Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

248 Chapter 6

6.6.2.1. Bimetallic Catalysts

During the 1950s and 1960s the performance of platinum reforming catalysts

was considerably improved as operating procedures were optimized. More eco-

nomic catalysts became available as the platinum content was reduced from

more than 0.6 wt% to about 0.35 wt% and the sulfur content of feed naphtha

was considerably reduced. Higher purity alumina supports, with better activity

and lower levels of sodium and iron could be synthesized reproducibly. An op-

timum amount of chloride added to the support limited coke formation.

More signiﬁcant changes, however, began in 1968 when Chevron intro-

duced its bimetallic catalyst containing platinum and rhenium. The main ad-

vantage of the new catalyst was its relative stability, which extended operation

before regeneration was necessary. The separation of small platinum crystallites

by rhenium made the catalyst less sensitive to carbon deposition. In fact, bime-

tallic catalysts doubled the period before regeneration was necessary. This im-

provement allowed operation at lower pressures with increased yields. Improved

dehydrocyclization and isomerization selectivity also increased the volume of

aromatics produced and the octane number of the gasoline. The level of hy-

drocracking decreased and the production of light gases was reduced.

The success of the platinum/rhenium catalyst was evident, as more than

80% of all catalyst replacement charges were bimetallic by 1972. By then 30%

of all reformers were using the rhenium catalyst.

Acceptance continued in all

reforming processes so that more than two-thirds of installed catalyst capacity

was bimetallic by 1979.

Catalyst improvements also continued during this pe-

riod, and the original Chevron catalyst, Grade A, was replaced several times up

to Grade F.

Most of the changes were probably related to metals concentration but the

original bimetallic catalyst contained equal weight percentages of platinum and

rhenium and is now known as a balanced catalyst. Since the late 1980s, rhenium

catalysts have contained about twice as much rhenium as platinum and are

known as skewed catalysts. The skewed catalysts are even more stable, particu-

larly when loaded to reactors at high bulk density, and have extended the operat-

ing cycle even further before regeneration is required. This is partly due to the

ability of the catalyst to operate with increased coke content.

A disadvantage of platinum/rhenium catalysts is their sensitivity to sulfur

poisoning. This requires that naphtha feed must be carefully hydrotreated before

use.

The very few plants still using platinum catalysts (with no rhenium) are

those without the necessary hydrotreating facilities.

It is necessary to pre-sulﬁde new or regenerated catalysts with a suitable

sulfur compound prior to use when working with a low sulfur-content feed, to

reduce excessive hydrogenolysis activity of the catalyst. Presulﬁding causes

deactivation of the super-active edges on the metal sites that are responsible for

Refinery Catalysts 249

the bulk of the hydrogenolysis activity which would otherwise lead to premature

coke formation and the hot spots that cause a reduction in reforming activity.

The addition of 100–200 ppm of dimethyl sulﬁde to the recirculating gas raises

the sulfur level of the catalyst to about 0.25% by weight. During subsequent

operation, only the sulfur on the superactive edges remains in the catalyst, the

rest being quickly removed. Subsequent operation of the catalyst is not affected

and the presence of rhenium sulﬁdes may limit the sintering of the platinum

crystallites.

The information in Table 6.20 shows how the use of rhenium has improved

operation and indicates the coke content before regeneration is necessary. The

coke content shown is for the third bed in a semi-regenerative unit (the ﬁrst and

second beds contain considerably less coke.) It has been found that even 1 ppm

of sulfur in the naphtha feed reduces the cycle time with skewed and balanced

catalysts by about 25% and 35%, respectively.

The ﬁrst CCR unit in 1971 used a platinum/rhenium catalyst, but in 1973

UOP introduced a catalyst in which the platinum was promoted by tin.

The

new catalyst not only resulted in lower coke levels at the low operating pressure

but was easily regenerated throughout its long life. Tin reacts with the surface

hydroxyl groups on the γ-alumina support. This gives better stability and in-

creases the amount of coke deposited on the support rather than on the platinum.

Satisfactory operation, however, depends on a good distribution of both plati-

num and tin. γ-Alumina is stabilized by the tin, which reacts with the surface

hydroxyl groups. The first UOP catalytic reformer plant at 1000 tonnes per ca-

pacity is shown in Figure 6.6.

Iridium has also been used as a promoter in platinum reforming catalysts

since it was introduced by Exxon in 1971–1972.

This was reported by Chemi-

cal Week in one of their Catalyst Reviews, when they noted a sharp increase in

the reported sales of iridium.

Catalysts containing iridium have not been wide-

ly used except in plants designed by Exxon and Amoco. Extremely long periods

between regenerations have been reported and may be due to the more efﬁcient

Table 6.20. Improved Operation with Platinum/Rhenium Catalysts.

Catalyst %Coke before regeneration Relative cycle time

Pt/Al

10 1

Pt/Re/Al

(balanced) 15 x3

Pt/Re/Al

(skewed)

high density

20 x6

Sulfur in feed (ppm)

Relative cycle time

Balanced Skewed

0 1 2

0.5 0.9 1.7

1.0 0.7 1.5

250

Figure

roduc

hydro

e wel

tempe

sinteri

6.6.2.2

Mono

suppor

solutio

acid pl

in sol

The a

strong

with

Chapter 6

6.6. The first

d with permissi

enation and r

distributed a

ature during t

. Catalyst Pre

etallic refor

with chlorop

usually con

ays an import

tion, by chlor

id strength u

enough to dis

ydroxyl grou

OP catalytic re

n from UOP,

moval of cok

bimetallic cl

e oxidation a

aration

ing catalysts

latinic acid in

ains the exac

nt role in the

ding the alu

ed is proport

solve the alu

s on the alu

orming plant at

LC.

precursors. T

sters and grea

d reduction o

are prepared

a dilute solut

amount of p

catalyst prepa

ina and ensur

onal to the a

ina. Chloropl

ina surface.

1000 tonne pe

he active met

care must be

f catalyst at s

y impregnati

on of hydroc

atinum requir

ration by kee

ng good plati

umina surfac

atinate and c

The replace

day capacity.

ls are claime

taken to limit

-up to prev

g the γ-alum

loric acid.

d. Hydrochl

ing the platin

um distributi

area but is

loride ions re

ent of hydro

ric

Refinery Catalysts 251

groups with chloride gives neighboring hydroxyl groups increased acidity with-

out increasing the total number of acid sites.

Impregnated alumina is carefully drained and dried at about 150

C to re-

move residual water and then oxidized in dry air at about 500

C. Drying and

oxidation can replace some of the chlorine atoms from the adsorbed chloroplati-

nate ions on active sites with oxygen. At the same time, the chloride content of

the alumina is decreased depending on the partial pressure of water vapor pre-

sent in the air stream. Loss of chloride should, therefore, be carefully controlled

by ensuring a low drying temperature and using dried air during oxidation to

maintain the required chlorine content in the finished catalyst. The properties of

the γ-alumina support are extremely important and the gradual improvements

made to the catalytic reforming process up to about 1960 came mainly from an

understanding of the catalyst properties and the use of purer, more stable alumi-

nas.

Bimetallic catalysts have the same basic characteristics as the platinum/γ-

alumina catalysts, with good dispersion of small platinum crystallites. The most

widely used bimetallic catalysts are the platinum/rhenium formulations. The

preparation of these catalysts is usually by impregnation of the γ-alumina sup-

port with a solution of chloroplatinic acid and perrhenic acid, or its ammonium

salt in hydrochloric acid.

Successive impregnations with chloroplatinic and

perrhenic acid solutions have also been used. The catalysts are then carefully

dried, oxidized, and reduced. The volatility of the rhenium complexes or oxides

requires careful control of temperature to avoid metal loss by vaporisation and to

prevent alloy formation during reduction.

Platinum/iridium catalysts are made by impregnating γ-alumina with chlo-

roplatinic and chloroiridic acids. Heating to temperatures about 375

C produces

the highly dispersed bimetallic clusters that form the catalyst. Great care is re-

quired to achieve the necessary metal dispersion because iridium oxide is vola-

tile, and it has even been suggested that a third metal such as chromium is re-

quired to promote support-metal interaction.

Platinum/tin catalysts are made by impregnating the support with tin chlo-

ride and calcining at 500

C, before impregnation with chloroplatinic acid. The

catalyst is then dried, oxidized, and reduced. Temperature control is necessary to

avoid reduction of the tin chloride to metallic tin, which would then alloy with

the platinum. Tin oxide combines with acid sites on the support and increases

resistance to deactivation. Typical catalyst formulations are shown in Table

6.19.

6.6.3. Catalyst Regeneration

When the yield from the reformer falls to an uneconomic level the catalyst must

be regenerated.

This is the point where the C

yield falls from the initial 76–

77% to about 73% and the purity of recovered hydrogen declines. Regeneration

252 Chapter 6

removes the coke deposited on the active sites, restores the catalyst acidity, and

redisperses the active metal into smaller crystallites. Different quantities of coke

are deposited in each bed of a semi-regenerative reformer. Beds 1, 2, and 3 con-

tain about 1–3 wt%, 5–15 wt%, and 8–20 wt%, respectively. Great care must be

taken to maintain a steady temperature proﬁle in each bed during the regenera-

tion procedure.

Catalyst is regenerated in the continuous catalytic reduction process using

the same basic procedures. About 200 lb h

-1

were reduced in the ﬁrst CCR units

in 1971 compared with 4000 lb h

-1

in modern plants. These quantities are

equivalent to regenerating the whole charge of catalyst in the reformer—a total

of 12 and 150 times a year, respectively.

6.6.3.1. Carbon Burn

Coke is removed by burning in an air/nitrogen gas stream, containing up to 0.5

vol% of oxygen at a space velocity of about 5000 h

-1

. Bed temperatures during

the ﬁrst stage should always be less than 350

C. If necessary the oxygen content

can be decreased to control hot spots. These conditions are maintained until the

outlet temperature falls to the same level as the inlet temperature. At this point

the oxygen content can be increased gradually and the bed temperature slowly

raised to 500

C. The ﬁnal carbon content of the catalyst is normally less than 0.2

wt%.

6.6.3.2. Oxychlorination

After carbon has been removed from the catalyst, the chlorine content usually

falls to 0.6–0.8 wt%. The loss of chlorine must, therefore, be compensated by

passing dry air that contains a suitable chlorine compound such as carbon tetra-

chloride over the catalyst at 450

C. Sufﬁcient chlorine must be used to restore

the original 0.9–1.3 wt% chlorine level. About 70% of the total chlorine in the

airstream is retained by the catalyst. A high space velocity of up to 5000 h

-1

can

be used.

6.6.3.3. Platinum Re-Dispersal

Platinum that has sintered during plant operation must be redispersed into the

small crystallites required for active catalysis. This is carried out by heating in a

dry air flow at 500

C. When the air leaving the reactor contains less than 5 ppm

water the air should be purged from the catalyst bed until the oxygen content is

less than 1 ppm. It is thought that the redispersion treatment breaks down the

larger platinum particles by producing volatile oxide or oxychloride complexes.

These then react with surface hydroxyl groups to form smaller crystallites and

restore activity.

Refinery Catalysts 253

6.6.3.4. Catalyst Reduction

Before use the new or regenerated catalyst is reduced in dry hydrogen at 500 C

at a space velocity of 400 h

-1

for about 2 h. This converts the platinum oxides

into finely divided metal particles.

6.6.4. Catalyst Life

Most of the initial activity reforming catalysts can be recovered during regen-

eration, although it is inevitable that some activity is lost during the procedure

and catalysts are eventually replaced. The catalyst in semi-regenerative reform-

ers, which can be operated for 9 to 12 months before being regenerated, will not

be changed for about 8 to 10 years. Catalyst in cyclic reformers, which are de-

signed for regeneration after only a few days, has a shorter life of only 4 to 5

years. These estimates are only approximate and may be less depending on the

feeds and operating conditions used.

Although catalyst life in the continuous catalytic regeneration reformers is

probably up to seven years, it is usually progressively replaced gradually as

catalyst activity declines. It may also be necessary from time to time to remove

catalyst from the reactor to clear any accumulated dust. This is usually recog-

nized by increased pressure drop or maldistribution of ﬂow giving hot spots. The

catalyst can be replaced for further use after sieving. Any heel catalyst, contain-

ing up to 50% carbon, which has not been properly regenerated in isolated dead

spots of the bed, can be separated and discarded.

6.7. OCTANE BOOSTING

Shape-selective reactions have been extensively studied since zeolites were ﬁrst

used in catalytic crackers during 1967

and Mobil has introduced several oc-

tane-boosting processes since 1968.

Gasoline composition and octane number

were reviewed when the Clean Air Act was passed in 1970. This mandated the

phased removal of tetraethyl lead from gasoline as catalytic converters (see

Chapter 11) were introduced to treat automobile exhaust gas. Octane boosting

was the ﬁrst in a series of measures that led to reformulated gasoline and im-

proved exhaust emissions standards.

6.7.1. Selectoforming

Shape-selective zeolite catalysts were ﬁrst used in the Selectoforming process by

Mobil in 1968. Hydrogen-exchanged natural erionite, containing some nickel,

254 Chapter 6

was used to crack and isomerize low-octane normal paraffins in reformate to

increase the concentration of isoparaffins and boost the octane number. The

nickel minimized carbon formation but did not saturate aromatics. The erionite

was later replaced by synthetic zeolite T, which has the twelve-membered of-

fretite ring containing about 4% of eight-ring erionite and effectively prevents

isoparafﬁns from entering the pores.

6.7.2. M-Forming

By about 1970 the beneﬁts of ZSM-5 zeolite in cracking low-octane normal

parafﬁns in FCC units were well known. By adding 1% of ZSM-5 as a promoter

to reforming catalysts it was possible to increase the octane number of reformate

at the expense of some liquid yield.

In the 1970’s, the octane number of the product from the catalytic reforming

of light naphthas was only RON 84. This was too low to satisfy the changing

market, and the M-forming process was developed by Mobil to use ZSM-5 zeo-

lites containing nickel as catalyst. Octane number was increased to RON 93 be-

cause the pores of ZSM-5 zeolite crack single-branched paraffins at about

315

C. Higher-octane products are formed by oligomerization, aromatization,

and alkylation reactions. There is an increase in yield and less gas formation

compared with the Selectoforming process.

6.8. AROMATICS PRODUCTION

Reformers can also be operated to maximize the production of aromatics. When

benzene is the required product, a narrow fraction boiling in the range 60–90

is used under severe operating conditions. If the feed is changed to a fraction

boiling in the range 110–140

C under similar high severity conditions, the main

product is a mixture of toluene and mixed xylenes. Normally however, the main

object of reforming is to maximize the octane rating of the product.

6.8.1. Aromatics Process

L-zeolite, potassium exchanged and also containing some platinum, is an active

catalyst for the conversion of n-hexane to aromatic compounds. L-zeolite is a

wide-pore zeolite, with a silica/alumina ratio of six and a unidimensional pore

structure.

The commercial catalysts described and used in the Chevron Aromax pro-

cess were partly exchanged with barium ions and contained platinum (0.6–0.8%)

as small particles, corresponding to an approximate formula

Pt/Ba

Direct dehydrocyclization of paraffins is an alternative to

isomerization and reforming to increase octane number of light straight-run

Refinery Catalysts 255

naphthas. These zeolite catalysts are, however, more easily poisoned by sulfur

compounds because they are based on platinum and feedstock must be desulfu-

rized.

6.8.2. Cyclar Process

BP and UOP developed a process to convert liquid petroleum gas (LPG), a mix-

ture of propane and butane, into aromatics.

A plant using this process began

operating in 1989 at the BP reﬁnery in Grangemouth, Scotland, and a second

followed in 1999 operated by Saudi Basic Industries (SABIC) in Saudi Arabia.

The catalyst is based on a ZSM-5 zeolite containing gallium. The yield of

aromatics is about 60%, with more than 90% as benzene, toluene, with some

xylene and relatively small amounts of C

aromatics. Up to 2000 standard cubic

feet of hydrogen per barrel of feed is produced. Reaction proceeds at 425

C by

dehydrogenation of the paraffins followed by oligomerization and cyclization. A

four-bed UOP continuous catalytic regeneration reactor is used with interbed

heating of the reaction mixture. Aromatics are separated from the remaining

aliphatic hydrocarbons by distillation.

Japanese research on converting propane to aromatics in the presence of

on ZSM-5 catalyst loaded with zinc showed that equilibrium was improved

as hydrogen was removed by the reverse water gas shift reaction:

+ CO

O + CO (6.6)

6.8.3. M2-Forming Process

The M2-Forming process introduced by Mobil uses a ZSM-5 zeolite to convert

both virgin naphtha (C

–110

C) and oleﬁnic gasolines (C

–110

C) from FCC

and thermal crackers into aromatic compounds. It operates at a higher tempera-

ture than the M-Forming process and at about 550

C the selectivity to aromatics

approaches 100%.

6.9. CATALYTIC DEWAXING

The viscosity of high-boiling fuels and lube oils can be reduced by shape-

selective cracking of long-chain paraffins in the presence of hydrogen. Several

zeolites introduced during the 1990s have been used: H-mordenite containing

platinum by BP; ferrierite containing palladium by Shell; ZSM-5 by Mobil. The

exchange of zeolite with the precious metal promoted the hydrogenation of coke

precursors at the high operating pressures,

and extended catalyst lifetimes.

256 Chapter 6

Typical operating conditions are: reactor temperature, 300–350

C; pressure,

20– 130 atm; hydrogen partial pressure, 15–100 atm.

6.10. ISOMERIZATION

The catalytic reforming of straight-run naphtha to gasoline is usually restricted

to feeds in the boiling range 100–180

C. This is because the equilibrium conver-

sion of the lower-boiling normal C

–C

parafﬁns under typical operating condi-

tions is too low to obtain a reasonable improvement in higher-octane-number

branched isomers.

A separate process to isomerize the straight-run C

–C

cut is used to com-

plement catalytic reforming and produce gasoline with good octane qualities

over the whole boiling range. The change in octane number as a range of C

–C

hydrocarbons is isomerized, is shown in Table 6.21. To achieve the most suita-

ble product, however, a high–activity catalyst must operate at the lowest possi-

ble temperature. This improves equilibrium conversion and increases the propor-

tion of dimethyl isomers rather than single-branch isomers.

6.10.1. Isomerization Catalysts

Isomerization processes developed slowly because of the low demand for higher

octane numbers and operational problems with aluminum chloride, the ﬁrst cata-

lyst to be developed. Despite high activity at 115–120

C, aluminum chloride has

the major disadvantages of the formation of sludges and acid corrosion of

equipment.

TABLE 6.21. Isomerization of C

–C

Hydrocarbons.

Wt% Feed Product Octane number

n-Pentane 33 15 62

i-Pentane 22 40 93

n-Hexane 20 6 31

2-Methyl pentane 12 14 74

3-Methyl pentane 10 7 74

2,2-Methyl butane 1 11 96

2,3-Methyl butane 2 5 105

–C

Hydrocarbons

RON 70 82

After n-C

recycle 83–85

After n-C

,n-C

recycle 87–90

Note: Extract n-hydrocarbons with a shape-selected zeolite.

Refinery Catalysts 257

The introduction of dual-function catalysts for naphtha reforming and the

demand for high-octane gasoline led to further interest in isomerization. The

platinum/alumina (chlorided) catalysts were a success despite the resulting low-

er conversion to high-octane products from the need to operate at higher temper-

atures. The Shell Hysomer process, which used a 5A-zeolite to separate low-

octane parafﬁns from the product allowed operators to recycle unconverted feed

and achieve almost 100% conversion.

A further important development was the use of hydrogen-exchanged mor-

denite zeolite as the catalyst support. The zeolite was more stable and water tol-

erant than the chlorided alumina and did not need chloride addition during oper-

ation. The zeolite catalyst had a lower activity than that of alumina, but the need

to operate at a slightly higher operating temperature was acceptable since the

normal paraffins could be recycled. Other zeolites, such as β-zeolite, have also

been used as the acid support.

Operating conditions for the different catalysts are given in Table 6.22. Hy-

drogen is usually added to minimize coke formation but is not used in the reac-

tions taking place.

6.10.2. Reaction Mechanism

The isomerization of normal C

–C

parafﬁns proceeds by dehydrogenation on

the platinum surface followed by isomerization of the oleﬁn formed on the acid

support. The isomer is then hydrogenated on the platinum to give the higher-

octane branched paraffin.

Platinum in the dual-function catalyst is important in reducing the steady-

state oleﬁn concentration, otherwise polymerization and cracking would lead to

extensive coke formation. Some coke, however, does deposit slowly and the

catalyst has to be regenerated at intervals. Mordenite supports are more stable

than chlorided aluminas and can operate for longer at more than 60% conversion

and about 97% selectivity. Catalysts are reported to operate for as long as seven

years before being replaced.

TABLE 6.22. Isomerization Operating Conditions.

Conditions Pt/Cl-Al

Pt/H-mordenite

Temperature (

C) 150 250

Pressure (atm) 15 30

Hydrogen/hydrocarbon 2:1 2:1

LHSV (h

−1

)

1.5 2.5

Yield (%) > 98 >98

Feed (RON ~70) (150

C) (250

No recycle RON 84-85 RON 82-83

With recycle RON 90-91 RON 89-90