane. The preferred normal paraffins are n-pentane and n-hexane. Normal paraffins are strong desorbents and n-hexane is actually the strongest desorbent of these compounds. A blend of normal paraffins and cycloparaffins or of normal paraffins and isooctane, is often desirable to adjust the strength of the desorbent stream. These blends may contain from 10 to 90 vol-% cycloparaffin or isooctane, with the remainder being the normal paraffin.

The extract stream comprises paraffins having a total number of carbon atoms per paraffin molecule of generally from 8 to 28, preferably from 8 to 15, and more preferably from 10 to 15 carbon atoms. The extract stream contains a higher concentration of lightly branched paraffins, based on the total paraffins in the extract stream, than the concentration of lightly branched paraffins in the

feed mixture, based on the total paraffins in the feed mixture. The lightly branched paraffins having either two alkyl group branches or four primary carbon atoms comprise generally less than 60 mol-%, preferably less than 30 mol-%, and more preferably less than 15 mol-%, of the total lightly branched paraffins in the portion of the extract stream that passes to the dehydrogenation zone of the process. The lightly branched paraffins having either one alkyl group branch and more desirably a methyl group comprise preferably more than 85 mol-% of the total lightly branched paraffins in the portion of the extract stream charged to the dehydrogenation zone. When present in the extract stream with the lightly branched paraffins, the linear paraffin content should be as no more than, 75 mol-% of the total paraffins in that portion of the extract stream that is charged to the dehydrogenation zone. Paraffin molecules consisting of at least one quaternary carbon atom generally comprise less than 10 mol-%, preferably less than 5 mol-%, more preferably less than 2 mol-%, and most preferably less than 1 mol-%, of that portion of the extract stream that passes to the dehydrogenation zone.

The dehydrogenation section may be configured substantially in the manner shown in the drawing. Briefly, a stream containing paraffins combines with recycled hydrogen to form a dehydrogenation reactant stream that is heated and contacted with a dehydrogenation catalyst in a fixed bed maintained at dehydrogenation conditions. The effluent of the fixed catalyst bed, which is referred to herein as the dehydrogenation reactor effluent stream, is cooled, partially condensed, and passed to a vapor-liquid separator. The vapor-liquid separator produces a hydrogen-rich vapor phase and a hydrocarbon-rich liquid phase. The condensed liquid phase recovered from the separator passes to a stripping column, which removes all compounds which are more volatile than the lightest hydrocarbon which is desired to be passed to the alkylation section.

The olefin-containing net stream that passes from the dehydrogenation section to the alkylation section of the process is referred to herein as the dehydrogenated product stream.

This invention is not limited to any one particular flow scheme for the dehydrogenation section which may include moving or fixed bed

dehydrogenation catalyst, catalyst-containing reaction zones with heat exchangers there between and introduction of hot hydrogen-rich gas streams.

Hydrocarbons may contact any catalyst bed in an upward-, downward-, or radial- flow fashion.

Dehydrogenation catalysts are well known in the prior art as exemplified by US-A-3, 274,287 ; US-A-3, 315,007 ; US-A-3,315, 008; US-A-3,745, 112; US-A- US-A-4, 430,517 ; US-A-4,716, 143; US-A-4,762, 960; US-A-4,786, 625; and US-A- 4,827, 072. It is believed that the choice of a particular dehydrogenation catalyst is not critical to the success of this invention. The preferred catalyst is a layered composition of an inner core bonded to an outer layer comprising a refractory inorganic oxide containing at least one uniformly dispersed platinum group (Group VIII (IUPAC 8-10)) metal and at least one promoter metal, and where at least one modifier metal is dispersed on the catalyst composition. Preferably, the outer layer is bonded to the inner core to the extent that the attrition loss is less than 10 wt-% based on the weight of the outer layer.

The dehydrogenation conditions are selected to minimize cracking and polyolefin by-products. Typical dehydrogenation conditions will not result in any appreciable isomerization of the hydrocarbons in the dehydrogenation reactor.

The hydrocarbon may contact the catalyst the liquid phase, mixed vapor-liquid phase, or preferably in the vapor phase. Dehydrogenation conditions include a temperature of generally from 400°C (752°F) to 900°C (1652°F) and preferably from 400°C (752°F) to 525°C (977°F), a pressure of generally from 1 kPa (g) (0.15 psi (g) ) to 1013 kPa (g) (147 psi (g) ), and a LHSV of from 0.1 to 100 hr'. As used herein, the abbreviation"LHSV"means liquid hourly space velocity, which is defined as the volumetric flow rate of liquid per hour divided by the catalyst volume. Generally for normal paraffins, the lower the molecular weight the higher the temperature required for comparable conversion. The dehydrogenation zone maintain pressure as low as practicable, usually less than 345 kPa (g) (50 psi (g) ), consistent with equipment limitations, to maximize chemical equilibrium advantages.

The extract stream may be admixed with a diluent material before, while, or after flowing to the dehydrogenation zone. The diluent material may be

hydrogen, steam, methane, ethane, carbon dioxide, nitrogen, argon, and the like, or a mixture thereof. Hydrogen is the preferred diluent. Ordinarily, when hydrogen is utilized as the diluent it is utilized in amounts sufficient to ensure a hydrogen to hydrocarbon mole ratio of 0.1 : 1 to 40: 1.

Water or a material decomposable to water at dehydrogenation conditions such as an alcohol, aldehyde, ether, or ketone may be added to the dehydrogenation zone continuously or intermittently in an amount calculated on the basis of equivalent water of 1 to 20,000 weight ppm of the extract stream. 1 to 10,000 weight ppm of water addition gives best results when dehydrogenating paraffins having from 2 to 30 or more carbon atoms.

The dehydrogenated product stream is typically a mixture of unreacted paraffins, linear (unbranched) olefins, and branched monoolefins including lightly branched monoolefins. Typically from 0 to 75 mol-%, and preferably from 0 to 50 mol-%, of the olefins in the monoolefin-containing stream from the paraffin dehydrogenation process are linear (unbranched) olefins. The dehydrogenated product may also contain monoolefins having a total number of carbon atoms of from 8 to 28 with preferably less than 10 mol-%, and preferably less than 1 mol- %, of the monoolefins having quaternary carbon atoms.

The dehydrogenated product stream may comprise a highly branched monoolefin or a linear (unbranched) olefin, but is preferably a lightly branched monoolefin. The term"lightly branched monoolefin,"as used herein, refers to a monoolefin having a total number of carbon atoms and of from 8 to 28, of which three or four of the carbon atoms are primary carbon atoms and none of the remaining carbon atoms are quaternary carbon atoms. Preferably, the lightly branched monoolefin has a total number of from 8 to 15 carbon atoms, and more preferably from 10 to 15 carbon atoms.

The lightly branched monoolefin generally comprises an aliphatic alkene having the general formula of (pj-alkylj) i-q-alkene. The lightly branched monoolefin may be an alpha monoolefin or a vinylidene monoolefin, but is normally an internal monoolefin. As used herein, the term"alpha olefins"refers to olefins having the chemical formula, R-CH=CH2. The term"internal olefins,"

as used herein, includes di-substituted internal olefins having the chemical formula R-CH=CH-R; tri-substituted internal olefins having the chemical formula R-C (R) =CH-R; and tetra-substituted olefins having the chemical formula R- C (R) =C (R) -R. The di-substituted internal olefins include beta internal olefins having the chemical formula R-CH=CH-CH3. As used herein, the term "vinylidene olefins"refers to olefins having the chemical formula R-C (R) =CH2.

Suitable lightly branched monoolefins include octenes, nonenes, decenes, undecenes, dodecenes, tridecenes, tetradecenes, pentadecenes, hexadecenes, heptadecenes, octadecenes, nonadecenes, eicosenes, heneicosenes, docosenes, tricosenes, tetracosenes, pentacosenes, hexacosenes, heptacosenes, and octacosenes.

For lightly branched monoolefins the alkyl group branch or branches of the lightly branched monoolefin are generally selected from methyl, ethyl, and propyl groups, with shorter and normal branches being preferred. For all lightly branched monoolefins passed to the alkylation section, preferably the lightly branched monoolefin has only one alkyl group branch, but two alkyl group branches are also possible. Lightly branched monoolefins having either two alkyl group branches or four primary carbon atoms comprise generally less than 30 mol-%, and preferably less than 15 mol-%, of the total lightly branched monoolefins passed to the alkylation section, with the remainder of the lightly branched monoolefins passed to the alkylation section having one alkyl group branch. Monoolefins having either two alkyl group branches or four primary carbon atoms and a quaternary carbon atom comprise generally less than 10 mol-%, and preferably less than 1 mol-%, of the total lightly branched monoolefins passed to the alkylation section. Lightly branched monoolefins having either one alkyl group branch or three primary carbon atoms comprise preferably more than 85 mol-% of the total lightly branched monoolefins passed to the alkylation section. Lightly branched monoolefins having only one alkyl group branch and where the sole alkyl group branch is a methyl group are referred to herein as monomethyl-alkenes and are a preferred component of the dehydrogenated product stream.

Although vinylidene monoolefins may be present in the dehydrogenated product stream, they are normally a minor component and have a concentration of usually less than 0.5 mol-%, and more commonly less than 0.1 mol-%, of the olefins in the dehydrogenated product stream.

The skeletal structures of the monoolefins in a mixture comprising lightly branched monoolefins can be determined by analytical methods that are well- known to a person of ordinary skill in the art of gas chromatography and need not be described here in detail. A person of ordinary skill in the art can modify the apparatus and method in the previously mentioned article by Schulz et al. to equip the injector with a hydrogenator insert tube in order to hydrogenate the lightly branched monoolefins to lightly branched paraffins in the injector. The lightly branched paraffins are then separated and identified using essentially the apparatus and method described in the article by Schulz et al. This apparatus and method, however, will not determine the location of the carbon-carbon double bond in any of the monoolefins in the mixture.

In addition to the lightly branched monoolefin, other acyclic compounds may be charged to the alkylation section via the dehydrogenated product stream. One of the advantages of this invention is that the stream containing the lightly branched monoolefins can be passed directly to the alkylation reaction section despite the fact that that stream also contains acyclic paraffins having the same number of carbon atoms as the lightly branched monoolefins. Thus, this invention avoids the need to separate the paraffins from the monoolefins prior to passing to the alkylation section. Other acyclic compounds include nonbranched (linear) olefins and monoolefins. Nonbranched (linear) olefins which may be charged have a total number of carbon atoms per paraffin molecule of generally from 8 to 28, preferably from 8 to 15, and more preferably from 10 to 13 carbon atoms. The nonbranched olefin may be an alpha monoolefin but is preferably an internal monoolefin. When present in the dehydrogenated product stream with the lightly branched monoolefins, the linear olefin content should be no more than, 75 mol-% of the total monoolefins in the dehydrogenated product stream, but is generally less than 60 mol-% of the total monoolefins in the dehydrogenated product stream.

Because of the possible presence in the dehydrogenated product stream of linear monoolefins, in addition to the lightly branched monoolefins, the bulk dehydrogenated product stream may contain, on average, fewer than 3, or between 3 and 3.4, primary carbon atoms per monoolefin molecule in the dehydrogenated product stream. Depending on the relative proportions of linear and lightly branched monoolefins, the dehydrogenated product stream, or the sum of all the monoolefins that pass to the alkylation zone, may have from 2.25 to 3.4 primary carbon atoms per monoolefin molecule.

Linear and/or nonlinear paraffins which pass to the alkylation section, via the dehydrogenated product stream, have a total number of carbon atoms per paraffin molecule of generally from 8 to 28, preferably from 8 to 15, and more preferably from 10 to 13 carbon atoms. The nonlinear paraffins in the dehydrogenated product stream may include lightly branched paraffins and may also include paraffins having at least one quaternary carbon atom. Such linear and nonlinear paraffins are expected to act as a diluent in the alkylation step and not to materially interfere with the alkylation step. Monoolefin molecules consisting of at least one quaternary carbon atom generally comprise less than 10 mol-%, preferably less than 5 mol-%, more preferably less than 2 mol-%, and most preferably less than 1 mol-% of the dehydrogenated product stream or of the sum of all the monoolefins that pass to the alkylation zone.

The alkylation section reacts the monoolefins in the dehydrogenated product stream with a phenyl compound. In the general case, the monoolefins could be reacted with benzene or substituted derivatives of benzene including toluene and ethylbenzene. For detergent alkylation, the preferred phenyl compound is benzene. A solid alkylation catalyst typically reacts the phenyl compound and the monoolefins.

Only minimal skeletal isomerization of the olefins is believed to occur in the alkylation section. Minimal skeletal isomerization of the monoolefins means that generally less than 25 mol-%, and preferably less than 10 mol-%, of the olefin, the aliphatic alkyl chain, and any reaction intermediate undergoes skeletal isomerization. Thus, the extent of light branching of the lightly branched

monoolefin is identical to the extent of light branching in the aliphatic alkyl chain in the phenyl-alkane product molecule and the number of primary carbon atoms also remains essentially the same. Finally, although the formation of 1-phenyl- alkane product is not significant at alkylation conditions the number of primary carbon atoms in the phenyl-alkane product will be slightly less than the number of primary carbon atoms in the lightly branched monoolefin.

The alkylation of the phenyl compound with the lightly branched monoolefins produces (mj-alkylj) i-n-phenyl-alkanes, where the aliphatic alkyl group has two, three, or four primary carbon atoms per phenyl-alkane molecule.

Preferably, the aliphatic alkyl group has three primary carbon atoms per phenyl- alkane molecule, and more preferably one of the three primary carbon atoms is in a methyl group at one end of the aliphatic alkyl chain, the second primary carbon atom is in a methyl group at the other end of the chain, and the third primary carbon atom is in a single methyl group branch attached to the chain.

Generally from 0 mol-% to 75 mol-%, and preferably from 0 mol-% to 50 mol-%, of the (mj-alkylj) i-n-phenyl-alkanes produced may have 2 primary carbon atoms per phenyl-alkane molecule. Typically from 25 mol-% to 100 mol-%, of the (mi- alkylj) i-n-phenyl-alkanes produced may have 3 primary carbon atoms per phenyl- alkane molecule. Generally from 0 mol-% to 40 mol-% of the (mj-alkylj) i-n- phenyl-alkanes produced may have 4 primary carbon atoms. Thus, (m-methyl)- n-phenyl-alkanes having only one methyl group branch are preferred and are referred to herein as monomethyl-phenyl-alkanes. It is expected that the number of primary, secondary, and tertiary carbon atoms per product phenyl- alkane molecule can be determined by high resolution multipulse nuclear magnetic resonance (NMR) spectrum editing and distortionless enhancement by polarization transfer (DEPT), which is described in the brochure entitled"High Resolution Multipulse NMR Spectrum Editing and DEPT, "which is distributed by Bruker Instruments, Inc., Manning Park, Billerica, Massachusetts, USA, and which is incorporated herein by reference.

The alkylation of the phenyl compound with the monoolefins has a selectivity of 2-phenyl-alkanes of generally from 40 to 100 and preferably from

60 to 100, and an internal quaternary phenyl-alkane selectivity of generally less than 10 and preferably less than 5.

The alkylation of the phenyl compound with the monoolefins has a selectivity to phenyl-alkanes containing non-quat quaternary carbons of less than 10, and preferably less than 1. A suitable approximation of the selectivity to such quaternary phenyl-alkanes can be arrived at by using the following formula : where T = selectivity to non-quats quarternary carbons CQO = moles of monoolefins having a quaternary carbon atom entering the selective alkylation zone Cl-moles of monoolefins entering the selective alkylation zone The values of CQO and Co can be determined using the molar flow rate of monoolefins entering the selective alkylation zone and the previously mentioned modified apparatus and method of Schulz et al. The selectivity, T, can be estimated using this formula if each monoolefin entering the selective alkylation zone has an equal probability of alkylating the phenyl compound, regardless of whether the monoolefin has a quaternary carbon atom. As a first approximation, this condition is met when more than 40 wt-% of the monoolefins entering the selective alkylation zone are lightly branched monoolefins or normal monoolefins.

Alkylation of the phenyl compound by the monoolefins may be conducted as a batch method but a continuous method is preferred. The alkylation catalyst may be used as a packed bed or a fluidized bed in upflow, downflow, or horizontal flow mode. The benzene and the dehydrogenated product stream containing the lightly branched monoolefins that enters the alkylation zone typically has a total phenyl compound: monoolefin molar ratio of between 2.5 : 1 and 50: 1, and more typically between 8: 1 and 35: 1. Portions of the

dehydrogenation product stream may feed into several discrete points within the alkylation reaction zone, and at each zone the phenyl compound: monoolefin molar ratio exceed 50: 1. However, the total benzene: olefin ratio used in the foregoing variant of this invention will remain within the stated range. The total combined feed passes through the packed bed at a liquid hourly space velocity (LHSV) between 0.3 and 6 hr'. The alkylation conditions include a temperature in the range between 80°C (176°F) and 225°C (437°F). Preferably the alkylation occurs in at least partial liquid phase, and preferably in either an all-liquid phase or at supercritical conditions. The requisite pressure necessarily depends upon the olefin, the phenyl compound, and temperature, but normally ranges from 1379-6895 kPa (g) (200-1000 psi (g) ), and most usually 2069-3448 kPa (g) (300- 500 psi (g) ). The alkylation reaction usually leaves little unreacted olefin and typically goes to at least 98% conversion based on the monoolefin.

Any alkylation catalyst that meets the requirements for conversion, selectivity, and activity may be used. Preferred alkylation catalysts comprise zeolites having a zeolite structures BEA, MOR, MTW, or NES. Such zeolites include mordenite, ZSM-4, ZSM-12, ZSM-20, offretite, gmelinite, beta, NU-87, and gottardiite. These zeolite structure types, the term"zeolite structure type," and the term"isotypic framework structure"are used herein as they are defined and used in the Atlas of Zeolite Structure Types, by W. M. Meier, et al., published on behalf of the Structure Commission of the International Zeolite Association by Elsevier, Boston, Massachusetts, USA, Fourth Revised Edition, 1996. Alkylations using NU-87 and NU-85, which is an intergrowth of zeolites EU-1 and NU-87, are described in US-A-5,041, 402 and US-A-5,446, 234, respectively. Gottardiite, which has an isotypic framework structure of the NES zeolite structure type, is described in the articles by A. Alberti et al., in Eur. J.

Mineral. , 8,69-75 (1996), and by E. Galli et al., in Eur. J. Mineral., 8,687-693 (1996). Most preferably, the alkylation catalyst comprises mordenite.

Useful zeolites for the alkylation catalyst in the present invention generally have at least 10 percent of the cationic sites thereof occupied by ions other than alkali or alkaline-earth metals. Such other ions include aluminum, zinc, copper, aluminum, and preferably ammonium, hydrogen, rare earth, or combinations

thereof. In a preferred embodiment, the zeolites are converted to the predominantly hydrogen form, generally by replacement of the originally present ions with hydrogen ion precursors, e. g. , ammonium ions, which upon calcination yield the hydrogen form. This exchange is conveniently carried out by contact of the zeolite with an ammonium salt solution, e. g. , ammonium chloride, utilizing well known ion exchange techniques. In certain embodiments the replacement produces a zeolite material in which at least 50 percent of the cationic sites are occupied by hydrogen ions. Although the hydrogen form of the zeolite catalyzes the reaction successfully, the zeolite may also be partly in the alkali metal form.

The zeolites may be subjected to various chemical treatments, including alumina extraction (dealumination) and combination with one or more metal components, such as the metals of Groups IIIB (IUPAC 3), IVB (IUPAC 4), VIB (IUPAC 6), VIIB (IUPAC 7), VIII (IUPAC 8-10), and IIB (IUPAC 12). The zeolites may, in some instances, be subjected to thermal treatment. including steaming or calcination in air, hydrogen, or an inert gas, e. g. nitrogen or helium. A suitable steaming treatment comprises contacting the zeolite with an atmosphere containing from 5 to 100% steam at a temperature of from 250°C (482°F) to 1000°C (1832°F) for 0.25 and 100 hours at pressures ranging from sub-atmospheric to several hundred atmospheres.

A matrix material or binder that is resistant to the temperature and other conditions used in the process may contain the zeolite. Suitable matrix materials include synthetic substances, naturally occurring substances, and inorganic materials such as clay, silica, and/or metal oxides. Naturally occurring clays for compositing with the zeolite include those of the montmorillonite and kaolin families. In addition to the foregoing materials, the zeolite used in this invention may be compounded with a porous matrix material, such as alumina, silica-alumina, silica-magnesia, silica-zirconia, silica-thoria, silica-beryllia, silica- titania, and aluminum phosphate as well as ternary combinations, such as silica- alumina-thoria, silica-alumina-zirconia, silica-alumina-magnesia, and silica- magnesia-zirconia. The relative proportions of and matrix material may vary widely, with the zeolite weight content ranging from between 1 % to 99%,

typically from 5% to 80% by weight, and preferably from 30% to 80%, of the combined weight of zeolite and matrix material.

The zeolites in the alkylation catalyst generally have a framework silica : alumina molar ratio of from 5: 1 to 100: 1. A mordenite for use in alkylation has a framework silica : alumina molar ratio generally of from 12: 1 to 90: 1, and preferably of from 12: 1 to 25: 1. The term"framework silica : alumina molar ratio" means the molar ratio of Si02 per A1203, in the zeolite framework.

Zeolites prepared in the presence of organic cations may not be sufficiently catalytically active for alkylation. Insufficient catalytic activity is believed to result from the organic cations of the forming solution occupying the intracrystalline free space. Heating in an inert atmosphere at 540°C (1004°F) for one hour, ion exchanging with ammonium salts, and calcining at 540°C (1004°F) in air may activate such catalysts. Calcination temperatures higher than 540°C (1004°F) may ensure decomposition of any ammonia on the catalyst.

The alkylation reaction zone produces an alkylation reaction effluent that enters separation facilities for the recovery of products and recyclable feed compounds. The alkylation reaction effluent passes into a benzene column to produce an overhead stream containing benzene for recycled to the alkylation reaction zone and a bottoms stream containing the phenyl-alkane product. This bottoms stream passes into a paraffin column to produce an overhead stream containing unreacted paraffins and a bottoms stream containing the product phenyl-alkanes and any higher molecular weight by-product hydrocarbons formed in the alkylation reaction zone. The paraffin column bottoms stream may pass to a rerun column to produce an overhead phenyl-alkane product stream containing the MAB and a rerun column bottoms stream containing polymerized olefins and polyalkylated benzenes (heavy alkylate). Alternatively, a sufficiently low heavy alkylate content of the paraffin column bottoms stream renders the rerun column unnecessary and the paraffin column bottoms stream may be recovered as the net MAB stream, which may be subsequently sulfonate to produce MABS.

Several variants of the subject process are possible. Selective hydrogenation may saturate diolefins in the dehydrogenated product stream to desired monoolefins. The process may selectively remove deleterious aromatic by-products contained in the dehydrogenated product stream which may be formed during the catalytic dehydrogenation of paraffins and that may cause deactivation of the catalyst in the alkylation section, decreased selectivity to the desired phenyl-alkanes, and accumulation to unacceptable concentrations. A selective removal zone may take the aromatic by-products from the extract stream, the dehydrogenated product steam, the overhead liquid stream of the paraffin column, the dehydrogenation zone or, where present, the selective diolefin hydrogenation product stream.

This process can produce a preferred MAB composition by adsorptively separating paraffins having an average weight between the weight of a C, 0 paraffin and a C, 3 paraffin to produce extract paraffins having an average level of branching of from 0.25 to 1.3, or of from 0.4 to 1.3, alkyl group branches per paraffin molecule. These extract paraffins primarily comprise linear paraffins and mono-branched paraffins, and the alkyl group branches on the aliphatic alkyl chain of the extract paraffins primarily comprise small substituents, such as methyl group branches, ethyl group branches, or propyl group branches. The extract paraffins are dehydrogenated to produce the corresponding mono- olefins, which alkylate a phenyl compound to produce phenyl-alkanes. The resultant phenyl-alkanes have the characteristics that the phenyl-alkanes having the phenyl group attached to the 2-and/or 3-position of the aliphatic alkyl group comprise greater than 55 wt-% of the phenyl-alkanes, and the phenyl-alkanes having at least one quaternary carbon atom on the aliphatic alkyl group comprise less than 20% of the phenyl-alkanes.

Sulfonation of the phenyl-alkanes produced by the processes of this invention can be accomplished by contacting the phenyl-alkane compounds with any of the well-know sulfonation systems, including those described in Detergent Manufacture Including Zeolite Builders and Other New Materials, by Marshal Sittig, Noyes Data Corporation, Park Ridge, New Jersey, 1979, and in Volume 56 of"Surfactant Science"series, Marcel Dekker, Inc., New York, NY,

1996. Sulfonation of the phenyl-alkane compounds produces a sulfonate product comprising phenyl-alkane sulfonic acids. After sulfonation, neutralization of the sulfonate product with any suitable alkali, such as sodium, potassium, ammonium, magnesium, calcium, and substituted ammonium alkalis, and mixtures thereof produces a neutralized product comprising phenyl-alkane sulfonates.

In yet another aspect of the present invention, this invention is the use of the MAB compositions produced by the processes disclosed herein as lubricants. These phenyl-alkanes are believed to have properties of viscosity, temperature-dependence of viscosity, and density that make them advantageous for use as petroleum lubricants. The use of phenyl-alkanes as lubricants is described, for example, in the article by E. R. Booser in Kirk-Othmer Encyclopedia of Chemical Technoloay, Fourth Edition, Volume 15, John Wiley and Sons, New York, New York, USA, 1995, pp. 463-517, to which reference is made for a description of such lubricants and their use.

In still another aspect, this invention is the use of the MABS compositions produced by the processes disclosed herein as lubricant additives. It is believed that phenyl-alkane sulfonates, either in the form of normal salts or basic salts of phenyl-alkane sulfonic acids, produced as disclosed herein, have the ability to reduce or prevent deposits in engines operating at high temperatures. The term "normal salt"of an acid means a salt which contains the stoichiometric amount of metal required for the neutralization of the acidic group or groups present, and the term"basic salt"means a salt which contains more metal than is required for the neutralization reaction. The excess metal in the form of basic salts is believed to be capable of neutralizing oil oxidation combustion products and "blow-by"fuel combustion products. Phenyl-alkane sulfonates and their use as lubricant additives, in particular as detergents, is described, for example, in the above-mentioned Booser article; in Lubricant Additives, by C. V. Smalheer and R. K. Smith, The Lezius-Hiles Co. , Cleveland, Ohio, USA, 1967, pp. 2-3; and in the article by R. W. Watson and T. F. McDonnell, Jr., entitled"Additives-The Right Stuff for Automotive Engine Oils,"in Fuels and Lubricants Technology : An

Overview SP-603, Society of Automotive Engineers, Warrendale, Pennsylvania, USA, October 1984, pp. 17-28.

The drawing shows a preferred arrangement for an integrated separation- dehydrogenation-alkylation scheme of this invention. The following description of the drawing is not meant to preclude other arrangements for the process flow of this invention and is not intended to limit this invention as set forth in the claims.

Referring now to the drawing, a line 12 charges a feed mixture comprising an admixture of C1O-C13, including lightly branched paraffins, more highly branched paraffins, and normal (nonbranched) paraffins to an adsorptive separation zone 20 which employs normal paraffin and/or cycloparaffin and/or isooctane as the desorbent. A line 20 passes a raffinate stream comprising more highly branched paraffins and a cycloparaffin or isooctane from the adsorptive separation zone 20 to a raffinate column 30. The conditions of the raffinate column 30 produce an overhead stream 32 comprising desorbent material and a bottom stream comprising more highly branched paraffins in a line 28. Line 34 passes an extract stream comprising lightly branched paraffins, normal paraffins, and desorbent materials from adsorptive separation zone 20 to an extract column 40. The extract column 40 recovers adsorbent materials in line 46. The overhead streams in lines 32 and 46 combined recycle desorbent via a line 44. The extract column 40 also produces a bottom stream comprising lightly branched paraffins and normal paraffins in a line 52.

Line 52 admixes the bottom stream from the extract column 40 with recycled hydrogen from a line 82 to form a mixture of paraffins and hydrogen that flows through a line 56. An indirect heat exchanger 58 heats the contents of line 56 that pass via a line 62 to a fired heater 60. A line 64 carries the heated stream to a dehydrogenation reactor 70. Dehydrogenation reactor 70 contacts the paraffins with a dehydrogenation catalyst at conditions that effect significant conversion of the paraffins to the corresponding olefins. Lines 66 and 68 cool a dehydrogenation reactor effluent stream comprising a mixture of hydrogen, paraffins, monoolefins including lightly branched monoolefins, diolefins, Cs-

minus hydrocarbons, and aromatic hydrocarbons by passage serially through heat exchangers 58 and 72. This cooling condenses substantially all of the C4- plus hydrocarbons into a liquid phase stream and separates the liquid phase stream from the remaining hydrogen-rich vapor. Line 74 passes the dehydrogenation reactor effluent to vapor-liquid separation vessel 80 that divides it into a hydrogen-rich vapor phase stream removed through a line 76 and a dehydrogenation product stream removed through a line 84. Line 76 supplies recycle hydrogen to line 82 and line 78 recovers a net hydrogen product stream.

Line 84 passes the bottom of the separation vessel 80, containing normal paraffins, lightly branched paraffins, normal monoolefins, lightly branched monoolefins, Cg-minus hydrocarbons, diolefins, aromatic by-products, and some dissolved hydrogen, to a selective hydrogenation reactor 86. Selective hydrogenation reactor 86 contacts the dehydrogenated product stream with a selective hydrogenation catalyst at conditions that convert a significant amount of the diolefins to the corresponding monoolefins and uses dissolved hydrogen in the dehydrogenated product stream and/or additional make-up hydrogen (not shown) for the conversion. Line 88 carries a selective hydrogenation reactor effluent stream comprising a mixture of hydrogen, normal paraffins, lightly paraffins, normal monoolefins, lightly branched monoolefins, Cg-minus hydrocarbons, and aromatic by-product hydrocarbons to a stripping column 90.

Stripping column 90 recovers a net overhead stream in line 94 that comprises the Cg-minus hydrocarbons produced in the dehydrogenation reactor as by- products and any remaining dissolved hydrogen. are separated from the and concentrated into a net overhead stream removed from the process through a line 94.

A line 96 passes the remaining C, 0-plus hydrocarbons from stripping column 90 to an aromatics removal zone 100 that contacts this effluent with an adsorbent to remove the aromatic by-products. Line 98 transfers from the aromatics removal zone 100 an effluent comprising an admixture of the normal paraffins, lightly branched paraffins, normal monoolefins, and lightly branched monoolefins, and which has a greatly reduced concentration of aromatic by-

products compared to the stripping effluent stream. After addition of benzene via a line 112 a line 102 passes the admixture into an alkylation reactor 104 where the admixture contacts an alkylation catalyst at alkylation-promoting conditions to produce phenyl-alkanes.

A line 106 carries the alkylation reactor effluent stream into a benzene fractionation column 110 by a line 106. This stream comprises an admixture of benzene, normal paraffins, lightly branched paraffins, and MAB. The phenyl- alkanes characterize the MAB which in addition to the phenyl portion have one aliphatic alkyl portion with either 1 or 2 primary carbon atoms, or with 2,3, or 4 primary carbon atoms and no quaternary carbon atoms except for quats.

Column 110 separates the effluent into a bottom stream and an overhead stream comprising benzene and possibly light gases. A line 107 combines the overhead stream of line 107 with make-up benzene from a line 109 into a fine 108 that passes the combined stream to a separator drum 120. A line 114 removes any noncondensed light gases from drum 120 while a line 116 supplies condensed liquid as reflux to column 110 via a line 118 and supplies benzene for recycle by a line 112. A line 122 carries the remainder of the alkylation effluent stream from column 110 to a paraffin column 124 from which a line 48 withdraws an overhead stream containing a mixture of paraffins and generally less then 0.3 wt-% monoolefins. A line 126 takes the paraffin column bottom stream containing the phenyl-alkanes and heavy alkylate by-products to a rerun column 130 that separates paraffin column bottom stream into a bottom stream 132 comprising heavy alkylate and an overhead alkylate product stream 128 containing the phenyl-alkane compounds. Sulfonation of the phenyl-alkane compounds in the overhead alkylate product stream 128 will produce phenyl- alkane sulfonic acids, which can be neutralized.

EXAMPLE 1 A"pulse test"procedure tests adsorbents with a particular feed mixture and desorbent material to measure adsorptive capacity, selectivity, resolution and exchange rate of the adsorbent. The basic pulse test apparatus consists of a tubular adsorbent chamber of approximately 70 cc volume having an inlet and

outlet at opposite ends of the chamber. The chamber is contained within a temperature control means and pressure control equipment maintains the chamber at a constant predetermined pressure. Attachment of quantitative and qualitative analytical equipment such as refractometers, polarimeters and chromatographs to an outlet line of the chamber and will detect quantitatively and/or determine qualitatively one or more components in the effluent stream leaving the adsorbent chamber. A pulse test begins by passing the desorbent material through the adsorbent chamber to fill the adsorbent to equilibrium with a particular desorbent material, then injecting a pulse of the feed mixture, sometimes diluted in desorbent, for a duration of one or more minutes and then resuming desorbent flow elute the feed mixture components as in a liquid-solid chromatographic operation. The effluent can be analyzed on-stream and/or effluent fractions can be collected and later analyzed separately to permit plotting the traces of the envelopes of corresponding component peaks in terms of component concentration versus quantity of effluent.

From information derived from the pulse test the adsorbent/desorbent system performance can normally be rated in terms of retention volume for an extract or a raffinate component, selectivity for one component with respect to the other, stage time, the resolution between the components and the rate of desorption of an extract component by the desorbent. The distance between the center of the peak envelope of an extract or a raffinate component and the peak envelope of a tracer component or some other known reference point will determine retention volume of an extract or a raffinate in terms of the volume in cubic centimeters of desorbent pumped during the time interval corresponding to the distance between the peak envelopes.

Table 1 lists variables and results of small scale"pulse tests"performed to evaluate various desorbents and conditions on several feed mixtures. The materials labeled Raffinate A and B are the raffinate streams of commercial adsorptive separation units which recover normal paraffins from a Ciao-14 hydrocarbon fraction. The desorbent column in Table 1 indicates the volume percent of each component of the desorbent as specified by the footnotes.

Table 1 Run No. Feed Mixture Temp, °C Desorbent * 9577-77 Raffinate A 150 50 C5/50 N6 9577-85 Raffinate A 100 50 C5/50 N6 9937-01 Raffinate A 150 100 C6 9937-06 Raffinate A 150 70 C6/30 N6 9937-17 Raffinate A 150 50 C6/50 N6 9937-25 Raffinate B 150 50 C6/50 N6 9953-06 Kerosene 150 50 C6/50 N6

*C5 indicates cyclopentane C6 indicates cyclohexane N6 indicates n-hexane All tests used the adsorbent comprised 80% silicalite and 20% silica binder and the chromatographic column having an adsorbent volume of 70 ml. The flow rate through the column was 1.21 cc/min.

In view of the very large number of different compounds in the feed mixture pulse and the limitations inherent of the simple pulse test procedure the effectiveness of the separation was determined by collecting fractions of the effluent every two minutes and analyzing each fraction. The initial fractions had high concentrations of desorbent and were followed by fractions having high concentrations of the more highly branched nonnormal hydrocarbons. The desired acyclic hydrocarbons having only 3 primary carbon atoms (i. e., monomethyl hydrocarbons) tended to concentrate in the fractions collected at the end of the pulse. Table 2 gives the concentration (wt percent) of acyclic paraffins having only 3 primary carbon atoms (i. e., monomethyl paraffins) present in several different fractions of Run No. 9937-06.

Table 2 Acyclic Paraffins Having Only 3 Fraction No. Primary Carbon Atoms, % 18 34 19 48 20 60 22 55 24 60 26 77 28 77 32 79 38 96 Run No. 9937-06 combined liquid collected as fractions No. 19 to 100.

Analysis found the combined liquid to contain the weight percentages of different structural classes of compounds on a desorbent free basis as shown in Table 3.

Table 3 acyclic paraffins having only 3 primary carbon 64% atoms (monomethyl branched) acyclic paraffins having only 4 primary carbon 2.7% atoms (dimethyl branched) acyclic paraffins having only 2 primary carbon 4.6% atoms (normal paraffins) Aromatics 4. 1 % Naphthenes 13. 9% Unknowns 10. 7%

Run No. 9937-17 combined liquid collected as fractions No. 23 to 50.

Analysis found the liquid to contain the weight percentages of different structural classes of compounds shown in Table 4: Table 4 acyclic paraffins having only 3 primary carbon 77% atoms (monomethyl branched) acyclic paraffins having only 4 primary carbon 0. 1 % atoms (dimethyl branched) acyclic paraffins having only 2 primary carbon atoms (normal) Aromatic Naphthene Unknowns 4. 8%

Analysis of a sample formed by combining liquid from fractions 23 to 48 of Run 9953-6 found 67% acyclic paraffins having only 3 primary carbon atoms (i. e., monomethyl branched compounds) and 9.3% acyclic paraffins having only 2 primary carbon atoms (i. e., normal paraffins).

In comparing this data to the performance normally desired in commercial separations requires recognition that better selectivity will result from optimization in terms of adsorbent composition, desorbent composition and operating conditions. Further the use of simulated moving bed (SMB) technology or even better batch separation technology will improve the performance of the process.

EXAMPLE 2 This example subjects a representative mixture of C, o pure components to a pulse test procedure that uses of a pre-pulse of C8 isoparaffin. The test used a feed mixture containing equal volumes of 3,3, 5-trimethylheptane, 2,6- dimethyloctane, 2-methylnonane, normal decane, and 1,3, 5-trimethylbenzene in pulse test column having a volume of 70 cc and temperature maintained at

120°C (248°F). The flow rate through the column was 1.1 cc/min. The test used a silicalite adsorbent and a 70/30 volume % mixture of normal heptane and isooctane as desorbent. The test injected a pre-pulse of 40 ml isooctane into the test loop immediately before the feed mixture injection.

Figure 2 graphically represents the results in a plot of the relative concentrations of the components versus time, as measured by the volume of collected effluent. Figure 2 shows a useful separation between the monomethyl paraffin and the normal paraffin on the one hand and the di-and tri-methyl paraffins on the other hand. Although the use of the pre-pulse is believed to improve the separation of the monomethyl paraffin band in the effluent, a useful separate band of the monomethyl paraffin is produced as a result of the presence of isooctane in the desorbent, even in the absence of a pre-pulse.

EXAMPLE 3 An olefinic stream was used.

Table 5: Composition of Olefinic Stream Olefin Component Content (wt-%) Lights'0. 64 Linear olefins 30.11 6-methyl undecene 7. 66 5-methyl undecene 15. 33 4-methyl undecene 11.82 3-methyl undecene 12.95 2-methyl undecene 8. 87 Other alkyl olefins 9.05 Heavies 3. 53 Total 100

1 Lights include olefins having fewer than 12 carbon atoms.

2 Linear olefins include C, 2 linear olefins.

3 Other alkyl olefins include dimethyl, trimethyl, and other C12 olefins 4 Heavies include C, 2 olefin dimers and trimers.

Mixture of an olefinic stream comprising a blend of monomethyl C12 olefins and having the composition shown in Table 5 with benzene produced a combined stream consisting of 93.3 wt-% benzene and 6.7 wt-% olefinic stream,

which corresponds to a molar ratio of benzene per olefin of 30: 1. A cylindrical reactor with an inside diameter of 0.875 in (22.2 mm), was loaded with 75 cc (53.0 g) of a mordenite-alumina extruded catalyst prepared from the hydrogen form of a mordenite having a SiO2/AI203 of 18.

The combined stream passed to the reactor and contacted the extrudate at a LHSV of 2.0 hr', a total pressure of 500 psi (g) (3447 kPa (g) ), and a reactor inlet temperature of 125°C (257°F). At these conditions, the reactor lined out over a period of 24 hours and then a selective liquid product was collected over the period of the next 6 hours.

The selective liquid product was analyzed by'3C nuclear magnetic resonance (NMR) in order to determine the selectivity to 2-phenyl-alkanes and end quaternary phenyl-alkanes. The NMR analytical method typically consists of diluting a 0.5 g sample of phenyl-alkane mixture to 1.5 g with anhydrous deuterated chloroform, mixing a 0.3 milliliter aliquot of the diluted phenyl-alkane mixture with 0.3 milliliter of 0.1 M chromium (III) acetylacetonate in deuterated chloroform in a 5 mm NMR tube, adding A small amount of tetramethylsilane (TMS) to the mixture as a 0.0 ppm chemical shift reference and running the carbon spectrum on a Bruker ACP-300 FT-NMR spectrometer, available from Bruker Instruments, Inc., Billerica, Massachusetts, USA at a field strength of 7.05 Tesla or 75.469 MHz in a 5 mm QNP probe with a sweep width of 22727 Hz (301.1 ppm) to collect 65000 data points. The quantitative carbon spectrum is obtained using gated on-acquisition'H decoupling (inverse gated decoupling).

The quantitative 13C spectrum is run with 7.99 microsecond (90°) pulses, 1.442 second acquisition time, a 5 second delay between pulses, a decoupler power, using composite pulse decoupling (CPD), of 18H with a pulse width of 105 microseconds (90°) and at least 2880 scans. The number of scans depends on whether benzene is stripped from the liquid product prior to taking the above- mentioned 0.5 g sample. The data processing is done with The Bruker PC software WINNMR-1 D, Version 6.0, does data scanning. During data processing a line broadening of 1 Hz is applied to the data. Specific peaks are integrated in the region between 152 ppm and 142 ppm. Table 6 shows then NMR peak identifications of the chemical shifts of the benzylic carbon of the phenyl-alkane isomers. The term"benzylic carbon"means the carbon in the ring of the phenyl group bound to the aliphatic alkyl group.

Table 6 : 13C NMR Peak Identifications

Chemical Shift of the Benzylic Carbon (ppm) 149.6 2-methyl-2-phenyl End 4-methyl-2-phenyl NQ 148. 3 I m-methyl-m-phenyl, m>3 Internal 148.0 5-methyl-2-phenyl NQ m-methyl-2-phenyl, m>5 NQ 5-methyl-2-phenyl NQ 147. 8 I 2-phenyl (linear) NQ 3-methyl-3-phenyl I nternal 147.6 4-methyl-2-phenyl NQ 147.2 3-methyl-2-phenyl NQ 146.6 3-methyl-2-phenyl NQ 146.2-146. 3 m-methyl-4-phenyl, mx4 NQ 145.9-146. 2 m-methyl-3-phenyl, m>5 NQ 145.9 3-phenyl (linear) NQ 'NQ= Nonquat

The peak at 148.3 ppm is identified both with 4-methyl-2-phenyl-alkanes and with m-methyl-m-phenyl-alkanes (m>3). However, the presence of m-methyl-m- phenyl-alkanes (m>3) at more than 1 %, reveals them as a distinct peak at 0.03 ppm upfield of the peak for the 4-methyl-2-phenyl-alkanes. The peak at 147.8 ppm is identified with the 2-phenyl-alkanes as shown in Table 6, with possible interference from 3-methyl-3-phenyl-alkanes.

The end quaternary phenyl-alkane selectivity is computed by dividing the integral of the peak at 149.6 ppm by the sum of the integrals of all of the peaks listed in Table 6, and multiplying by 100. The 2-phenyl-alkane selectivity can be

estimated if the amount of internal quaternary phenyl-alkanes contributing to the peaks at 148.3 ppm and 147.8 ppm is less than 2%, as determined by the hereinafter-described gas chromatography/mass spectrometry method. As a first approximation, this condition is met when the sum of the integrals of the 4-phenyl-alkane and 3-phenyl-alkane peaks at 146.2-146. 3 ppm and 145.9- 146.2 ppm (respectively) is small relative to the sum of the integrals of all the peaks from 145.9 ppm to 149.6 ppm and the end quaternary phenyl-alkane selectivity is less than 10. In such case, the 2-phenyl-alkane selectivity is computed by dividing the sum of integrals of the peaks from 149.6 to 146.6 ppm by the sum of the integrals of all of the peaks listed in Table 6, and multiplying by 100.

Analysis by gas chromatography/mass spectrometry of the selective liquid product also determines the selectivity to internal quaternary phenyl-alkanes.

The gas chromatography/mass spectrometry analytical method typically consists of analyzing the liquid product by an HP 5890 Series II gas chromatograph (GC) equipped with an HP 7673 autosampler and an HP 5972 mass spectrometer (MS) detector used with an HP Chemstation to control the data acquisition and analysis, both available from Hewlett Packard Company, Palo Alto, California, USA. The GC has a 30 meter x 0.25 mm DB1HT (df = 0.1 pm) column or equivalent obtained from J&W Scientific Incorporated, 91 Blue Ravine Road, Folsom, California, USA. Helium carrier gas at 15 psi (g) (103 kPa (g) ) and 70°C (158°F) passes in constant pressure mode as the injector temperature holds at 275°C (527°F). The transfer line and MS source temperatures are held at 250°C (482°F). An oven temperature program of 70°C (158°F) for 1 minute, then to 180°C (356°F) at 1°C per minute (1.8 °F per minute), then to 275°C (527°F) at 10°C per minute (18°F per minute), then hold at 275°C (527°F) for 5 minutes is used. The HP Chemstation software tunes the MS with the software set to standard spectra autotune. The MS detector is scanned from 50-550 Da with a threshold = 50.

The concentrations of internal quaternary phenyl-alkanes in the selective liquid product are determined (i. e. , the selective liquid product is quantitated)

using the method of standard addition. Chapter 7 of the book entitled, Samples and Standards, by B. W. Woodget et al., published on behalf of ACOL, London by John Wiley and Sons, New York, in 1987 provides background on this method.

First, a stock solution of internal quaternary phenyl-alkanes is prepared and quantitated by alkylating benzene with a monomethyl alkene using a nonselective catalyst such as aluminum chloride. The stock solution comprising this nonselective liquid product of this alkylation contains a blend of internal quaternary phenyl-alkanes. A standard GC methodology identifies the largest peaks corresponding to the stock solution and the concentrations of the internal quaternary phenyl-alkanes in the stock solution are determined (i. e. , the stock solution is quantitated) using a flame ionization detector (FID). The retention times of the peaks for the internal quaternary phenyl-alkanes decrease as the index m in the formula m-methyl-m-phenyl-alkane increases and as the number of carbon atoms in the aliphatic alkyl group of the internal quaternary phenyl- alkane decreases. The concentration of each internal quaternary phenyl-alkane is computed by dividing the area of the peak of that internal quaternary phenyl- alkane by the sum of the areas of all of the peaks.

Next, a spiking solution of internal quaternary phenyl-alkanes is prepared by diluting an aliquot portion of the stock solution with dichloromethane (methylene chloride) to attain a nominal concentration of 100 wppm of one particular internal quaternary phenyl-alkane of interest (e. g., 3-methyl-3-phenyl decane). The concentration of any other particular internal quaternary phenyl- alkane in the spiking solution may be greater or less than 100 wppm, depending on the concentration of that internal quaternary phenyl-alkane in the stock solution.

Third, a sample solution is prepared by adding 0.05 g of an aliquot portion of the selective liquid product to a 10 milliliter volumetric flask and then diluting its contents with dichloromethane up to the 10 milliliter mark.

Fourth, a resultant solution is prepared by adding 0.05 g of an aliquot portion of the selective liquid product to a 10 milliliter volumetric flask and

diluting its contents with spiking solution up to the 10 milliliter mark to dilute the contents.

Both the sample solution and the resultant solution are analyzed by GC/MS under the above-described conditions. Table 7 lists the extracted from the full MS scan, plotted, and integrated using the HP Chemstation software.

The HP Chemstation software determines the individual extracted ion peak areas that correspond to the internal quats listed in Table 7.

TABLE 7: Ratio of Mass to Charge of ion for Peaks of Extracted ions Internal Number of Carbon Ratio of Mass to Charge Quaternary Atoms in Aliphatic (m/z) of Two Extracted Phenyl-Alkane Alkyl Group of the lons Corresponding to Internal Quaternary Internal Quaternary Phenyl-Alkane Phenyl-Alkane 11 133 and 203 3-methyl-3-phenyl 12 133 and 217 13 133 and 231 11 147 and 189 4-methyl-4-phenyl 12 147 and 203 13 147 and 217 11 161 and 175 5-methyl-5-phenyl 12 161 and 189 13 161 and 203 The concentration of each internal quaternary phenyl-alkane in Table 7 is computed using the following formula :

where C = concentration of internal quaternary phenyl-alkane in sample solution, wt-% S = concentration of internal quaternary phenyl-alkane in spiking solution, wt-% A, = peak area of internal quaternary phenyl-alkane in sample solution, area units A2 = peak area of internal quaternary phenyl-alkane in resultant solution, area units Using consistent units the concentration of each internal quaternary phenyl- alkane in the selective liquid product is computed from the concentration of that internal quaternary phenyl-alkane in the sample solution by accounting for the dilution effect of the dichloromethane in the sample solution. In this manner, the concentration in the selective liquid product of each of the internal quaternary phenyl-alkanes in Table 7 is computed. The total concentration of internal quaternary phenyl-alkanes in the selective liquid product, CQpA, is computed by summing the individual concentrations of each of the internal quaternary phenyl- alkanes in Table 7.

The selective liquid product may contain internal quaternary phenyl- alkanes other than those listed in Table 7, such as m-methyl-m-phenyl-alkanes where m > 5, depending on the number of carbon atoms in the aliphatic alkyl groups of the phenyl-alkanes. It is believed that, with the C12 olefinic stream and the conditions of this example, the concentrations of such other internal quaternary phenyl-alkanes are relatively low compared to those of the internal quaternary phenyl-alkanes listed in Table 7. For this example, the total concentration of internal quaternary phenyl-alkanes in the selective liquid

product, CIQPA, is computed by summing only the individual concentrations of each of the internal quaternary phenyl-alkanes in Table 7. However, if the olefinic stream had comprised olefins having, say, up to 28 carbon atoms, then the total concentration of internal quaternary phenyl-alkanes in the selective liquid product, CIQPA, would be computed by summing individual concentrations of m-methyl-m-phenyl-alkanes, where m is from 3 to 13. In more general terms, if the olefinic stream contains olefins having x carbon atoms, then the total concentration of internal quaternary phenyl-alkanes in the selective liquid product, CfpQA, is computed by summing individual concentrations of m-methyl- m-phenyl-alkanes where m is from 3 to x/2. Without undue experimentation at least one peak with a ratio of mass to charge (m/z) of an extracted ion corresponding to each internal quaternary phenyl-alkane is identifiable, so that the concentration of all internal quaternary phenyl-alkanes may be determined and then summed to arrive at CIQpA.

The selectivity to internal quaternary phenyl-alkanes in the selective liquid product is computed using the following formula: where Q = selectivity to internal quaternary phenyl-alkanes CIQPA = concentration of internal quaternary phenyl-alkanes in selective liquid product, wt-% CMAB = concentration of modified alkylbenzenes in selective liquid product, wt-% The concentration of modified alkylbenzenes, CMAB, in the selective liquid product is determined using a gas chromatography method to find the concentration of impurities in the selective liquid product. In the determinination of CMAB, "impurities"means components of the selective liquid product that lie outside a specific retention time range used in the gas chromatography method.

"Impurities"generally includes benzene, some dial kylbenzenes, olefins, paraffins, etc.

A gas chromatography method determines the amount of impurities from the selective liquid product. Equivalent equipment, equivalent sample preparation, and equivalent GC parameters that are different but that produce equivalent results to those described below may also be used to determine the amount of impurities in the selective liquid product.

Equipment : Hewlett Packard Gas Chromatograph HP 5890 Series II equipped with a split/splitless injector and flame-ionization detector (FID) * J&W Scientific capillary column DB-1 HT, 30 meter length, 0.25 mm inside diameter, 0.1 micro-meter film thickness, catalog no. 1221131.

Restek Red lite Septa 11 mm, catalog no. 22306. (Available from Restek Corporation, 110 Benner Circle, Bellefonte, Pennsylvania, USA).

Restek 4 mm Gooseneck inlet sleeve with a carbofrit, catalog no.

20799-209.5.

O-ring for inlet liner Hewlett Packard, catalog no. 5180-4182.

J. T. Baker HPLC grade methylene chloride, catalog no. 9315-33, or equivalent. (Available from J. T. Baker Co. , 222 Red School Lane, Phillipsburg, New Jersey, USA).

2 ml gas chromatograph autosampler vials with crimp tops, or equivalent.

Sample Preparation: Weigh 4-5 mg of sample into a 2 ml GC autosampler vial.

Add 1 mi methylene chloride to the GC vial ; seal with 11 mm crimp vial Teflon lined closures (caps), HP part no. 5181-1210 (available from Hewlett Packard Company), using crimper tool, HP part no.

8710-0979 (available from Hewlett Packard Company); and mix well.

The sample is now ready for injection into the GC.

GC Parameters: Carrier gas: hydrogen.

Column head pressure: 9 psi.

Flows : column flow, 1 ml/min ; split vent, 3 ml/min ; septum purge, 1 ml/min.

Injection : HP 7673 Autosampler, 10 microliter syringe, 1 microliter injection.

Injector temperature: 350°C (662°F) Detector temperature: 400°C (752°F) Oven temperature program: initial hold at 70°C (158°F) for 1 minute; heating rate of 1 °C per minute (1. 8°F per minute); final hold at 180°C (356°F) for 10 minutes.

This gas chromatography method requires two standards freshly distilled to a purity of more than 98 mol-% and generally comprising a 2-phenyl-alkane.

The"light standard"of the 2-phenyl-alkane has at least one fewer carbon atom in its aliphatic alkyl group than that of the olefin in the olefinic stream charged to the alkylation zone that has the fewest number of carbon atoms. The"heavy standard"of the 2-phenyl-alkane standard has at least one more carbon atom in its aliphatic alkyl group than that of the olefin in the olefinic stream charged to the alkylation zone that has the most number of carbon atoms. For example, if the olefins in the olefinic stream charged to the alkylation zone have from 10 to 14 carbon atoms, then the suitable standards include 2-phenyl-octane as the light standard and 2-phenyl-pentadecane as the heavy standard.

Gas chromatography methods using the conditions specified above determine retention time of each standard, and the two standard retention times in turn define a retention time range. Then, an aliquot sample of the selective liquid product is analyzed by the gas chromatography method using the above conditions. If more than 90% of the total GC area is within the retention time range, then the impurities in the selective liquid product are deemed to be not more than 10 wt-% of the selective liquid product, and, for the sole purpose of

computing the selectivity to internal quaternary phenyl-alkanes, CMAB is assumed to be 100 wt-%.

If the percent of the total GC area within the retention time range is not more than 90%, then the impurities in the selective liquid product are deemed to be more than 10 wt-% of the selective liquid product. In this case, to determine CMAB, impurities are distilled from the selective liquid product by adding 2200 to 2300 g of the selective liquid product to 5-liter, 3-necked round bottom flask with 24/40 joints and containing a magnetic stir bar and a few boiling chips. A 9-1/2 inch (24.1 cm) long Vigreux condenser with a 24/40 joint is placed in the center neck of the flask and a water cooled condenser fitted with a calibrated thermometer attaches to the top of the Vigreux. A vacuum receiving flask attaches to the end of the condenser. A stopper occupies one side arm of the flask and a calibrated thermometer extends from the other side arm. Aluminum foil wrapping surrounds the flask and the Vigreux. A vacuum line applies vacuum to the receiving flask as the selective liquid product is stirred in the flask and upon reaching maximum vacuum (at least 25.4 mm Hg by gauge or less), an electric heating mantle heats the selective liquid product.

After the heating begins"fraction A"is first collected from 25°C (77°F) to the light standard boiling point temperature at the pressure at the top of the Vigreux, as measured by the calibrated thermometer at the top of the Vigreux.

"Fraction B"is then collected from the temperature of the light standard boiling point to the heavy standard boiling point temperature both measured by the calibrated thermometer at the top of the Vigreux for the pressure at the top of the Vigreux. Low-boiling fraction A and high-boiling pot residues are discarded.

Fraction B contains the modified alkylbenzenes of interest, and is weighed.

Appropriate temperatures for collecting fractions A and B can be determined from the article written by Samuel B. Lippincott and Margaret M. Lyman, published in Industrial and Engineering Chemistry, Vol. 38, in 1946, and starting at page 320. Using the Lippincott et al. Article.

Next, an aliquot sample of fraction B is analyzed by the gas chromatography method using the above conditions. If more than 90% of the

total GC area for fraction B is within the retention time range, then the impurities in fraction B are deemed to be not more than 10 wt-% of the selective liquid product, and, for the sole purpose of computing the selectivity to internal quaternary phenyl-alkanes, CMAB is computed by dividing the weight of fraction B collected by the weight of the aliquot portion of the selective liquid product charged to the 5-liter flask in the above distillation method. If the percent of the total GC area for fraction B within the retention time range is less than 90%, then the impurities in fraction B are deemed greater than 10 wt-% of fraction B and impurities removal again follows using the above distillation method to separate a low-boiling fraction C and pot residues from a fraction D containing the modified alkylbenzenes of interest. Fraction D is recovered, weighed, and analyzed by the gas chromatography method to determine if it meets the same 90% criteria for the total GC area for fraction within the retention time range. If so, CMAB is computed by previously described procedure. If not, the distillation and gas chromatography methods are repeated for fraction D.

The above-described distillation and gas chromatography methods can be repeated until a fraction containing the modified alkylbenzenes of interest and having less than 10 wt-% impurities is collected, provided that sufficient quantity of material remains after each distillation for further testing by these methods.

Then, once CMAB is determined, the selectivity to internal quaternary phenyl- alkanes, Q, is computed using the above formula.

The results of these analyses are shown in the Table 8: Table 8: Liquid Product Analysis Internal Quaternary 2-Phenyl-Alkane End Quaternary Phenyl- Selectivity Alkane Selectivity Phenyl-Alkane Selectivity 82.0% 6.98% 1.9% In the absence of shape selectivity, most of the 2-methyl undecene would be expected to form 2-methyl-2-phenyl undecane (that is, an end quat).

Likewise, most of the 6-methyl undecene, 5-methyl undecene, 4-methyl undecene, and 3-methyl undecene would be expected to form internal quats.

The linear olefins would be expected to produce a statistical distribution of 2- phenyl-dodecane, 3-phenyl-dodecane, 4-phenyl-dodecane, 5-phenyl-dodecane, and 6-phenyl-dodecane. Thus, if the lights, the heavies, and the other alkyl olefins listed in Table 5 are excluded from the computations, the 2-phenyl- alkane selectivity would be no greater than 17 and the internal quaternary phenyl-alkane selectivity would approach 55. Table 8 shows that the 2-phenyl- alkane selectivity is significantly higher than expected in the absence of shape selectivity and that the internal quaternary alkylbenzene selectivity obtained using the mordenite catalyst is much less than the internal quaternary alkylbenzene selectivity that would be expected in the absence of shape selectivity.