Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

268 Chapter 7

+ KOH → CH

=CHCl + KCl + H

O (7.9)

The first reference to the compound as vinyl chloride was by Kolbe in 1854.

Baumann later described polyvinyl chloride in 1872 when he gave a good sum-

mary of its properties.

Around 1900, when industry began to take an interest in acetylene chemis-

try, Fritz Klatte produced vinyl chloride at 180

C from acetylene and hydrogen

chloride:

CH≡CH + HCl → CH

=CHCl (7.10)

Patents for his process were issued to Griesheim–Electron in 1912/1913.

9,10

They described a mercuric chloride catalyst supported on coke or pumice. The

early catalysts described in 1912 could only operate for short periods because

mercury compounds sublime at reaction temperature. Although the 1913 patent

described reaction in aqueous solution with the mercuric chloride catalyst, this

process was never used industrially. A further patent in 1913 claimed that the

process only took place in the presence of the mercury catalyst that accelerated

the reaction.

The volatility of mercury compounds led the Consortium für Electrochem-

ische Industrie to patent a catalyst in 1928 based on mercury or bismuth com-

pounds supported on active carbon or silica gel,

which retained its activity for

100 hours. B. F. Goodrich also claimed that complex salts, such as HgCl

.2KCl

or HgCl

.BaCl

, were not only less volatile but more active.

None of the later

catalysts was as efficient as mercuric chloride on charcoal in terms of the space

time yield, despite their longer lives. The early catalysts gave better results at

lower temperatures with good temperature control. Developments prior to 1940

were reviewed by Kaufmann

and showed that vinyl chloride could be produced

from either ethylene or acetylene. Early processes, however, still used acetylene,

which was more readily available, although at that time demand was not great.

In the 1920s, Union Carbide recovered by-product ethylene dichloride from

its ethylene oxide plant for conversion to vinyl chloride.

Vinylite, a relatively

soft copolymer of vinyl chloride and vinyl acetate, was used to produce gramo-

phone records and other useful items. Plasticizers were then developed to pro-

vide the softer grades of polyvinyl chloride that were widely used during the

World War II. By 1939, the acetylene-based vinyl chloride process had been

developed commercially and was used to supply PVC for the war effort. During

the period from 1935 to 1940 the only industrial producers were Germany, with

about 110 tons a month, and the United States, where, in 1936, production

reached 950 tons a year. The German catalysts and processes were described in

various BIOS, FIAT, and CIOS reports and other publications after the end of

Petrochemical Catalysts 269

TABLE 7.3. Production of Vinyl Chloride 1939–1945.

Operating conditions

Feed Acetylene (C

) and HCl

Reactor Tubes 50–80 mm diameter × 3 m long (vessel 2 m diameter)

Temperature 120

C new to 200

C old

Pressure ~1 atm

Catalyst life

10 months

Catalyst composition

A: Catalyst (I.G. Farben) 10% HgCl

/charcoal—1/4-in. cubes

Yield (on C

) 96–98%

(on HCl) 80–90%

Conversion 100%

B: Catalyst (Wacker)

30% BaCl

; 1% HgCl

/charcoal

HCl available from cracking carbon tetrachloride to trichloroethylene over BaCl2 catalyst.

the war; operation from 1939 to 1945 is shown in Table 7.3.

No similar details

were released for the US process. The same mercuric chloride catalyst supported

on charcoal was still being used, acetylene was still the preferred feedstock, and

the use of PVC was rapidly expanding.

The demand for larger vinyl chloride plants coincided with the development

of hydrocarbon steam crackers. Ethylene reacts directly with chlorine forming

ethylene dichloride, which was decomposed thermally to produce vinyl chloride.

The byproduct hydrogen chloride could then be used to produce more vinyl

chloride as a feedstock for the acetylene process. This was described as the bal-

anced ethylene–acetylene process:

=CH

+ Cl

→ CH

Cl.CH

Cl (direct addition) (7.11)

Cl.CH

Cl → CH

=CHCl + HCl (thermal reaction and HCl recycle) (7.12)

CH≡CH + HCl → CH

=CHCl (mercuric chloride catalyst) (7.13)

The ethylene–acetylene balanced process could, therefore, solve the prob-

lem of hydrogen chloride disposal, although only half of the vinyl chloride was

produced from ethylene. The manufacture of acetylene is a very energy-

intensive process, and compared to ethylene, acetylene is very expensive to pro-

duce. Ethylene would, therefore always be the preferred feedstock for the pro-

duction of petrochemicals, provided that suitable technology was available.

Development of the oxychlorination process, which used a copper chloride

catalyst, improved the situation because it only required ethylene as feedstock.

The oxychlorination of ethylene for the manufacture of vinyl chloride is a three-

stage process. The first stage is the addition of chlorine gas to ethylene to pro-

duce 1-2 dichloroethane. The second stage is the cracking of the dichloroethane

270 Chapter 7

to vinyl chloride and hydrogen chloride. The third stage is the oxychlorination

reaction itself, in which hydrogen chloride from the cracking stage is reacted

with air, or in later processes, oxygen, to produce more dichloroethane, which is

then recycled to the second stage.

=CH

+ Cl

→ CH

Cl.CH

Cl (direct addition) (7.14)

Cl.CH

Cl → CH

=CHCl + HCl (thermal reaction and HCl recycle) (7.15)

=CH

+ ½ O

+ 2 HCl → CH

Cl.CH

Cl + H

O (copper chloride catalyst)

(7.16)

Cl.CH

Cl → CH

=CHCl + HCl (thermal reaction and HCl recycle) (7.17)

The cupric chloride catalyst was promoted with potassium chloride and

supported on activated alumina. It was related to the Deacon catalyst, which was

developed for chlorine production in 1887.

The first oxychlorination plant was built by B. F. Goodrich in 1964

and

produced 30 million lb.year

−1

of vinyl chloride in a fluid bed. At about the same

time the Stauffer Chemical Company built a unit using three tube-cooled ox-

ychlorination reactors in series. This was the beginning of the end for routes

based on acetylene.

7.1.2.1. The Oxychlorination Reaction

The need to recover hydrogen chloride from the ethylene dichloride cracking

reaction introduced during the early 1960s led to a renewed interest in the Dea-

con process, although the low equilibrium concentration of chlorine in hydrogen

chloride/air mixtures at a reasonable temperature made economic recovery im-

possible.

It was found, however, that the Deacon catalysts could chlorinate ethylene

with a mixture of hydrogen chloride and air. The Deacon reaction did not pro-

duce molecular chlorine but chlorinated ethylene directly as chlorine species

formed on the catalyst surface. Low equilibrium conversion to chlorine and its

slow removal from the catalyst surface were no longer limitations and complete

chlorination of the ethylene was achieved at temperatures in the range 210

–

240

C. Free chlorine was never found even with low ethylene concentrations.

7.1.2.2. Oxychlorination Catalyst

Suitable catalysts are similar to the original Deacon catalyst. The copper chlo-

ride is supported on medium-surface-area alumina with pore sizes in the range

80–600 Å.

While copper chloride is more active than other metals it is also

Petrochemical Catalysts 271

volatile at reaction temperature. Potassium chloride is added to decrease the loss

of copper and reduce the formation of by-product ethyl chloride, even though it

also reduces the overall activity. The copper chloride content of the catalyst

must be limited because an excess would lead to catalyst caking during reaction,

since the complexes KCuCl

and K

CuCl

have naturally low melting points.

The eutectic mixture with copper chloride melts at 150°C. Catalysts may also

contain other alkali and rare earth chlorides and promoters that help to control

the reaction or inhibit side reactions.

The reaction mechanism was originally thought to be as follows:

+ 2 CuCl

→ CH

Cl.CH

Cl + Cu

(7.18)

+ ½ O

→ CuO.CuCl

(7.19)

CuO.CuCl

+ 2 HCl → 2 CuCl

+ H

O (7.20)

The overall reaction is:

+ 2HCl + 1/2O

→CH

Cl.CH

Cl + H

O (7.21)

Catalysts are usually prepared on a large scale by impregnating a suitable

alumina support with aqueous solutions of cupric chloride and potassium chlo-

ride. Other alkali metals or rare earth chlorides have been used as promoters or

to inhibit by-product formation.

7.1.2.3. Catalyst Operation

The overall reaction is very exothermic, and temperature control is extremely

important in vinyl chloride production to maintain a satisfactory selectivity. This

is relatively easy in fluidized beds, which is probably the most widely used pro-

cess, but more difficult when fixed bed tubular reactors are used. High bed tem-

peratures result in the dehydrochlorination of the ethylene dichloride to give

vinyl chloride, and this can lead to the formation of other chlorinated products.

Increased oxidation of the ethylene feed to carbon dioxide also occurs as the

temperature increases. Coke formation and copper loss by sublimation also in-

crease with higher bed temperatures.

In tubular reactors the proportions and concentration of the reactants can be

changed to control the reaction or the tube diameter design can be decreased.

The most common ways to control the reaction are to dilute the catalyst with

inert balls or to vary the copper chloride content of the catalyst. There is a rapid

increase in catalyst temperature at the top of the tubes where most of the exo-

thermic reaction takes place giving a significant hot spot. This can be controlled

to a certain extent by splitting the total air addition between the three beds. Good

272 Chapter 7

control at the inlet to the first bed is necessary in any case to ensure that the ex-

plosive limit is not exceeded. The addition of fused alumina, graphite, or other

inert diluents to the catalyst can limit the temperature rise, although there are

practical problems in mixing catalyst with inert materials.

Stauffer

described a process consisting of a single reactor containing up to

four different layers of a catalyst containing 8.5% cupric chloride. Each succes-

sive layer was diluted with an inert substance to control the exothermical reac-

tion. The first layer only contained 7% of catalyst, the others containing 15%,

then 40% and the final layer contained 100% catalyst. However, better control

with diluted catalyst was obtained in a series of several separate reactors. A top

layer of special catalyst to initiate or strike the reaction was used in the first re-

actor. An even better degree of control was obtained with three different catalyst

compositions containing varying amounts of cupric chloride. Typically, catalysts

containing 6%, 10%, and finally 18% cupric chloride, promoted with 3%, 3%,

and then 2% potassium chloride respectively, each packed in layers, were the

optimum. Beds 1 and 2 contained 40% of both the 6% copper chloride and 10%

copper chloride, respectively, on top of the 18% copper chloride catalyst. Bed 3

was completely filled with the more concentrated catalyst.

Catalyst life in tube-cooled reactors depends on the operating conditions,

but in reactor 1 the catalyst is usually changed after one year. Life increases to

one and a half years and three years in reactors 2 and 3, respectively, where

conditions are less severe and the intensity of the temperature hot spots de-

crease. There are usually hot spots at the position in tubes where the catalyst

type changes. Up to 50% of the air can be added to the first reactor and must be

carefully controlled to avoid the explosive limit. The balance of air is added to

the second and third reactors to minimize the hot spot and to maintain conver-

sion. Oxygen can replace air when convenient because this reduces the gas to be

vented from the system, simplifies the incineration of noxious products and

makes environmental control easier.

In fluidized bed reactors the preheated gas mixture enters the base of the

vessel and temperature is controlled either by cooling coils or a bundle of cool-

ing tubes in the fluid bed. The oxychlorination reaction takes place at a relative-

ly low temperature, around 220–225

C, and at a low pressure of about 2 bar. The

catalyst support consists of microspheroidal γ-alumina particles, with a median

particle size around 50–60 microns. The particles are impregnated with an aque-

ous mixture of cupric and potassium chlorides, to give a copper content around

12%. One of the main process problems with oxychlorination

is a possible

tendency for the catalyst particles to cake or to stick to the cooling bundles with-

in the reactor. This results in poor temperature control, local overheating, and a

loss in selectivity. The surface area of the alumina is in the range 150–250

−

, with 94% of the particles greater than 6 μm and 24% greater than 80

μm.

Fresh catalyst is added to replace fines lost from the bed. An excess of air

is added to maximize hydrochloric acid conversion and unreacted ethylene is

Petrochemical Catalysts 273

recycled. Yields of 98% on hydrochloric acid and 96% on ethylene can be ob-

tained. By-products are chloral, 1,1,2-trichloroethane, chloroform, 1,2-dichloro-

ethylene, and ethyl chloride. In modern processes pure oxygen is used instead of

air. This enables much easier recovery of product and catalyst fines, also ena-

bling a reduction in the volume of vent gas by at least 95% and a reduction in

both energy and capital costs.

Since the oxychlorination process was intro-

duced plant capacities have increased from about 10,000 to more than 600,000

tonnes a year.

7.2. SYNTHETIC RUBBER FROM BUTADIENE AND STYRENE

In the 1930s, a wide range of hydrogenation and dehydrogenation processes was

already developed by companies in the petroleum industry to increase gasoline

yield. Olefins had been used to provide polygasoline and alkylates, as well as

many other petrochemical intermediates. A form of synthetic rubber had been

made in Germany since the 1914–1918 war and in 1929 two I. G. Farben chem-

ists from Leverkusen patented a variety they named Buna-S.

It was a co-

polymer of butadiene containing 25% styrene and, as a result of many improve-

ments, is still being used. However, at the time it was difficult to process, ther-

mally unstable during use, and was expensive to produce. Other synthetic rub-

bers soon followed. Neoprene, produced from chlorobutadiene and originally

named Duprene, was developed by DuPont and used in several applications.

G. Farben also produced Buna-N, replacing the styrene in Buna-S with acryloni-

trile. At the time Buna-N was unsuccessful, although it did eventually lead to the

present large-scale production of acrylonitrile. By 1940, when the

Reserve Company was formed in the USA, two of the most significant rubber

intermediates were butadiene and styrene. The potential wartime shortage of

natural rubber then led the Rubber Reserve Company to make plans for four

plants to produce synthetic rubber, which was known as GR-S rubber (Govern-

ment Rubber-Styrene), and by November 1943 a total of 15 plants were operat-

ing.

Production of GR-S rubber increased from about 4000 tonnes in 1942 to

almost 1 million tonnes in 1946. Production fell as the war ended but increased

again to more than 600,000 tonnes a year from 1951 to 1953, while the Korean

War was in progress.

Production on such a large scale led to an increasing demand for the rela-

tively new petrochemicals butadiene and styrene. When wartime rubber produc-

tion was being planned most commercial butadiene was produced as a by-

product from the high-temperature steam cracking of naphtha or catalytic crack-

ing of oil fractions. The quantity, however, was very small. It was therefore ap-

parent that, in addition to making more of the quickie butadiene, other methods

would have to be developed. Two sources were potentially available. One possi-

bility was a process that had been used in Russia to produce butadiene from

274 Chapter 7

butane and n-butene by processes that were still being developed.

Nine dehydrogenation plants were eventually authorized by the Rubber Re-

serve Company to make 330,000 tonnes year

−

of butadiene.

Two were based

on butane and seven on butene. During construction of these plants it was also

decided to make about 220,000 tonnes.year

−1

of butadiene by the Russian Lebe-

dev process based on ethanol:

→ C

+ 2H

(7.22)

→ C

+ H

(7.23)

Little is known about the Lebedev process, which was developed in Russia

in 1927 during the investigation of synthetic rubber production. It was a one-

step process that converted ethanol to butadiene by simultaneous dehydrogena-

tion, dehydration, and condensation stages:

2 C

OH → C

+ 2 H

O + H

(7.24)

The catalyst used was nominally a mixture of 75% magnesia and 25% silica

or siliceous earth, which are well known as dehydrogenation and dehydration

catalysts respectively. Up to 3% chromium oxide was included to inhibit the

formation of magnesium silicate.

Carbon formed on the catalyst during operation at 270°–300°C, so that re-

generation was required every 12–16 h. The catalyst was soon deactivated and

had to be changed after only 30 days. A maximum butadiene yield of 60% was

achieved.

As with most catalysts, the original recipe must have been modified

as revealed in BIOS reports on the German use of the catalyst.

Whereas the

composition was reported to be 100 parts magnesium oxide with 15 parts kaolin

and 1–3 parts chromia, the analysis of an authentic sample was somewhat dif-

ferent.

The catalyst was made by wet mixing the components, extruding the

mixture, and calcining at 500°C for 2–3 h before use.

Commercial dehydrogena-

tion processes are usually limited to propane and butane feeds because they op-

erate below the thermal steam cracking temperature. Nevertheless, a high reac-

tion temperature is still required to achieve a reasonable conversion close to the

thermodynamic equilibrium. Frey and Huppke indicated that chromia catalysts

gave a good yield of olefins from the corresponding alkanes.

Later, Burgin and

Groll patented chromia/alumina catalysts containing 6–40% Cr

and it was

found that the optimum chromium content to dehydrogenate propane was lower

than that required to dehydrogenate n-butane. Promoters such as potash or mag-

nesia were used to reduce carbon decomposition during operation.

n-alcohol and the second was to produce butadiene by the dehydrogenation of

Petrochemical Catalysts 275

7.2.1. Butadiene from Butane

All butane dehydrogenation processes used chromium catalysts supported on γ-

alumina. Catalysts were generally made by impregnating a preformed support

with chromic acid solution that also contained the magnesium or potassium

promoter. Processes operated at a high temperature and low butane partial pres-

sure to compensate for the high hydrogen content of reaction gases and to im-

prove the equilibrium conversion. The Houdry or Catadiene process was used in

only two of the Rubber Reserve Company plants and was the only process able

to produce butadiene directly from butane. The chromia/alumina catalyst, con-

taining 15–20% chromia, was used in several adiabatic reactors operating in

parallel. Despite being diluted with a dense inert material the catalyst was deac-

tivated by carbon deposition after about 7–15 min of operation at 620°–650°C

and a reduced pressure of 3 psig. Liquid butane space velocity was in the range

1–2 h

−1

. The reactor containing deactivated catalyst was taken out of service and

replaced by a spare reactor to give continuous operation. Deactivated catalyst

was then regenerated in air and diluted with an inert gas at 500°C before reuse.

Overall selectivity to butadiene was only 55–60%, with conversions in the

range 30–40%. The feed gas was preheated using the heat evolved from the

combustion of carbon that had been laid down on the surface of the catalyst dur-

ing operation. Thus, carbon lay-down and catalyst regeneration became an inte-

gral part of the process.

Butane was also dehydrogenated, usually to butene, by the Phillips and

UOP processes. Both processes used the same sort of chromia/alumina catalysts,

loaded in tubular reactors, but were able to operate at atmospheric pressure be-

cause less hydrogen was produced during the reaction. In some cases the butene

produced was subsequently dehydrogenated to butadiene in a separate unit.

ICI operated a plant to produce isobutene from isobutane in a UOP plant.

Yields of 75–80% isobutene were obtained at 30–40% conversion in a tubular

reactor and residual isobutane was recycled.

I. G. Farben also produced butene by butane dehydrogenation.

A moving

catalyst bed system was used in a tubular reactor. The total catalyst charge was

1.5 tonnes with a residence time of 4 h in the tubes. Yields of 85% at 20–25%

conversion were obtained at a liquid space velocity of 2 h

−1

and 620°–650°C

operating temperature. This was an impressive result for a new reactor design

that has now been developed as the continuous catalyst regeneration process and

is widely used in refineries.

7.2.2. Butadiene from Butenes

Butene dehydrogenation was studied extensively in Russia, Germany, and the

United States in the 1930s, and several promising processes were selected for

276 Chapter 7

the Rubber Reserve Company program. In all cases a steam ratio of up to 20:1

was used during operation. The steam diluent provided several benefits by:

• Decreasing the partial pressure of hydrocarbon and increasing equilibri-

um conversion at atmospheric pressure.

• Acting as a heat carrier so that the feed was not directly heated to the re-

action temperature.

• Decreasing coke deposition to extend the operating cycle.

For these reasons the catalyst could be used in adiabatic beds rather than in

tubular reactors. The inlet temperature lay in the range 600°–675°C and, more

practically from a process point of view, the process was operated at atmospher-

ic pressure. The use of less water-sensitive catalysts also meant that regeneration

was possible with steam rather than using air at high-temperature.

A catalyst developed by the Standard Oil Development Company and first

used in 1940 was used in six of the Government butadiene plants.

The compo-

sition of catalyst 1707 is shown in Table 7.4.

It was shown that iron oxide promoted dehydrogenation activity while pot-

ash reduced the extent of carbon deposition during operation and also assisted in

carbon removal during regeneration. The catalyst gave 85% selectivity at 20%

conversion.

Phillips produced a bauxite catalyst containing 5% barium hydroxide that

was later impregnated with potash. Natural bauxite always contains a significant

amount of iron impurities, and there was probably an unexpected beneficial ef-

fect from the iron impurity. The Shell Development Company introduced an

iron catalyst promoted with chromia and containing potash, which by this time

was recognized as essential to reduce the effects of carbon formation.

The

compositions of the Shell catalyst 105 and the later Shell 205, which gradually

replaced 1707 in wartime plants, are also shown in Table 7.4. It seems likely

that the potash content of Shell 105 was gradually increased to 7% K

TABLE 7.4. Early Butene Dehydrogenation Catalysts.

Wt% Standard

Oil 1707

Shell 105

(1940s)

Shell 205

(1950s)

Dow

(1950s)

Magnesium oxide 72.4

⎯

—

Ferric oxide 18.4 93 62.5

Ni(PO

)

Copper oxide 4.6

Chromium oxide — 5 2.2 2

Potassium 4.6 (K

O) 2 KOH) 33.3 (K

)

Operating temperature (

C) 625/650 650 625/650

Steam/butane 14:1 18:1 20:1

Conversion % 28–35 37 30–50

Selectivity %

70–75 76 86–90

Lebedev catalyst 40–50% MgO; 10–12% SiO

; 2–3% Al

; ~3% Cr

; balance CO

Petrochemical Catalysts 277

1954 the Shell catalyst 205, containing 33% potassium carbonate, became the

preferred catalyst for butadiene production.

The Shell catalysts were made by dry mixing powdered iron oxide, chromic

acid, and potassium carbonate, mixing with water, extruding before drying, and

then calcining at 800°–950°C. The high-temperature calcinations was critical

and later work indicated that it optimized surface area and pore size as the iron

oxide was converted to α-hematite.

Dow introduced a calcium/nickel phosphate catalyst (see Table 7.4), which

was used in several butene dehydrogenation units.

It gave a better yield at a

higher steam ratio than the other catalysts but needed to be regenerated with air

at intervals of about 15 min to remove the carbon deposits which collected in the

catalyst bed, not on the catalyst itself. The operating life was about seven

months before replacement, which was much shorter than that of the other cata-

lysts.

Butadiene production from these dehydrogenation processes did not contin-

ue beyond the 1960s, as butadiene from steam cracking of naphtha became

available. Later, however, from about 1990, as butadiene and butene shortages

developed, several commercial processes were revived.

7.2.2.1. Oxidative Dehydrogenation

The addition of oxygen to the butene and steam reaction mixture improves con-

version and selectivity during the dehydrogenation reaction by removing the

hydrogen as water:

+ ½ O

→ C

+ H

O (7.25)

Not only does this improve the equilibrium conversion but the exothermic oxi-

dation reaction also supplies heat to balance the endothermic dehydrogenation

reaction. At 550°–600°C more than 90% selectivity is possible at conversions in

the range 65–80% with steam mole ratios up to 12. Catalysts in the Petrotex

(Mobay) process are zinc/chromium and magnesium/chromium ferrites.

7.2.3. Propylene from Propane

The dehydrogenation processes used for the production of butadiene from n-

butane and n-butenes during the development of GR-S rubber were modified in

the 1990s for the dehydrogenation of propane to propylene. This compensated

for the short supply of steam cracked propylene used to produce polypropylene.

The new processes can also be used for the dehydrogenation of other paraffins.

The old Houdry process has been developed by UCI/Lummus to become

the Catofin process while Phillips and UOP have introduced their Star and Ole-

flex processes. Linde in Germany has also introduced a process. A range of dif-