Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

218

Figure

Exxon

TAB

Year

1938

1946

1957

1978

1986

1993

1998

Chapter 6

6.2. Outline flo

obil Research

E 6.7. Import

sheet of alkyl

nd Engineerin

nce of the H

% H

tion plant. Rep

Company.

/HF Alkylatio

oduced with pe

Processes in t

% HF

—

mission from

e United States

Total productio

(million bpd)

Small

0.17

0.27

1.00

1.05

1.07

1.10

compressor

recycle

refrigerant

recycle acid

makeup

isobutane

alkylate

product

butane

product

propane

product

olefin

feed

recycle isobutane

reactor

settler

deisobutanizer

depropanizer

debutanizer

Refinery Catalysts 219

TABLE 6.8. Sulfuric Acid and Hydrogen Fluoride Alkylation.

6.3.1. Liquid Acid Processes

All commercial plants operate with a high proportion of isobutane in the cir-

culating gas and conversion depends on the residence time of the gas in the reac-

tor. Typical feeds include oleﬁns derived from catalytic cracking and, more re-

cently, the refnates from MTBE or TAME units. Typical operating conditions

are summarized in Table 6.8. Some of the acid catalyst must be removed on a

continuous basis for recovery as it becomes diluted with inactive hydrocarbons.

Sulfuric acid must be reused in other processes or recovered off-site for resale, a

process which costs more than fresh acid. Hydrogen fluoride is usually recov-

ered on site and reused.

6.3.2. The Mechanism of Alkylation with an Acid Catalyst

The generally accepted mechanism for the alkylation reaction is as follows:

• Initiation. Butylene combines with a hydrogen ion to form a secondary

butyl carbonium ion:

CH=CHCH

+ H

→ CH

CHCH

(6.2)

• Isomerism. The secondary butyl carbenium ion can then either isomerize

to form a tertiary butyl carbonium ion:

CHCH

→ CH

(6.3)

Temperature (◦C) 5–10 21–38

Isobutane/oleﬁn feed ratio (vol) 5–8 10–14

Oleﬁn space velocity (vol oleﬁn/vol acid) 0.2–0.3

Contact time minutes 20–30 10–15

Acid strength (%) 90–92 85–95

Acid in emulsion (%) 50–60 25–60

Acid used (lb barrel

−1

)

13–30 0.1–0.3

Alkylate from C

feed (%):

Trimethyl pentane 73

Dimethyl hexane 10

Heavy alkylate C

Light ends C

−

220 Chapter 6

or react with an isobutane molecule to give a butane molecule and a ter-

tiary butyl carbonium ion:

| |

CHCH

+ CH

CH → CH

+ CH

(6.4)

| |

• Addition. The tertiary butyl carbonium ion reacts with a butylene mole-

cule to form branched C

trimethyl pentene carbonium ions:

| |

+ CH

=CH

→ CH

C CH

CHCH

(6.5)

| |

• Propagation. The C

carbonium ion reacts with an isobutane molecule by

hydrogen transfer to give the C

paraffin and to generate a tertiary C

carbonium ion that continues the sequence of alkylation reactions.

A number of side reactions can also take place such as:

• The decomposition of a C

carbonium ion to the C

oleﬁn and a hydro-

gen ion. This is likely at either a low isobutane/oleﬁn ratio or low acid

strength.

• Reaction of a C

carbonium ion with another C

oleﬁn molecule to form

a C

carbonium ion. This can react with an isobutane molecule to give a

parafﬁn, but in the presence of excess oleﬁn, further chain growth can

take place.

6.3.3. Liquid Acid Operating Conditions

During operation hydrocarbons are emulsiﬁed in acid, ensuring that the oleﬁn is

not in excess, and reacts quickly. Operating pressure is adjusted depending on

the boiling point of the hydrocarbon mixture and the need to dissolve the hydro-

carbons. The space velocity of the mixed oleﬁns and the catalyst is regulated,

depending on the octane level of the product needed and operating temperature.

Octane levels increase if the acid strength is low, which means that more spent

acid has to be removed from the system. In sulfuric acid alkylation the optimum

acid strength decreases with propylene, butylene, and pentylene feeds.

Refinery Catalysts 221

6.3.4. Processes Using Solid-State Acid Catalysts

Several solid acid catalysts were tested in the laboratory for the alkylation of

ethylene with isobutane following the introduction of the sulfuric acid alkylation

process.

These were mostly derived from aluminum chloride or boron triﬂuo-

ride and were never used in the full-scale production of alkylates.

Operators have always hoped that a successful solid acid catalyst would be

discovered because of the problems and hazards involved in handling large

quantities of sulfuric and hydroﬂuoric acids. This situation reached a climax in

1990, when California reﬁneries were faced with the possibility of having to

close down units using HF catalysts. From about that time, potential solid-state

acid catalysts were investigated by catalyst producers and several have been

tested in pilot units.

In the meantime, although modifications were introduced

to HF units and additives were used to reduce the volatility of the acid, the pro-

cess has been less popular than the sulfuric acid process.

The short-term alter-

native of converting HF units to the use of sulfuric acid was much more expen-

sive than the use of additives.

Suitable solid acids have been difﬁcult to ﬁnd. Those used for benzene al-

kylation are not sufficiently acidic for economic operation and, also, the oleﬁn

can polymerize on the catalyst surface. This blocks the active sites and prevents

adequate hydrogen transfer. Frequent regeneration of the catalyst is not econom-

ic. The use of antimony pentaﬂuoride slurried with the hydrocarbons is one of

the options that has been investigated in recent years.

The best solution appears to be the use of an almost insoluble liquid catalyst

held within the pores of a suitable inert support. Supported liquid catalysts are

well known and can be used with a continuous catalytic regeneration system

similar to that developed for catalytic reforming processes. Haldor Topsøe has

successfully tested triﬂuoromethane sulfonic acid in this way since 1993 with a

variety of oleﬁn feeds.

No formal regeneration was necessary apart from peri-

odic removal of some catalyst for reimpregnation and the recovery of dissolved

acid from the alkylate. Both catalyst and support are, therefore, recirculated. The

small quantity of polymeric by-products formed (acid soluble oil) appears to be

less than that formed in the sulfuric acid process, but slightly more than in the

HF process.

6.4. HYDROTREATING

Sulfur impurities in products manufactured from crude oil products are undesir-

able because hydrogen sulﬁde, sulfur dioxide, etc., formed during product use.

For many years it was possible to obtain acceptable quality gasoline and kero-

sene by selecting low-sulfur, or sweet, crude oils. Sour crude oils contain dis-

222 Chapter 6

solved hydrogen sulﬁde, mercaptans, organic sulﬁdes, thiophenes, and elemental

sulfur in varying amounts. These could be sweetened by a number of chemical

processes. High-sulfur crude oils were more difﬁcult to desulfurize and the

chemical and solvent extraction processes were combined with or replaced by

relatively cheap and more efﬁcient catalytic processes that could also remove

gum-forming compounds from cracked gasolines.

For a short time from 1946 bauxite or fuller’s earth was used without the

addition of hydrogen.

It was found that sulﬁdes and mercaptans reacted with

impurities in the bauxite and that this together with some mild cracking of hy-

drocarbons produced sufﬁcient hydrogen to hydrogenate thiophenes. It was soon

realized that the catalytic desulfurization was actually a mild, selective hydro-

genation process that did not saturate aromatics.

During the 1950s, co-

balt/molybdate catalysts supported on bauxite or Fuller’s earth were used. These

were similar to the catalysts developed for coal hydrogenation and which were

also used to desulfurize steam reforming feeds. The new catalysts were most

effective when hydrogen was added to the feed. This also had the effect of re-

ducing the deposition of carbon, and allowed for longer operating cycles before

regeneration was necessary. More effective cobalt/molybdate catalysts were

developed using γ-alumina as support. The activation step for the catalyst in-

volved the formation of metal sulﬁdes, and when the catalyst was pre-sulﬁded

before use, it was found that light distillates, kerosene and even crude oils could

be treated effectively with these catalysts.

Operating conditions depended on

the boiling range of the fraction being treated. Catalyst temperature was usually

limited to about 400

C in order to avoid excessive carbon deposition while total

pressure was increased from 300–500 psig for low-boiling distillates and up to

700–1000 psig for higher-boiling or cracked feeds. Liquid space velocity was

usually up to 8 h

-1

, with a hydrogenn/oil ratio of about 1000 scf of hydrogen per

barrel of feed for low-sulfur distillates. Lower space velocities, in the range from

0.5–3 h

-1

, with hydrogen/oil ratios up to 10,000 scf per barrel, needed to be used

for higher-boiling residues. In the hydrotreating of heavy feeds, more carbon

was deposited by thermal cracking than in the hydrotreating of lighter feeds.

Catalyst regeneration was required after operation for less than 24h.

The use of hydrodesulfurization became more widespread as catalytic naph-

tha reforming processes were introduced. The operation of platinum catalysts

needed an increasingly strict sulfur speciﬁcation for the naphtha, and as a bonus,

the cheap by-product hydrogen from the reforming process could be used to

hydrotreat other reﬁnery product streams.

By the early 1960s, hydrotreating capacity in the United States had in-

creased to 2.5 million bpd, with catalyst lifetimes averaging about 5 years. The

use of hydrotreating was extended to kerosene, gas oil, and vacuum gas oils as

government regulations on sulfur emissions became more stringent and as better

cobalt molybdate catalysts became available. By the late 1970s, when atmos-

pheric and vacuum residues were also being desulfurized, US hydrotreating ca-

Refinery Catalysts 223

pacity had increased to about eight million bpd. By 1999 US capacity exceeded

10 million bpd, with more than 35 million bpd being treated worldwide.

6.4.1. What Is Hydrotreating?

The term hydrodesulfurization was used to describe processes that removed sul-

fur compounds from crude oil fractions by reaction with hydrogen. As the pro-

cesses evolved to include nitrogen and oxygen removal, together with the hy-

drogenation of aromatics and oleﬁns, the group of processes became known as

hydrotreating. Hydrotreatment simply results in the conversion of organic sul-

fur, nitrogen, and oxygen compounds to hydrocarbons and hydrogen sulﬁde,

ammonia, or water, respectively. At the same time oleﬁns and aromatics may be

converted to saturated hydrocarbons without any cracking of the hydrocarbons.

When high-boiling crude oil fractions are hydrotreated under more severe condi-

tions a proportion of the heavy molecules may crack as impurities are removed.

For this reason it is now common to refer to processes that crack a relatively

small proportion of the light feeds while sulfur or other impurities are removed

as hydroreﬁning.

Deliberate reduction in the molecular weight of heavy feeds was introduced

with the hydrocracking process (Section 6.5). This operates at high pressure, and

heavy gas oils and residual fractions are converted to gasoline or other desirable

products. Obviously there can be an overlap in deﬁning processes where some

precracking is an advantage to reﬁneries. When an overall description is re-

quired, the term hydroprocessing is often used.

6.4.2. Hydrotreating Processes

In the early hydrotreating processes, sulfur compounds were removed from the

light hydrocarbon fractions used in gasoline by hydrogenation over co-

balt/molybdate catalysts to produce hydrogen sulﬁde and a saturated hydrocar-

bon. Around the same time, it was found that nickel/molybdate catalysts were

more active for the hydrogenation of nitrogen compounds to ammonia and a

hydrocarbon while also giving some saturation of oleﬁns and aromatics.

In modern reﬁneries both cobalt/molybdate and nickel/molybdate catalysts

are now widely used in the puriﬁcation of various crude oil fractions. These in-

clude:

• Straight-run naphthas, used as feedstock for catalytic reforming and

steam reforming processes. They must contain less than 1 ppm of sulfur

and nitrogen to avoid poisoning platinum or nickel catalysts.

• Cracked gasoline, to hydrogenate undesirable sulfur and nitrogen com-

pounds as well as oleﬁns.

224 Chapter 6

• Middle distillates such as diesel fuel, kerosene, jet fuel, domestic heating

oil, and other gas oils, to remove sulfur for environmental reasons. Hy-

drotreating is also used to increase the smoke point or cetane number by

hydrogenating aromatic components.

• Vacuum gas oils, used as catalytic cracker or hydrocracker feeds, to re-

move sulfur, nitrogen, and metal impurities.

• Atmospheric and vacuum residues, to remove as much sulfur as possible

to provide low-sulfur fuel oils. It is also used to hydrogenate asphaltenes

and porphyrins to reduce both Conradson carbon and metal contents.

As the boiling point and the speciﬁc gravity of the fractions increase, more

severe hydrotreating operating conditions are needed. A lower space velocity

and more extensive hydrogen recycle are needed to limit deactivation of the cat-

alyst by deposition of coke. The catalyst must be regenerated after shorter inter-

vals and discarded more often than when using light fractions.

A very brief summary of typical operating conditions is shown in Table 6.9

although these can vary signiﬁcantly for different crude oils and blends contain-

ing cracked fractions.

6.4.2.1. Catalyst Production and Operation

Hydrotreating catalysts are usually produced by impregnating preformed γ-

alumina supports with aqueous solutions of ammonium molybdate. Particles are

then dried and calcined. Molybdenum oxide forms a uniform surface layer on

the alumina by reaction with the surface hydroxyl groups. It is important that the

surface area of the alumina is consistent with the amount of molybdenum oxide

required for the hydrotreating catalyst speciﬁcation (typically 1 wt% MoO

≡ 12

-1

area).

Table 6.9. Typical Operating Conditions.

Fraction

Naphtha Gas oil Vacuum gas oil Atmospheric residue

Boiling Range (

C) 66–200 240–380 350–560 560+

Liquid space velocity (h

−1

)

6–10 2–4 1–2 0.2–0.5

Hydrogen/oil ratio 60 240 350 7500

Temperature (

280–320 340–360 360–400 400–420

Pressure (atm) 10–25 20–40 30–60 80–100

Regeneration interval 1–3 years 1–2 years 1 year 0.5–1 year

Sulfur content feed (%) 0.05–0.15 1–2 2–3 4–5

Sulfur content product (%) < 1 ppm 0.1–0.3 0.2–0.4 0.6–0.8

Nitrogen removal (%) 99 45 40–45 45

Refinery Catalysts 225

The molybdenum-treated alumina is then impregnated with either cobalt or

nickel nitrate solutions and again dried and calcined. Several reactions can take

place during the calcination step. At temperatures in the range 400–500

C, co-

balt and nickel nitrates decompose to form oxides on the molybdena/alumina

catalyst surface. The deposition of cobalt and nickel in contact with the molyb-

denum on the catalyst surface is important in forming active sites. As the tem-

peratures are increased both cobalt and nickel oxides react to form aluminates

close to the alumina surface. Cobalt aluminate (CoAl

) produces the charac-

teristic blue color of cobalt/molybdate catalysts. The nickel and cobalt ions mi-

grate into the alumina particles as the temperature is increased to 650–700

forming bulk aluminates. These are not active catalyst precursors and the tem-

perature of calcinations must be carefully limited to avoid their formation.

A typical support can be prepared by mixing boehmite [AlO(OH)] and wa-

ter to form a smooth paste, often adding a small volume of dilute nitric acid to

peptize the alumina to give increased strength. The paste is extruded to form

small-diameter, relatively long, spaghetti-like cylinders, rings, and other shapes,

or the boehmite can be granulated to produce spheres. The supports are dried

and then calcined at up to 600

C to produce γ-alumina. It is common practice to

add up to 1% of silica, often as montmorillonite or bentonite, to stabilize the γ-

alumina at high operating temperatures and to act as a lubricant during extru-

sion. Addition of phosphoric acid appears to produce phosphate groups on the

alumina surface that inhibit coke formation as well as to improve molybdenum

dispersion.

Finally, small amounts of potash can be added to reduce surface

acidity and to control coke formation. Typical catalyst specifications are given

in Table 6.10 and some catalyst are shown in Figure 6.3.

When using heavy feeds, particularly residues, the catalyst pores at the top

of the bed soon become blocked with metals and high molecular-weight hydro-

carbons present in the feed. This reduces catalyst activity and the catalyst must

be regenerated. In an attempt to overcome the problem and to increase the cycle

time, so-called bimodal catalysts were introduced containing both large and

small pores. Small particles of cellulose, combustible polymers, or carbon were

mixed with the boehmite and burned out during the ﬁnal calcination step of the

formation of the γ-alumina support. By using different techniques it was possi-

ble to generate pores shaped like bottles which could contain large molecules,

and large pores which gave access to small subsidiary pores. These special sup-

ports extended the operating cycle before the catalyst had to be regenerated or

discarded.

6.4.2.2. Catalyst Handling

The catalyst is supported by a layer of relatively large inert spheres, and must be

packed carefully into the reactor. Dense-loading procedures to increase the

packing density of the catalyst are now available that use a rotating disk at the

226

end of

dates,

An ad

ﬂat as

izontal

resulti

catalys

pletely

uids t

lane

differe

idea is

ids an

shapes

distrib

Chapter 6

a hose to dis

nd trilobe sh

antage of the

oading proce

y and this im

g in better c

t is packed

ﬂat, but usua

en run prefer

erformance.

t catalyst typ

to have large

impurities in

and activities

layer of cera

tion of the fe

ure 6.3. Ty

produced wit

psøe A/S.

ribute the cat

pes can be in

dense-loadin

ds. Catalyst p

roves the dist

ntrol of temp

anually into

ly has rather

ntially down

In the design

es in a two-o

iameter parti

the feed. Diff

can also be us

ic balls is pla

d and to catc

ical hydrotreat

permission

lyst. The pac

reased by 8

technique is

rticles, which

ibution of liq

rature distrib

converter, t

convex or c

ome channels

or modern hy

three-

ed re

les at the top

rent catalysts

ed on top of t

some of the s

ng catalysts.

rom Haldor

ing density

, 12%, and 2

hat the cataly

are very long

id residue th

tion and con

e surface lay

ncave orienta

, resulting in

drotreating pr

ctor system

of the reactor

having differ

e catalyst be

lid particles i

f spheres, ext

%, respectiv

t layer is alw

and thin, lie h

oughout the b

version. Whe

r is rarely co

ion. Process l

oor catalyst

cesses, up to

ay be used.

to cope with s

nt compositio

to help impr

troduced by t

ly.

iq-

ol-

Refinery Catalysts 227

Table 6.10. Typical Hydrotreating Catalyst Speciﬁcation.

ﬂow of the reactants. In older plants, wire mesh baskets were often placed on top

of the catalyst bed and ﬁlled with pebbles. These were part of the vessel design

to catch solids and to distribute liquid feed, and were known as trash baskets.

Although still in use, better ﬁltering of the feed and graded catalyst layers with

the larger catalyst particles at the top of the beds have largely replaced the use of

baskets.

6.4.2.3. Activating the Catalyst

Hydrotreating catalyst is activated by presulﬁding with hydrogen sulﬁde. This is

achieved by circulating sulfur compounds in a stream of hydrogen at tempera-

tures too low to reduce the oxides to metals. Small molybdenum disulﬁde plate-

lets (truncated hexagons) that incorporate cobalt or nickel ions at the plate edges

are formed to produce the active CoMoS or NiMoS sites. The surface area of the

alumina support is critical to achieve the appropriate distribution of molyb-

denum, and the optimum ratio of nickel or cobalt to molybdenum is approxi-

mately 1:4. In the case of cobalt molybdate, if the atomic ratio exceeds 30:1,

cobalt sulﬁde (Co

) crystals form on the molybdenum disulﬁde in preference

Cobalt molybdate Nickel molybdate

Composition

Cobalt oxide % 4–5 —

Nickel oxide % — 4–6

Molybdenum oxide % 12–20 15–20

Phosphorus % Up to 2 Up to 2

Alumina Balance Balance

Impurities SO

, Na

O, Fe

1%, 0.06%, 0.6%,

respectively

1%, 0.06%, 0.6%,

respectively

Properties

Loss on ignition 500◦C (wt%) 2–4 2–4

Bulk density (kg liter

−1

) 0.5–0.7 0.5–0.7

Surface area (m

−1

) 200–300 200–300

Pore volume (ml g

−1

) 0.5–0.7 0.5–0.7

Mean pore radius (nm) 4.5–5.0 4.5–5.0

Shapes Size (mm)

Extrudates 1.6, 1.3, 0.8 diameter × 3–5 long

Rings 5 × 5, 3 × 3

Trilobes 1.6, 1.3 × 5

Quadrilobes 1.6, 1.3 × 5

Used catalyst Before regeneration After regeneration

Sulfur (wt%) Up to 10 < 0.5

Carbon (wt%) 8–12 0.1–0.3

Loss on ignition (wt%) 17–20 < 0.5