Jones D.S.J., Pujado P.R. Handbook of Petroleum Processing

Подождите немного. Документ загружается.

9.2 Catalytic oleﬁn condensation

Peter R. Pujad´o and Dennis J. Ward (retired)*

Introduction

Catalytic oleﬁn condensation refers in general to the reaction of one molecule of an

oleﬁn with one or more molecules of the same oleﬁn or of other oleﬁns to yield heavier

oleﬁnic compounds. The term “condensation” reﬂects the fact that liquid products

are obtained from gaseous oleﬁns.

Oleﬁn condensation is in some ways similar to motor fuel alkylation, except that in

alkylation we react an oleﬁn with an isoparafﬁn to yield a parafﬁnic compound. In

oleﬁn condensation, on the other hand, the products that are made are oleﬁnic, but

may be later hydrogenated in order to meet the requirements and speciﬁcations of the

end products.

Oleﬁn condensation can be carried out in many ways and over a diversity of catalysts.

The common characteristic is that oleﬁn condensation takes place over acid catalysts.

These catalysts may be solid phosphoric acid (SPA), liquid acids, organometallic

catalysts, zeolites, sulfonic acid resins, etc.

Although oleﬁn condensation ﬁnds broad application in the petroleum reﬁnery en-

vironment, it is also extensively used in petrochemical applications (often combined

with reﬁnery operations) as, for instance, in the production of heptenes, nonenes,

dodecenes, etc. often used for the production of alkylaromatic derivatives. In fact,

some of the same catalysts (e.g., SPA) can also be used for the alkylation of oleﬁns

with aromatics as in the production of cumene (isopropyl-benzene), ethylbenzene,

etc. This review will concentrate on reﬁnery and oleﬁn oligomer applications and

will not discuss aromatic alkylation applications.

*UOP LLC.

372

GASOLINE COMPONENTS 373

History

The discovery of acid catalysis, and in particular the development of solid phosphoric

acid (SPA) catalysts, is attributed to the pioneering work of Vladimir Ipatieff at

what was then called Universal Oil Products and is today known as UOP. The ﬁrst

commercial unit for the production of octenes by the dimerization of butenes came

on-stream in 1935. By introducing this new ingredient in the formulation of gasolines,

it greatly contributed to the popularization of the automobile that had hitherto been

constrained by the limited availability of straight-run gasolines from the fractionation

of petroleum crudes. In addition, the octenes obtained by butene dimerization had a

higher octane number than naturally occurring straight-run gasolines. Several hundred

units were soon built to satisfy the growing demand for gasoline. Commercially,

the dimerization and oligomerization of oleﬁns over SPA catalysts is known as the

Catalytic Condensation process. This is still the dominant process for the dimerization

and oligomerization of oleﬁns. It is only recently that more work has been done on

the development of acid zeolite catalysts and, in some cases, of sulfonic acid resin

catalysts.

The other type of catalysts most often found in reﬁnery applications for the produc-

tion of gasoline components are organometallic catalysts developed by the Institute

Fran¸cais du P´etrole (IFP) and used, for example, in the dimerization of propylene

to isohexenes (Dimersol G process), or in the oligomerization of light oleﬁns from

cracked gases (Dimersol E), or in the production of heptenes and octenes from propy-

lene and/or butenes (Dimersol X).

Because the production of oleﬁn dimers and oligomers almost invariably produces a

blend of isomers and often also a lighter and a heavier fraction, the production of these

oleﬁns for petrochemical applications cannot be easily segregated from their appli-

cation in gasoline and diesel fuel uses. However, it is usually the case that catalysts

can be tailored within certain limits so as to yield more branched compounds (higher

octane) for gasoline applications and less branched products for petrochemical appli-

cations. Because the products from such processes always have some branching, the

quality of the products in the diesel range is somewhat poor (low cetane numbers)

and their use in diesel applications is not that extensive.

Catalytic condensation process

The catalytic condensation process was developed in the early 1930s as a means to

convert light gases produced by thermal cracking and thermal reforming into useful

products. In the early years of the 20th century, most operations in the petroleum

reﬁning industry were limited to the distillation of crude oils into gasoline boiling

374 CHAPTER 9

range materials (straight-run gasolines), kerosenes, and lamp fuels. The yields in the

gasoline fraction were low (sometimes as low as 5%, depending on the crude) and

did not meet the requirements of the growing automobile industry.

One of the early ways to increase gasoline yields was by thermal cracking. The

dominant thermal cracking technology was the Dubbs Process. In fact, out of the de-

velopment of the Dubbs Process emerged the Universal Oil Products company (now

UOP). A by-product from the Dubbs Process was a light gas stream that contained

saturated and oleﬁnic components that ranged from methane to butenes. Although

there was no immediate use for those gases, reﬁners knew that by circulating these

gases through a strong mineral acid it was possible to condense the oleﬁns into heav-

ier liquid products, and actually this was the principle applied in the Orsat apparatus

used to measure the oleﬁn content of those gases. Use of strong mineral acids (e.g.,

sulfuric acid) poses signiﬁcant corrosion problems and presents considerable opera-

tional difﬁculties. Ideally, it was desired to operate with a single carbon steel vessel

in which the strong acid could be contained. R. E. Schaad and V. N. Ipatieff dis-

covered that phosphoric acid mixed with kieselguhr (an abundant naturally occurring

siliceous diatomite mineral) solidiﬁed upon heating. When the solid mass was cooled,

broken up, and placed inside a reactor, the propylene and/or butenes condensed into

a liquid in the gasoline boiling range. This discovery was scaled up and eventually

commercialized (1).

Shortly after the introduction of the Catalytic Condensation process came the catalytic

cracking process, initially as a ﬁxed process and later as the ﬂuidized-bed catalytic

cracking process, or FCC. These catalytic cracking processes, and in particular FCC,

yielded even more oleﬁns than thermal cracking, principally propylene and butenes,

which at that time had very little commercial value. Thus, further application of the

Catalytic Condensation process ensued.

The early operations were, by today’s standards, primitive. To manufacture the catalyst,

phosphoric acid and kieselguhr were mixed together, formed into a cake that was

heated and then cooled, and broken up with a sledge hammer. The pieces were screened

into the proper size range and loaded into the reactor. The reactor operated at very

low pressures and produced such amounts of carbon that the catalyst had to be burnt

off every other week, and be completely replaced every four to ﬁve weeks. The

technology evolved by ﬁrst improving the catalyst both in terms of manufacture (e.g.,

by producing an extrudate) and of formulation and also by optimizing the operating

conditions.

Early units operated at about 10 atmospheres, since that was the pressure level of many

liqueﬁed petroleum gas (LPG) recovery systems. However, it was soon discovered

that, if the reactor was operated at 30 or more atmospheres, catalyst lives could be

extended by a factor of four to six times. Thus, began the evolution of higher pressure

GASOLINE COMPONENTS 375

units. Some units that employed tubular reactors at 60 atmospheres were constructed

and were found to be even more successful.

SPA catalyst was also tried for other reactions. In the late 1930s, it was discovered

that SPA would catalyze the reaction of oleﬁns with aromatic compounds to yield

mono-oleﬁnic alkyl aromatics. This fortunate discovery was widely used to make

cumene (isopropyl-benzene) for aviation fuel used in World War II. Cumene has a

very high performance number necessary for the aviation gasoline used in internal

combustion engines in aircraft. After the war, use of cumene in aviation fuel declined,

but cumene found an even more important outlet as feedstock for the production

of phenol and acetone—to this date, this petrochemical use of cumene continues

to dominate this industry, even though in recent years there has been a signiﬁcant

shift from SPA to acid zeolite catalysts for this aromatic alkylation. During World

War II there was also a high demand for ethylbenzene and styrene. Some units were

built to produce ethylbenzene by reacting ethylene with benzene over SPA catalyst.

The reaction was slow and fairly inefﬁcient, so SPA never gained relevance for this

application; most ethylbenzene units at the time made use of aluminum chloride

catalysts. Today, practically all existing capacity makes use of acid zeolite catalysts.

World War II also led to increased demand and production of polyvinylchloride (PVC).

PVC is an excellent plastic material but lacks the ﬂexibility required for many ap-

plications. It was discovered, however, that octyl phthalates and the esters of higher

alcohols could be used to “plasticize” PVC; that is, to make it more ﬂexible and supple.

Octyl to decyl alcohols could be obtained by the “oxo” reaction (hydroformylation

reaction) of oleﬁns with carbon monoxide and hydrogen. The corresponding oleﬁns,

with one carbon less than the desired alcohols, could be recovered as by-products

from the gasoline operation simply by fractionating the desired oleﬁns out of the

oleﬁnic gasoline blend obtained by the Catalytic Condensation process. To this date,

this is still the way that these oleﬁns, from heptenes to nonenes, are produced, except

that now these oleﬁns are the main products from this operation, and the remaining

lighter and heavier fractions are blended back into the gasoline pool.

A further application that derived from World War II was the advent of synthetic deter-

gents to overcome the shortage of soap caused by the diversion of glycerin supplies to

the manufacture of explosives. Though not readily biodegradable, one of the earliest

and still one of the most active detergent ingredients is dodecylbenzene sulfonate.

This was obtained by ﬁrst alkylating dodecene with benzene according to technology

developed independently by UOP and Standard Oil of California (now Chevron).

Dodecene, or propylene tetramer, was readily produced by the oligomerization of

propylene over SPA catalyst.

Still, in the apex of its success, the Catalytic Condensation process was used mostly for

the production of gasoline fractions. During World War II, an important component in

376 CHAPTER 9

the aviation fuels used in ﬁghter airplanes of that time was hydrocodimer, described

more fully below. This together with cumene, also derived from phosphoric acid

catalyst operations, was credited with being a signiﬁcant factor in Britain’s success

in the Battle of Britain. The typical reﬁnery product was known as polymer gasoline

or “poly gasoline” even though in reality it consisted mostly of fairly simple dimers

and trimers. At one time there were as many as 300–400 Catalytic Condensation units

producing gasoline alone. Today, production of polymer gasoline is less signiﬁcant,

but the Catalytic Condensation process is still the preferred route to oleﬁn dimers and

oligomers and, despite the inroads made by zeolitic catalysts, is still widely used for

the production of cumene. Based on the current demand for SPA catalyst, it may be

estimated that about 40% is for gasoline, 40% for higher oleﬁns (plasticizer oleﬁns),

and the balance 20% for cumene applications.

SPA catalyst may contain as much as 65 wt% phosphoric acid, but not all is active. Only

the so-called “free phosphoric acid” provides catalytic activity, and this is typically

around 16–20 wt% of the catalyst as manufactured. The activity of SPA is dictated

by its equilibrium with water. Anhydrous phosphoric acid, P

, is not catalytic

for it lacks the hydroxyl groups that are necessary for activity. Too much water, on

the other hand, may cause the catalyst to get “soupy” and lose mechanical strength

or even dissolve and come out of the unit. The successful operation of a Catalytic

Condensation unit with SPA catalyst requires a close control on the water balance

going into the reactor, both in order to obtain the desired level of activity and also in

order to achieve long catalyst lives. Use of SPA catalyst in aromatic alkylation units

may present additional constraints owing to the partial solubility of the phosphoric

acid in aromatic streams; a small purge stream of phosphoric acid is usually maintained

in these units.

Catalytic condensation process for gasoline production

Figure 9.2.1 illustrates a simpliﬁed process ﬂow diagram for gasoline production

using a tubular reactor. Figure 9.2.2 illustrates a similar diagram using an adiabatic

plug ﬂow reactor. Either type can be used, except that smaller size catalyst particles

are normally used with a tubular reactor so as to minimize wall effects.

The tubular reactor system has the catalyst in the tubes and condensate or boiler

feed water is used to remove the exothermic heat of reaction and to raise steam on

the shell side. The reactants are circulated through the catalyst and the efﬂuent is

cooled and passed to a stabilizer column. An LPG (propane–butane) recycle may be

maintained to control the concentration of oleﬁns in the feed and the heat release in

the system. Usually, the oleﬁn concentration in the feed to the reactor is maintained

at not more than about 50%. The gasoline produced is removed via the bottoms of the

stabilizer.

GASOLINE COMPONENTS 377

surge drum

reactorsteam

LPG recycle

steam LPG product

gasoline

olefinic feed

stabilizer column

Figure 9.2.1. Catalytic condensation with tubular reactor.

Adiabatic plug ﬂow reactors do not have provisions for steam generation within the

reactor. Instead, an LPG interbed quench is used to control the temperature rise within

the reactor with maybe as many as ﬁve catalyst beds being arranged within the vessel.

The product is sent to a ﬂash drum in which partial condensation occurs, and the

liquid is sent to a stabilizer column. The bottom from this column is the ﬁnal gasoline

product. With this type of reactor, the concentration of oleﬁns in the feed is usually

limited to about 30%. From an energy point of view, it is less efﬁcient than a tubular

surge drum

reactor

steam

LPG recycle

LPG product

olefinic feed

flash system,

stabilizer column

gasoline

quench

Figure 9.2.2. Catalytic condensation with plug ﬂow reactor and quench.

378 CHAPTER 9

Table 9.2.1. Oligomerization reactions

Propylene/butene feed: FCC–C

fraction

Operation: without polymer recycle

Wt% yield based on feed oleﬁns

Unreacted propylene 1.3 (97.3% conv.)

Unreacted butenes 3.7 (93.1% conv.)

oleﬁns 4.4

oleﬁns 41.3

oleﬁns 24.0

oleﬁns 15.8

+ oleﬁns 9.5

reactor, but it is more economical to build. For practical purposes, the products made

by either system are almost identical.

Table 9.2.1 illustrates a typical product composition distribution based on a mixed

-C

feed from an FCC unit expressed as weight percent of the oleﬁns in the charge

stock. Feed oleﬁn conversion is about 97% for propylene and 93% for butenes, but

this may vary somewhat depending on the state of the catalyst, the feed composition,

and the operating conditions. Table 9.2.2 provides typical distillation properties of

depentanized gasoline; actual composition will vary depending on what vapor pres-

sure adjustments are desired to adapt the product to the gasoline pool requirements.

This means that the end product may also contain some pentenes and butenes if the

higher vapor pressure is acceptable for blending into the gasoline pool.

Table 9.2.2 also serves to illustrate three different typical operations of these units:

(1) polymer gasoline made from propylene only with no attempt made to produce

higher oleﬁns beyond the gasoline range; (2) polymer gasoline from a mixed C

–C

feed; and (3) polymer gasoline from a unit operating with butenes only. The speciﬁc

Table 9.2.2. Typical polymer gasoline properties

C6+ basis

Polymer type: C3 poly C3/C4 poly C4 poly

Speciﬁc gravity 0.739 0.738 0.738

Engler distillation (

◦

IBP 57 93 62

10 114 101 103

30 132 109 113

50 141 119 119

70 151 137 127

90 191 168 181

EP 218 211 216

GASOLINE COMPONENTS 379

gravities of all three products are virtually identical. Although these data are from

actual commercial operations, some differences arise from unit to unit. Also, it may

be appreciated that the mid-point (50%) and the end point (100%) for a propylene feed

are higher than for mixed feeds or butene feeds. In actual practice, these end points can

be adjusted in order to meet product speciﬁcations. Often too the amount of polymer

gasoline blended into the pool is small, so that it does not signiﬁcantly alter the overall

properties of the pool; only seldom is the product rerun in order to meet more stringent

speciﬁcations. Interestingly, the product from a mixed C

-C

feed, although having

a higher molecular weight, has a lighter boiling range than that of the propylene

product. The butene product in Table 9.2.2 is typical of a codimer operation. This

type of operation was rare in the past, but because the octane number from this mode

of operating is exceptionally high, it is currently generating considerable interest as a

replacement for the production of methyl-tert-butyl ether (MTBE) in existing MTBE

units, and will be discussed later in more detail.

Hydrogenated versus nonhydrogenated polymer gasolines

from the catalytic condensation process

Polymer gasoline as obtained is almost totally oleﬁnic. Polymer oleﬁns, however, can

be easily hydrogenated to yield a very interesting range of products. Most polymer

gasolines blended in the gasoline pools have traditionally been oleﬁnic, without any

attempt of hydrogenation. When the Catalytic Condensation process was ﬁrst de-

veloped, polymer gasoline was far better than the natural straight-run materials then

available. Gasoline characteristics have changed dramatically since the 1930s, and

while polymer gasoline used to be the best material in the leaded gasoline pool be-

cause of its excellent lead susceptibility, it is now one of the poorest in today’s unleaded

gasolines. This perception, however, can be altered signiﬁcantly if, instead of con-

sidering untreated polymer gasoline, one looks at the properties of the hydrogenated

products.

Oleﬁnic products have a high Research (R) octane number (RON), but a low Mo-

tor (M) octane number (MON). As a result, the average octane index, (R + M)/2,

of unsaturated polymer gasolines is low. The difference (RON–MON) is called the

octane sensitivity; saturated or parafﬁnic products usually do not have a high octane

sensitivity and in general, but not always, have much higher octane indices.

Hydrogenation is frequently performed at conditions that lead to the skeletal isomer-

ization of the product. Polymer gasolines from SPA catalysts are highly branched,

much more than predicted from thermodynamic equilibrium. Therefore, skeletal iso-

merization of these products would tend to move in the direction to equilibrium and,

thus, to less branched products and to lower octane numbers. Thus, in general, hydro-

genation catalysts that avoid isomerization should be selected.

380 CHAPTER 9

Table 9.2.3. Octane numbers of hydrogenated polygasolines

Polymer type

C4 poly

C3 poly C3/C4 poly (codimer)

Oleﬁnic polygasoline

RON 93 97 99

MON 82 83 84

Hydrogenated polygasoline

RON 49 70 99

MON 59 76 94

Table 9.2.3 illustrates typical octane numbers of clear oleﬁnic and hydrogenated

products. These values were measured on actual commercial samples and correspond

to the trends that can be observed with these products. It is clear that hydrogenation

does not enhance the octane numbers in all cases, so polymer gasolines from propylene

feeds are best used without hydrogenation. It is interesting to note though that the MON

of the hydrogenated material is much higher than the RON—an octane sensitivity of

minus 10 is quite unusual in gasoline components of any kind.

Polymer gasoline from propylene–butene mixtures has higher RON typically in the

96–97 range, and a MON of about 83. In the past, when tetra ethyl lead (TEL)

was added, both of these numbers increased by about 3, so leaded polymer gasoline

from propylene–butylene had excellent characteristics with (R + M)/2 of about 93.

When this gasoline is hydrogenated, both the RON and MON drop signiﬁcantly, again

rendering this product unattractive for motor fuel usage, but here too we encounter a

reverse octane sensitivity of minus six. We may add that, when used in its leaded form,

the RON and MON of the hydrogenated product increased to 86 and 90, respectively,

thus making it an acceptable blending component for the older leaded pools.

Of more interest is perhaps the operation with butene feed components. The RON

and MON of the product, sometimes called “codimer,” are up to 99 and 84, re-

spectively. Contrary to the negative experience with propylene and propylene–butene

feedstocks, the hydrogenated product from an all-butene feed operation, sometimes

called “hydrocodimer,” has increased RON and MON of 99 and 94, respectively,

which is excellent for gasoline blending in clear pools. This response was even more

striking in the past when lead was added. With the addition of 0.5 g/L TEL, the RON

and MON increased to 110–112 and to about 103, respectively. However, even in the

absence of lead, both the oleﬁnic and the saturated products have excellent blending

characteristics.

Figure 9.2.3 illustrates an approximate linear correlation between the butene content

in the feed and the RON of the hydrogenated product. The more butene there is in

GASOLINE COMPONENTS 381

100

-10

20 30 40 50 60 70 80 90 100 110

wt % propylene reacted relative to total gasoline

RON of the hydrogenated product

Figure 9.2.3. Effect of propylene content on RON of the hydrogenated polygasoline.

the feed to the catalytic oleﬁn condensation unit, the higher will be the RON of the

hydrogenated product and the better the product will be as a parafﬁnic component of

the gasoline pool. Because of the absence of oleﬁns and aromatics, this product meets

or exceeds all current environmental regulations.

It is seldom the case that the various components that integrate the composition of a

gasoline pool are used straight without blending. Normally, a motor fuel gasoline is

a complex blend of hydrocarbons generated from a number of sources, for example:

straight-run gasoline (mostly parafﬁnic with some aromatics, depending on the cut),

alkylate (parafﬁnic), isomerate (parafﬁnic), reformate (aromatic), FCC gasoline (aro-

matic), polymer gasoline (oleﬁnic or parafﬁnic), and sometimes also hydrocracked

naphtha (mostly parafﬁnic). The octane numbers of these various components do not

blend as the octane numbers of the pure fractions; a “blending octane” has to be used

instead. Thus, for example, if two gasoline materials both with an octane number of

90 are blended together, there is no assurance that the mixture will have on octane

of 90; the octane of the blend may be the same, lower, or higher, depending on the

synergism among the various components as parametrized by the blending octane.

Table 9.2.4 illustrates the blending octanes of various polymer gasoline fractions

when blended with different base stocks. The reference base stock was a mixture of

approximately 60% FCC gasoline and 40% reformate, having a clear octane of 93

RON and 80.8 MON. Blend stocks consisting of the various polymer gasolines were

added at 10 or 20 vol% levels and the blending octanes were back calculated from

the octanes measured in the resulting mixtures. It must be noted, however, that small

errors in the determination of the octane values can lead to signiﬁcant ﬂuctuations in

the estimation of the blending octanes.