Reference Library

Alkylation will be key process in reformulated gasoline era

Regulations mandated gasoline reformulation to reduce motor vehicle emissions during the 1990s and boosted the importance of the alkylation process in refineries.Constituents in the alkylate produced do not contribute to ozone formation. The low vapor pressure of alkylate, a high octane blending component, helps refiners maintain volatility (Rvp) specifications. About 11% of the gasoline pool in American refineries is produced by alkylating isobutane with C3-C5 olefins. Current alkylation capacity in the United States totals about 960,000 b/d of alkylate. Although alkylation is much less important in other countries, its relative importance is growing. Worldwide alkylation capacity totals at least 1.15 million b/d.

Alkylation offers several key advantages to refiners, including the highest average quality of all components available to the gasoline pool, increased amounts of gasoline per volume of crude oil, and high heats of combustion. Alkylates permit use of internal combustion engines with higher compression ratios and hence the potential for increased miles per gallon. Alkylates burn freely, promote long engine life, and have low levels of undesired emissions.

Commercial alkylation plants use either sulfuric acid (H2S04) or hydrogen fluoride (HF) as catalysts. About 20 years ago, almost three times as much alkylate was produced using H2S04 as the catalyst as compared to processes using HF.

Since then, the relative importance of processes using HF has increased substantially, and currently these processes produce in the U.S. about 47% of the alkylate. In the last 5 years, however, more H2S04-type units have been built in the U.S. than HF-type units. Recent information that clarifies the dangers of HF are causing refineries that use HF to reconsider and/or improve their safety equipment and procedures.

This first installment of a two-part series reviews the complicated chemistry of alkylation as it applies to processes currently in commercial operation. The major H2S04 alkylation processes are described and compared, along with several process modifications that have occurred within the last few years. The second installment will discuss the processing details of hydrogen fluoride alkylation and some improvements, and the relative advantages of H2SO and HF alkylation will be compared.

Chemistry and fundamentals

Alkylation of isobutane with C3-C5 olefins involves numerous consecutive and simultaneous ionic reactions. The reaction sequences have within the last 10-20 years been found to be much more complicated than the relatively simple chain reaction sequences proposed many years ago. The olefins often react to a much higher degree in the early stages of the reactions than isobutane. The chemical sequences differ significantly when comparing H2S04 and HF or different olefins such as n-butenes and isobutylene.

The compounds: 1-butene, 2 butenes, and propylene generally react to a large extent first with the acid, and especially H2S04, to form isoalkyl esters. These esters later decompose to free an olefin and the acid.

Several key features of the chemistry are summarized because they help explain various industrial findings.

Alkylate production

Alkylates are produced by three mechanisms designated as Mechanisms 1, 2, and 3 (Figure 1.). In Mechanism 1, C3-C5 olefins react with t-C4H9+ (a key intermediate cation) to produce C7 to C9 cations. Hydride transfer to these cations produces C7 to C9 isoparaffins (or alkylate). When the olefin is propylene, 1-butene, 2-butenes, or iso-butylene, and C5 olefins (amylenes), the resulting isoparaffins produced are di-methylpentanes, dimethyl-hexanes (DMHs), trimethyl-pentanes (TMPs), and various C9 isoparaffins, respectively.

With 2-butenes or isobutylene, alkylates having research octane numbers (RONs) of about 102-103 are produced via Mechanism 1. With H2S04, most of the 1-butene isomerizes to 2-butenes before alkylation, so that the composition and quality of the alkylate produced from either 1-butene or 2-butenes are essentially identical. With HF, only part of the 1-butene isomerizes, hence, many DMHs are formed by Mechanism 1. The DMHs have very low RONs, even lower than those for dimethyl-pentanes or Cg isoparaffins.

Mechanism 2 involves the formation of relatively heavy cations, such as C12-C20 cations, as intermediates by polymerization-type reactions. These cations then fragment to produce C5-Cg cations and olefins. Subsequent hydride and proton transfer results in the formation of C5-C9 isoparaffins (or alkylate).

Light ends, often most of the DMHs, part of the TMPs, and much of the heavy ends, are produced by Mechanism 2. The alkylates produced in this manner have RONs of about 88-89.

Much of the isobutylene and branched C5 olefins alkylate via Mechanism 2, whereas smaller fractions of 1-butene, 2-butenes, propylenes, and n-pentanes do not. In the case of isobutylene, importance when H2S04 is used as a catalyst.

In Mechanism 3, heavy cations, such as formed in Mechanism 2, are converted by hydride transfer to heavy isoparaffins. The quality of these isoparaffins is relatively low, often in the 86-88 RON range.

The relative importance of Mechanisms, 1, 2, and 3 are estimated as a function of RON when H2S04 is used as the catalyst and when only C4 olefins are used to alkylate isobutane, as shown in Figure. 1. Mechanism 1 is the predominant one for alkylates with RON values of 96 or higher, but Mechanism 2 is predominant at 90 to about 95 RON.

Hydrogen transfer reactions are important when HF is used as a catalyst, but not so important for H2S04. Overall, 2 moles of isobutane and 1 mole of the olefin react to form 1 mole of a C8 isoparaffin (usually TMP) and 1 mole of a paraffin with the same number of carbon atoms as the olefin.

With propylene and HF, often about 20% of the propylene is converted to propane. UOP data indicate that a small amount of n-butane is produced from n-butenes.

Figure 1: Comparison of C4 alkylation mechanisms.

Acid consumption and conjunct polymer production

Conjunct polymers are a by-product in commercial alkylation processes. They are acid soluble, frequently being referred to as acid-soluble oils (ASOs) or red oil. They are excellent hydride transfer agents acting as a reservoir for the hydride ions being transferred from the isobutane, and they act as surfactants to promote increased interfacial surface areas. Too many conjunct polymers lead, however, to inactive catalysts which must then be regenerated.

About 0.4-0.6 lb of H2SO4 is frequently required to produce 1 gal of alkylate, but much lower values, such as 0.1-0.25 ppg, can be realized at preferred conditions. Acid costs frequently account for about one third of the total operating costs of alkylation when H2S04 is used as catalyst. There is hence considerable incentive to reduce H2S04 consumption

HF consumption is often in the range of 0.08-0.25 lb HF/gal of alkylate. Regeneration of used HF is relatively easy and cheap. Because most of the HF is recovered and recycled, the amount of make-up HF required is small, usually about 0.15-0.2 lb/bbl of alkylate. The conjunct polymers (or tarry bottoms) from the distillation column contain some HF. There are sometimes environmental problems in disposing of them. HF wastes can be treated with ASOs, or with calcium chloride and caustic.

The sulfuric acid composition has an important effect on alkylate quality and on the kinetics of the reactions. The optimum acid compositions differ appreciably. For 2-butene feed, the acid contains about 5-6% conjunct polymer, or red oil, and 0.5-1.0% water.

Water is always produced so that used H2S04 contains considerably more water by the time 5-6% conjunct polymers have formed. Reducing the water content in the acid is desirable. Acid consumption would likely be decreased by a factor of at least 2.

For HF, a water content of about 2.8% resulted in the best quality alkylate. Higher amounts of heavy ends, light ends, and DMHs are produced with lower concentrations of water.

Furthermore, alkylate equals and yields decrease the acid contains more man 1.0% conjunct polymers. Corrosion becomes a problem in the alkylation unit when the water content increases beyond 1.5%.

Acid consumption generally decreases with lower reaction temperature, higher ratios of isobutane to olefin, improved agitation of the two-phase reaction mixture, and use of acids that promote high-quality alkylates. Considerably less acid is consumed when 2-butenes and 1-butene are used as compared to isobutylene.

Some refineries use isobutylene to produce methyl tertiary butyl ether (MTBE), and consequently use an olefin stream for alkylation that is mainly n-butenes. In such cases, acid consumption is normally relatively low.

Acid consumption is also relatively high when appreciable amounts of propylene used. Based on recent information, normal or branched amylenes (or pentanes) result in acid consumption values lower than those of propylene. Unfortunately, C5 olefin streams in a refinery often contain significant amounts of cyclopentene, conjugated dienes, and acetylenic compounds which result in high acid consumption.

It was reported how various impurities in the hydrocarbon feeds affect H2S04 consumption. These values are approximate because acid consumption depends significantly on the operating conditions employed.

Reference used L. F. Albright: Alkylation will be key process in reformulated gasoline era, Oil & Gas Journal, 77, 79-92, 1990.

Commercial alkylation processes

The H2SO4 processes to be covered include: effluent refrigeration, cascade (or au-to refrigeration), and time-tank processes.

Effluent refrigeration

The effluent refrigeration process of Stratco Inc., (H2SO4 catalyst) is employed in many refineries throughout the world. The total capacity of these units is in excess of 400,000 b/d.

At least six plants have been built and put in operation since 1985. Other plants are currently in either the building or advanced planning stages.

This process accounts for more than 60% of the worldwide production of alkylate using H2S04 as the catalyst. The Stratco contactor reactor is a horizontal vessel with an impeller in one end and a tube bundle (or cooling coils) in the other. The hydrocarbon-acid suspension in the contactor is circulated through the outer shell and then passed through the tube bundle to remove the exothermic heat of alkylation and the energy provided by the impeller. Both the hydrocarbon feed mixture of isobutane and olefins and the acid feed are introduced to the eye of the impeller. The feed streams rapidly form a two-phase suspension which is maintained by the rapid circulation. Stratco contactors are usually sized to produce 2,000 b/d of alkylate. Changes were made in the contactor in the last few years to obtain improved quality alkylate and lower operating costs. Longer coils were included to obtain substantially more heat transfer surface. Improved positioning of the coils has resulted in 50% increases of the overall heat transfer coefficients. The contactor includes an improved pump/agitator system. The injection devices for introducing the hydrocarbon feeds and the acid into the contactor have been improved. The pump/agitator has been positioned below the center line in order to minimize partial settling of acid in the bottom of the contactor. Part of the two-phase suspension is continuously removed and sent to the acid settler (or decanter). The decanter was previously operated at residence times of about 1.0 hr to obtain complete separation of the two phases. Some undesired reactions, including conjunct polymer production, however, occurred. The decanter is currently operated with much lower residence time for the acid. Some hydrocarbons are recycled with the acid phase, but acid consumption is reduced and alkylate quality improved. The hydrocarbon phase, which is often at a temperature of about 10°C or less, is passed through a pressure-reducing valve. As a result, some of the more volatile hydrocarbons, and particularly isobutane, are vaporized and the temperature of the resulting liquid is decreased, often to about 0°C. This cold liquid is used as the coolant in the tube bundle of the contactor. The gas stream formed by partial flashing of the effluent hydrocarbon stream is mainly isobutane plus small amounts of propane and C5-C7 isoparaffins. This mixture is compressed and condensed using cooling water as the coolant. Propane is then removed by fractionation in the depropanizer. The bottom product of this column is partially flashed to produce a cold liquid which is recycled as part of the isobutane feedstreams to the contactor. The vapors from the last flashing step are once again compressed, condensed, and recycled. isobutane is hence both a refrigerant and a reactant. The liquid stream remaining after the flashing of the effluent hydrocarbon stream contains mostly isobutane and alkylate. It also contains a small amount of hydrocarbon-soluble sulfates such as di-sec-butyl and di-isopropyl sulfates. Three methods are employed to remove these sulfates. The liquid stream is washed first with fresh acid and then with a dilute caustic solution. During the acid washing, the sulfates react with H2SO4 to form acid-soluble sec-butyl acid sulfate and isopropyl acid sulfate. Acid washing results in slightly improved yields of alkylates and somewhat reduced acid consumption. Caustic washing is needed to remove the residual acid and sulfates from the hydrocarbon liquids.

The following method is also sometimes employed to remove hydrocarbon-soluble sulfates. The liquid hydrocarbon mixture is first washed with a fairly concentrated caustic solution and then with water. During the caustic wash, Na2S04 and a low-quality alkylate are formed. It is not easy to get intimate contact with the caustic solution, and sulfate removal is sometimes incomplete.

Water washing is needed to remove residual caustic. This technique is generally less preferred. The following sulfate removal technique is often preferred. The hydrocarbon liquids are first acid washed and then contacted with a bed of bauxite powder. The bauxite effectively adsorbs residual acid, sulfates, and water. A dry feed minimizes corrosion. This procedure has somewhat increased operating costs, but it is sometimes cost effective.

After the sulfates are removed from the hydrocarbon liquids, the resulting mixture is fed to the deisobutanizer (DIB). In larger plants, the DIB is generally a distillation column (and hence has a reflux condenser). In smaller plants, a stripping column is often used. The top product of the DIB is mainly isobutane which is cooled and fed as a liquid to the contactor. The stream is premixed with the olefin feed before entering the contactor. A side am rich in n-butane is generally withdrawn from the DIB. Steam is the heat source of the DIB. Better separations occur in distillation columns.

Cascade autorefrigeration

Cascade-type reactors are large horizontal drums. Generally, three to seven reactor stages are provided in one end of the reactor, and an acid settler (or decanter) is provided in the other end. Each stage is partially filled with liquid (an acid-hydrocarbon suspension), and a gas mixture is above the liquid. This process is widely used, and additional reactors have recently been installed in at least two refineries.

The isobutane and acid feedstreams are introduced to the first stage. Equal amounts of olefin are, however, added to each stage. Unreacted isobutane, alkylate, and acid flow over baffles (or weirs) from one stage to the next and eventually reach the acid settler. The temperature in each stage is controlled by vaporizing part of the more volatile hydrocarbons (mainly isobutane plus minor amounts or propane and C5-C7 isoparaf-fins). The impellers of a cascade reactor are positioned at the throat of venturi-shaped casings that open into cylindrical hollow housings. Several vertical tubes are positioned in the upper portion of the housings. The impeller produces an emulsion that is acid continuous, and the emulsion is pumped upward. The olefin feed to the stage is added just above the impeller and mixed into the emulsion. The emulsion is rapidly recirculated in and around the housing. The upper portion of the emulsion, and especially that outside the housing, is generally hydrocarbon continuous. Some of it is transferred to the next stage.

In a cascade reactor of different design, a turbine impeller is used, and the olefin feed is jetted into the eye of the impeller. The emulsion in this reactor presumably remains acid-continuous throughout the stage. Nevertheless, the average residence time of the acid phase in a stage is always greater than that of the hydrocarbon phase because of the density difference.

In a cascade reactor, the isobutane-olefin ratio in a five-stage cascade reactor have a combined ratio of 8:1, Stage 1 is about 35:1 and Stage 5 is 12:1.

This apparent advantage of a cascade reactor as compared to a Stratco contactor tends to be negated by the fact that the olefin feed is not premixed with isobutane. The olefin feeds tend to contact the acid phase first when the emulsions are acid-continuous at the injection points. Consequently, undesired reactions often occur that result in conjunct polymer formation plus polymerization reactions. The following conclusions apply to a cascade reactor:

  • The highest quality alkyl-ate is produced in Stage 1 and the poorest in Stage 5 because of the difference in isobutane-to-olefin ratios in the stages.
  • More acid-soluble hydro- carbons are produced in latter stages of the reactor.
  • Alkylate produced in the initial stages degrades to at (its a small extent in the final stages of the reactor.
  • The sec-butyl and isopropyl sulfates produced in the last stage of a cascade reactor react to a slightly lesser extent as compared to similar sulfates produced in a Stratco contactor. This conclusion is based on the shorter residence time in the last stage as compared to that of a Stratco contactor.

The remainder of the alkylation process using cascade reactors has a flowsheet very similar to the one for the effluent refrigeration process.

Time-tank process

Several refineries still operate time-tank processes, but apparently none have been built in the past 25 years. The following features are, however, typical:

  • Mixtures of isobutene and olefins are contacted with H2S04 in a fairly large cooled pipe near the entrance of a centrifugal pump. The acid and hydrocarbon phases then enter the pump which provides mixing and emulsification. Temperature rises because the exothermic reactions are moderate due to fairly high ratios of acid-to-hydrocarbon phases.
  • The emulsion enters a heat exchanger cooled by a refrigerant where the heat of reaction is removed.
  • The emulsion then is introduced to the middle section of a large vertical column called the reactor. Reactions occur here, but many have occurred before the mixture entered the reactor.

This vessel also acts as a decanter. Acid removed from the bottom of the reactor is recycled to the centrifugal pump.

  • The hydrocarbon steam from the top of the reactor is processed as follows: First, acid and dissolved are removed, and second, the hydrocarbons are separated in multiple distillation columns to recover un-reacted isobutane, alkylate, propane, and n-butane. The isobutane stream is recycled to the time-tank reactor.

H2SO4 process comparison

Of the three processes employing H2SO4 as catalyst, the effluent refrigeration process and the cascade process are generally considered to be superior to the time-tank process. One refiner who uses the time-tank process claims to produce a better quality alkylate and to consume less acid as compared to the cascade process. Slightly poorer performance compared to the effluent refrigeration and cascade processes results from higher energy costs for the time-tank process. All unreacted isobutane must be separated from the alkylate by distillation, and because the refrigeration method uses more energy.

Unless major improvements are made in the design and operation of time-tank reactors, additional plants will likely not be built in the future.

One major company has made a direct comparison between its cascade and effluent refrigeration units. The company attempted to use identical feeds and operating conditions in making this comparison. Details of the specific reactors used and the operating conditions employed are not known, but better alkylate quality and less acid consumption occurred in the effluent refrigeration process.

A recently built effluent refrigeration process uses C4 olefins. The alkylates produced in a series of test runs had 96.4-97.3 RON and 93.6-94.2 MON. Acid consumption was less than 0.3 lb acid/gal alkylate.

Computer models have been developed by various companies for cascade and effluent refrigeration processes. The reactor used in cascade process had modified internals and resulted in acid consumption values of 0.2 lb H2S04/gal of alkylate. Recent Stratco pilot plant results confirm that such low acid consumption occurs at preferred operating conditions and with preferred olefins. In many cases, removal of undesired impurities from the olefins is cost-effective.

The size and energy demands of both the compressors and DIB differ substantially for the effluent refrigeration and cascade processes. The compressor in each process serves two purposes: compresses and recycles the isobutane that is flashed in order to maintain the desired reaction temperature, and compresses the isobutane that is flashed as refrigerant in order to cool the feed hydrocarbons to the desired reaction temperature.

The two processes were compared at conditions that are considered preferred in both cases. The following operating conditions were assumed for each process: reaction temperature, 43° F.; alkylate yield, 1.753 bbl of true alkylate/bbl olefin; 68% isobutane in reactor effluent; refrigerant condensation temperature, 100° F.; and compressor efficiency, 75%. The temperature difference for heat transfer in the contactor was set at 11.5° F.

In summary, the compressor demands for the effluent refrigeration process are larger, more isobutane is vaporized, and a larger pressure increase occurs in the compressor.

Steam requirements in the DIB column of the cascade process are significantly higher because more isobutane is separated. The combined energy costs for the compressor and DIB column are most identical for the effluent refrigeration and cascade processes.

Calculations indicate that contactors which result in smaller temperature differences for heat transfer result in smaller energy requirements for the compressor and larger ones for the DIB tower. Substantial energy reductions occur for the compressors of both processes if the isobutane is condensed at lower pressures (and hence at lower temperatures). Such condensation can be realized if the temperature of the cooling water is relatively low as sometimes occurs in winter. The ratio of isobutane to olefins used also affects the comparison.

Improved operations in a cascade reactor would be expected if the isobutane and olefins were premixed before being introduced into the first stage, the olefins were introduced into the hydrocarbon phase in the reactor or into the hydrocarbon-continuous emulsion sometimes present, and jetting the olefin feed into the emulsion in order to provide better mixing.

Methods of H2SO4 addition

When a large alkylation plant is built, two or more reactors are generally needed. The reactors are frequently arranged with the acid being fed in series flow from one reactor to the next. In such a case, the acidity of the acid decreases from one reactor to the next. In one example, fresh acid is continuously introduced in the first of three cascade reactors, and used acid is continuously withdrawn from the third reactor. RONs for the alkylates a 95.4, 96.6, and 96.0 for Reactors 1, 2, and 3, respectively. With more than five reactors, combinations of series and parallel flows of acid are sometimes preferred. The objective is to employ acids with preferred compositions in as many reactors as possible.

For a plant using a single reactor, two methods can be used for adding fresh acid (and for removing the used acid). The new acid can be added continuously (and the old acid can be removed continuously).

Alternatively, a given batch of acid can be recycled until the acidity has dropped to a given end point and then most of the acid is replaced. The latter technique usually produces a better average quality alkylate.

Reference: L. F. Albright: Alkylation will be key process in reformulated gasoline era, Oil & Gas Journal, 77, 79-92, 1990.

History of Alkylation

Alkylation, first commercialized in 1938, experienced tremendous growth during the 1940s as a result of the demand for high-octane aviation fuel during World War II. During the mid-1950s, refiners' interest in alkylation shifted from the production of aviation fuel to the use of alkylate as a blending component in automotive motor fuel. Capacity remained relatively flat during the 1950s and 1960s due to the comparative cost of other blending components. The U.S. Environmental Protection Agency's lead phase-down program in the 1970s and 1980s further increased the demand for alkylate as a blending component for motor fuel. As additional environmental regulations are imposed on the worldwide refining community, the importance of alkylate as a blending component for motor fuel is once again being emphasized. Alkylation unit designs (grassroots and revamps) are no longer driven only by volume, but rather by a combination of volume, octane, and clean air specifications. Lower olefin, aromatic, sulfur, Reid vapor pressure, and drivability index specifications for finished gasoline blends have also become driving forces for increased alkylate demand in the United States and abroad.

The alkylation reaction combines isobutane with light olefins in the presence of a strong acid catalyst. The resulting highly branched, paraffinic product is a low-vapor-pressure, high-octane blending component. Although alkylation can take place at high temperatures without catalyst, the only processes of commercial importance today operate at low to moderate temperatures using either sulfuric or hydrofluoric acid catalysts. Several different companies are currently pursuing research to commercialize a solid alkylation catalyst. The reactions occurring in the alkylation process are complex and produce an alkylate product that has a wide boiling range. By optimizing operating conditions, the majority of the product is within the desired gasoline boiling range with motor octane numbers (MONs) up to 95 and research octane numbers (RONs) up to 98.

Reference: R. A. Myers: Handbook of petroleum refining processes, ISBN: 0-07-139109-6, 2003.

Process design aspects of alkylation

In designing alkylation processes, several aspects are especially important.


The level of agitation and the ability to provide intimate contact between the two liquid phases is always of great importance. For alkylation with H2S04, the main reactions occur at, or at least near, the interfaces between the liquid phase. The following design factors are important:

  • Both the type of agitator and the rate of agitation. In one comparison using H2S04 and 2-butenes, the alkylate quality increased by about 7.5 RON as the agitation speed increased from 1,000 to 3,000 rpm.
  • Baffling and the location of the impeller in the reactor vessel.
  • Continuous liquid phase in reactor. With H2S04, acid-continuous suspensions produce alkylates having higher quality by about 1-3 RON than hydrocarbon-continuous emulsions.
  • The olefin feed should preferentially be premixed with the isobutane before entering the reactor.

High levels of agitation are less important for alkylation using HF because HF has much lower viscosities than H2S04, and HF is slightly soluble in the hydrocarbon phase.

Temperature control

Alkylation reactions are highly exothermic, necessitating cooling. When H2S04 is used, improved quality alkylates are generally obtained as the reaction temperatures are reduced to about 41° F.

At lower temperatures, the viscosity of the acid increases and effective agitation is a problem. Refrigeration costs of course increase as the reaction temperatures are decreased.

When HF is used, alkylation occurs normally at 86-113° F, and cooling water is generally used as a coolant. This means a substantial cost saving compared to alkylation using H2S04.

Lower temperatures substantially improve the quality of the alkylates produced from 2-butenes and isobutylene when HF is used, but the quality is greatly decreased in the case of 1-butene.

Isobutane/olefin ratio

Higher ratios of isobutane to olefins in the feed streams to the reactor minimize the undesired polymerization reactions. The quality of the alkylate hence increases as the ratios increase (especially when the ratios are relatively small). Costs of separating and recycling the unreacted isobutane and cooling both become larger as the ratio increases. Alkylation plants employing H2S04 as the catalyst often operate in the range of 5:1 to 8:1. Plants employing HF generally operate at higher ratios, such as 10:1 to 15:1.

Residence time

Slightly higher quality alkylates and lower acid consumption normally occur if the residence time of the hydrocarbon-acid suspension in the reactor, is increased, i.e., the space velocities are decreased. Longer residence times, however, reduce the capacity of a reactor and increase operating expenses.


Plant results show that several additives to H2S04 result in improved production rates, improved alkylate quality, and/or decreased acid consumption. These additives, which presumably accumulate at the acid-hydrocarbon interface, result in improved hydride ion transfer and in improved interracial surface tensions. Fluorosulfonic acid promoters gave beneficial results when added to HF.

In summary, improved quality alkylate and generally lowered acid composition can be obtained by several operation changes, but unfortunately all tend to increase operating costs. Operating conditions for an alkylation plant should obtain maximum profits.

Reference used: L. F. Albright: Alkylation will be key process in reformulated gasoline era, Oil & Gas Journal, 77, 79-92, 1990.

Pump Shaft Sleeves

Pump shafts are usually protected from erosion, corrosion, and wear at seal chambers, leakage joints, internal bearings, and in the waterways by renewable sleeves.

The most common shaft sleeve function is that of protecting the shaft from wear at packing and mechanical seals. Shaft sleeves serving other functions are given specific names to indicate their purpose. For example, a shaft sleeve used between two multistage pump impellers in conjunction with the interstage bushing to form an interstage leakage joint is called an interstage or distance sleeve.

In medium-size centrifugal pumps with two external bearings on opposite sides of the casing (the common double-suction and multistage varieties), the favored shaft sleeve construction uses an external shaft nut to hold the sleeve in an axial position against the impeller hub. Sleeve rotation is prevented by a key, usually an extension of the impeller key. The axial thrust of the impeller is transmitted through the sleeve to the external shaft nut.

In larger high-head pumps, a high axial load on the sleeve is possible and a design similar to that shown in Figure 1 may be preferred. This design has the advantages of simplicity and ease of assembly and maintenance. It also provides space for a large seal chamber and cartridge-type mechanical seals. When shaft sleeve nuts are used to retain the sleeves and impellers axially, they are usually manufactured with right- and left-hand threads. The friction of the pumpage and inadvertent contact with stationary parts or bushings will tend to tighten the nuts against the sleeve and impeller hub (rather than loosen them). Usually, the shaft sleeves utilize extended impeller keys to prevent rotation.

Figure 1: A sleeve with an internal impeller nut, external shaft-sleeve nut, and a separate key for the sleeve.

Some manufacturers favor the sleeve, in which the impeller end of the sleeve is threaded and screwed to a matching thread on the shaft. A key cannot be used with this type of sleeve, and right- and left-hand threads are substituted so that the frictional grip of the packing on the sleeve will tighten it against the impeller hub. As a safety precaution, the external shaft nuts and the sleeve itself use set screws for a locking device.

In pumps with overhung impellers, various types of sleeves are used. Most pumps use mechanical seals, and the shaft sleeve is usually a part of the mechanical seal package supplied by the seal manufacturer. Many mechanical seals are of the cartridge design, which is set and may be bench-tested for leakage prior to installation in the pump.

For overhung impeller pumps that utilize packing for sealing, the packing sleeves generally extend from the impeller hub through the seal chambers (or stuffing boxes) to protect the pump shaft from wear. The sleeves are usually keyed to the shaft to prevent rotation. If a hook-type sleeve is used, the hook part of the sleeve is clamped between the impeller and a shaft shoulder to maintain the axial position of the sleeve. A hook-type sleeve used to be popular for overhung impeller pumps that operate at high temperatures because it is clamped at the impeller end and the rest of the sleeve is free to expand axially with temperature changes. But with the increased use of cartridge-type mechanical seals, the use of hook-type sleeves is diminishing.

In designs with a metal-to-metal joint between the sleeve and the impeller hub or shaft nut, a sealing device is required between the sleeve and the shaft to prevent leakage. Pumped liquid can leak into the clearance between the shaft and the sleeve when operating under a positive suction head and air can leak into the pump when operating under a negative suction head. This seal can be accomplished by means of an O-ring, or a flat gasket. For high temperature services, the sealing device must be either acceptable for the temperature to which it will be exposed, or it must be located outside the high temperature liquid environment. According to an alternative design used for some high-temperature process pumps, the contact surface of the hook-type sleeve and the shaft is ground at a 45-degree angle to form a metal-to-metal seal. That end of the sleeve is locked, but the other is free to expand with temperature changes.

When O-rings are used, any sealing surfaces must be properly finished to ensure a positive seal is achieved. All bores and changes in diameter over which O-rings must be passed should be properly radiused and chamfered to protect against damage during assembly. Guidelines for assembly dimensions and surface finish criteria are always listed in O-ring manufacturers’ catalogs.

Reference: I. J. Karassik, C.C. Heald: Centrifugal Pumps: Major Components, Pump Handbook, ISBN 0-07-034032-3, 2001.