Lloyd L. Handbook of Industrial Catalysts

Подождите немного. Документ загружается.

298 Chapter 7

cobalt hydridotricarbonyl, in the equilibrium reaction mixture with dicobalt oc-

tacarbonyl.

The usually accepted reaction mechanism with cobalt octacarbonyl can be

simplified as follows:

• An activation step gives an equilibrium mixture containing the catalyst,

which coordinates with the olefin forming a π-complex:

(CO)

+ H

↔ 2 HCo(CO)

+ 2 CO (7.45)

• The olefin is then inserted into the Co–H bond to form an alkyl species:

RCH=CH

+ HCo(CO)

↔ H(RCH=CH

)Co(CO)

(7.46)

H(RH=CH

)Co(CO)

↔ RCH

Co(CO)

(7.47)

• Molecular carbon monoxide is inserted into Co–C bond to form an acyl

species:

RCH

Co(CO)

+ CO ↔ RCH

COCo(CO)

(7.48)

• The final, irreversible step is hydrogenolysis of the acyl species to the al-

dehyde and catalyst regeneration:

RCH

COCo(CO)

+ H

→ RCH

CHO + HCo(CO)

(7.49)

The reaction steps are reversible and isomerization of the olefin, alkyl, or

acyl species can take place to allow the formation of isoaldehydes. The typical

4:1 product distribution of normal and isoaldehydes must be separated if the

mixture cannot be used commercially. Efforts were therefore made to increase

the proportion of useful normal aldehydes during operation. Partial success was

achieved by operating at lower temperatures with higher carbon monoxide par-

tial pressures, although this decreased conversion to aldehydes. A major prob-

lem with the cobalt catalyst was the tendency to decompose at high temperature

and to deposit metal onto the reactor walls. This led to loss of activity and low

catalyst recovery.

7.4.2. Phosphine Modified Catalysts

By the 1970s, the chemistry of the process was being investigated in an effort to

improve performance. The main objectives were as follows:

Petrochemical Catalysts 299

• Reduce by-product formation.

Catalysts derived from rhodium were found to be much more active than those

from cobalt, but rhodium was vastly more expensive than cobalt and gave a

lower yield of the normal aldehyde. Around the same time, it was shown in

work by Shell that thermally-stable complexes could be formed

by partial re-

placement of the carbonyl groups in the cobalt carbonyl catalyst by ligands such

as organophosphines. This increase in stability allowed the reaction temperature

to be increased to such an extent that the reaction pressure could be reduced with

no loss in the rate of reaction. This resulted in a lowering of both operating and

capital costs. However, a big disadvantage of the new catalyst was that the rate

of reaction decreased by about 80% and hydrogenation activity increased. This

led to lower yields, higher paraffin production, and more alcohol formation. The

process is still being used to produce 2-ethyl hexanol directly from propylene

and a range of higher-molecular weight detergent alcohols. Operating conditions

for unmodified and modified cobalt catalysts are shown in Table 7.9.

TABLE 7.9. Hydroformylation Catalyst Operation.

Variable

Catalyst: Cobalt

(Used for higher-molecular-weight oleﬁns: >C

)

Composition

(CO)

4(n-butyl)

Pressure (atm) 200–300 50–100

Temperature (

110–160 160–200

CO/H

1:1 1:2

% Metal/oleﬁn 0.1–1.0 0.6

Phosphine/metal

—

2:1

Normal/isoaldehyde 4:1 9:1

By-products (wt%) 5 1

Oleﬁn hydrogenation (vol%) <2 Up to 10

Catalyst: Rhodium

(Used for low-molecular-weight oleﬁns: up to C

)

Composition

HRh(CO)(Ph

HRh(CO)(P(m-sulfophenyl Na)

)

Pressure (atm) 12–15 Up to 50

Temperature (◦C)

~100 ~120

CO/H

1:1 1:1

% Metal/oleﬁn 0.01–0.1

Phosphine/metal 2:1 (+ excess) 3:1 (+ excess)

Normal/isoaldehyde Up to 16:1 15–20:1

By-products (wt%) Low 0.2–0.8

Oleﬁn hydrogenation (vol%) 5 —

Operates in a single or two-phase system.

• Reduce operating pressure and find a more stable and active catalyst.

• Increase the proportion of normal acetaldehyde produced.

300 Chapter 7

7.4.3. Low-Pressure Hydroformylation

Union Carbide Corporation was working on a low-pressure hydroformylation

process as early as 1967, and used a rhodium carbonyl catalyst modified with a

triphenyl phosphine ligand.

The ligand dramatically improved the normal/iso-

aldehyde ratio of the product when compared to the free metal carbonyl. Re-

duced operating temperatures and pressures, low by-product formation, and the

elimination of metallic rhodium deposition were also achieved. By 1975, a plant

was operating using this catalyst, and the process was being licensed for the

hydroformylation of propylene.

In the early version of the process, propylene and synthesis gas were circu-

lated through the catalyst solution to minimize handling of the catalyst and thus

to reduce catalyst losses. The product was recovered by using the reaction ex-

otherm to vaporize the product, which was then removed from the reactor by the

flow of the process gases. The life of the catalyst was relatively short in these

early plants, despite careful purification of the propylene and the synthesis gas.

The operating conditions are shown in Table 7.9.

The operating procedure was subsequently modified.

A single liquid phase

containing a solution of the catalyst in the form of rhodium carbonyl with a

modified ligand, together with excess ligand circulated through the reactor with

the propylene and synthesis gas. Solubilizing agents such as N-methyl pyrroli-

done or a polyalkylene glycol were added to ensure that both the catalyst and

products were completely miscible. This gave better mixing of the catalyst with

the propylene and synthesis gas thus promoting formation of a single liquid

phase. Other ligands such as modified diorganophosphates could also be used.

On leaving the reactor, the nonpolar products were separated from the cata-

lyst by adding water or methanol to split the reaction mixture into two liquid

phases. The two phases were separated by decantation, and the products thus

recovered. The catalyst was dried and purified before being returned to the reac-

tor. The lifetime of the catalyst was extended to several years, though make-up

quantities had to be added to replenish any small losses. Increasing the solubility

of the catalyst by use of more effective ligands was claimed to produce a cata-

lyst that was active for the hydroformylation of octane, dodecene and even sty-

rene, in addition to propylene and butene.

Ruhrchemie and Rhone Poulenc introduced a two-phase modification to the

basic propylene hydroformylation process in 1984.

The aqueous phase con-

tained the rhodium catalyst, and a phosphine derivative was circulated through

the reactor with the propylene and synthesis gas. The specific phosphine used

was triphenylphosphine in which the meta-position of the phenyl group had

been substituted by the sodium salt of the sulphonic acid. The products formed

an organic layer that was separated by decantation when the two immiscible

liquid phases were removed from the reactor. The aqueous layer was returned to

the reactor for further use. The catalyst is not affected by sulfur poisons, and can

Petrochemical Catalysts 301

operate for as long as 12 years provided that additional ligand is regularly sup-

plied to replace losses owing to degradation. The solubility of the propylene in

the reaction medium can become a limitation in the process.

7.4.4. Commercial Operation

Hydroformylation processes are used extensively throughout the world and the

single largest product is 2-ethyl hexanol, which is converted to dioctylphthalate,

which is used mainly as a plasticizer for polyvinylchloride. Cobalt carbonyl cat-

alysts are still important and produce large volumes of the higher-molecular-

weight aldehydes, most of which are converted to detergent alcohols. When al-

cohols are required, cobalt carbonyl catalysts, modified with phosphine ligands

are useful in that they provide a high proportion of the normal alcohol relative to

the iso-alcohol.

Rhodium catalysts with phosphine ligands are used only for propylene hy-

droformylation, to produce butyraldehyde and then 2-ethyl hexanal, but are not

used for the production of higher aldehydes. The available processes are com-

plementary and do not compete. The main factor in choosing a process is the

product required.

7.4.5. Acetic Acid

Acetic acid was originally produced by bacterial oxidation of ethanol, but from

around 1914, synthetic acetic acid was produced by the oxidation of acetalde-

hyde. Hydrocarbon oxidation processes using butane or naphtha as feedstock

were introduced in the 1950s. In a typical liquid phase oxidation process, a co-

balt acetate catalyst operating at a temperature of 175°C, and a pressure of 54

bar was used, and by-products could be recycled. Conditions could be modified

to produce methyl ethyl ketone.

The carbonylation of methanol based on Reppe chemistry was a major ad-

vance in acetic acid technology. BASF introduced this process in 1966, and used

an iodide-promoted cobalt catalyst at high pressures in the range 500–700 bar.

The reaction was thought to proceed via cobalt hydridotetracarbonyl and hydro-

gen iodide as intermediates. The hydrogen iodide was believed to react with

methanol to give methyl iodide, which underwent oxidative addition to the co-

balt carbonyl intermediate, thus forming a methyl cobalt derivative. Carbon

monoxide was inserted into the methyl cobalt bond to give an acyl cobalt spe-

cies which on hydrolysis formed acetic acid and more hydrogen iodide, thus

completing the catalytic cycle.

This process was rapidly rendered obsolete by the introduction of the low

pressure process by Monsanto

in 1971. The catalyst was similar to the Union

Carbide hydroformylation catalyst, and was based on rhodium carbonyl, pro-

moted with iodine. The operating pressure was only around 30–40 bar, with a

302 Chapter 7

lower reaction temperature of about 180°C. These mild reaction conditions led

to the formation of acetic acid at a selectivity of >99%, whereas the selectivity

of the cobalt route was only 90%, and of course a major reduction in both capital

and operating costs. Furthermore, the cobalt catalyst unlike the rhodium cata-

lysts was very sensitive to traces of hydrogen in the feed gases.

While reaction rates for cobalt catalysts depend on the methanol concentra-

tion and partial pressure of carbon monoxide, the rates with rhodium catalysts

are independent of both reactant and product concentrations. The reaction mech-

anism for cobalt catalyst in methanol carbonylation is similar to hydroformyla-

tion but is different for rhodium catalysts. The catalyst intermediate in the rho-

dium process is Rh(CO)

and has been identified at 100°C and 6 atm by infra-

red spectroscopy.

Reaction with methyl iodide in the rate-determining step

then forms a methyl rhodium complex that rapidly gives an acyl complex

(CH

CORh(CO)I

)

2−

from reaction with the high carbon monoxide content in

the reactor. The complex decomposes after reaction with water to produce acetic

acid, and with simultaneous regeneration of the catalyst:

OH + HI ↔ CH

I + H

O (7.50)

I + CO → CH

COI + H

O → CH

COOH + HI (7.51)

By 1999 more than two-thirds of worldwide acetic acid was made by the

Monsanto process using the rhodium catalyst. BP Chemicals purchased the

rights to the process in 1986 and in 1996 introduced a promoted iridium acetate

catalyst that was claimed to be more efficient with easier purification.

A com-

parison of the BASF and Monsanto processes is shown in Table 7.10.

TABLE 7.10. Methanol Carbonylation Catalyst Operation

Catalyst

Variable Cobalt Rhodium

Composition HCo(CO)

(Rh(CO)

)

−

Pressure (atm) 500–700 30–40

Temperature (

C) 230 180

Metal concentration 10

−1

−3

Promoter

Iodide Iodide

Selectivity (%) ~90 >99

By-products Methane, acetaldehyde, ethanol,

carbon dioxide

Virtually none

Reaction via methyl iodide intermediate.

Petrochemical Catalysts 303

7.4.6. Acetaldehyde

An interesting process for the production of acetaldehyde was based on the work

of F. C. Phillips, who showed that ethylene could be oxidized to acetaldehyde by

an aqueous palladium chloride solution.

The palladium chloride was reduced

to metallic palladium. During the late 1950’s, Wacker Chemie introduced a new

process for the manufacture of acetaldehyde by direct oxidation of ethylene with

air. The palladium metal was converted back to (PdCl

)

2−

by an acidic solution

of cupric chloride, which was, itself, reduced to cuprous chloride. The cupric

chloride was regenerated by reaction with air in hydrochloric acid solution. The

reaction sequence is shown in the following equations:

+ Pd(Cl

)

2−

+ H

O → CH

CHO + Pd

+ 2 HCl + 2 Cl

−

(7.52)

2 Cl

−

+ Pd

+ 2 CuCl

→ Pd(Cl

)

2−

+ 2 CuCl (7.53)

2 CuCl + ½ O

+ 2 HCl → 2 CuCl

+ H

O (7.54)

Overall: C

+ ½ O

→ CH

CHO (7.55)

In the Wacker two-stage process, the ethylene was completely oxidized

with palladium chloride/cupric chloride solution at 110°C and 10 bar, with the

stoichiometric volume of air in the reaction vessel. Acetaldehyde was recovered

from the circulating solution and the palladium oxidized with air at 100°C and

100 bar pressure in a second vessel before the catalyst solution was returned to

the reactor.

A one-stage process was then introduced by Hoechst. A mixture of almost

completely pure ethylene and oxygen (above the upper flammability limit) was

passed through the catalyst solution which contained about 200 g/liter cupric

chloride and 4 g/liter palladium chloride. At a temperature of about 120°–130°C

and a pressure of 3 bar, 35–45% of the ethylene was converted consuming all of

the oxygen. The acetaldehyde product distilled out of the reaction mixture using

the exotherm from the reaction, and the remaining ethylene was recycled. Selec-

tivity in both processes is about 95%, and the by-products are acetic acid, cro-

tonaldehyde, and methyl/ethyl chlorides, which must be removed periodically.

The process was still used to make most of the acetaldehyde required in the ear-

ly 1990s. In 1992, Catalytica suggested a possible replacement—phosphor-

molybdovanadate/polyoxoanions—for the cupric chloride used with palladium

chloride in the two-stage process, to avoid corrosion problems from hydrochlo-

ric acid.

304 Chapter 7

7.5. METATHESIS OF OLEFINS

Olefin metathesis is the disproportionation of two olefin molecules, via cleavage

of the double bonds as shown in the equation:

2 RCH=CH

→ RCH=CHR + CH

=CH

(7.56)

and

RCH=CH

+ R'CH=CH

→ RCH=CHR' + CH

=CH

(7.57)

The reaction was announced by Phillips Petroleum in 1964 and became known

as the Triolefin process because it was first used to balance the production of

olefins from naphtha steam crackers by converting propylene into ethylene and

butene-2.

As the scope of the reaction increased, the name Olefin Metathesis was in-

troduced by Goodyear, who also pioneered the use of homogeneous catalysts.

The reaction had first been recognized during experiments on the development

of a heterogeneous catalyst to replace mineral acids in alkylation reactions. Mo-

lybdenum hexacarbonyl catalyst supported on alumina produced 2-pentene

(40%) from mixed n-butenes together with propylene (51%) and hexene (9%).

Tungsten hexacarbonyl was less active than the molybdenum catalyst, and in

total contrast, it was found that chromium hexacarbonyl acted as a polymeriza-

tion catalyst.

Later experiments showed that metal oxides were more active than carbon-

yls and that the most active catalyst was actually a commercial cobalt/molybdate

hydrodesulfurization catalyst containing 3% cobalt oxide and 8% molybdenum

oxide supported on γ-alumina. Molybdenum and tungsten oxides supported on

high-surface-area silica, similar to the familiar Phillips chromium/silica polyeth-

ylene catalyst, were also active in metathesis reactions and more resistant to

poisons. The operating temperature range for a tungsten oxide/silica catalyst

was, however, significantly higher at 300°–400°C than for the cobalt/molybdate

catalyst which operated at 120°–210°C, although the catalyst volume required

was reduced by a factor of ten. Operating pressure was in the range 25–30 bar.

BP has developed an even more active rhenium oxide catalyst supported on

alumina that operates at 20°–50°C, with an equilibrium conversion higher than

60%. This follows from an observation by Phillips that alumina is a more suita-

ble support than silica.

7.5.1. Process Development

The first commercial use of the process was in Canada in 1966, to convert a

surplus of propylene into polymer-grade ethylene and 30,000 tonnes/year of

Petrochemical Catalysts 305

butene-2, at about 40% propylene conversion. The Phillips neohexene process is

used to convert di-isobutylene and ethylene to the product by a double-bond

isomerization/metathesis mechanism. All metathesis catalysts are gradually de-

activated by the formation of high-molecular-weight compounds and must be

regenerated by heating with air to 500°–600°C. The cycle time before regenera-

tion depends on the feed purity and the operating temperature.

The catalyst is installed in three reactors, two of which are on line at any

given time, while the third is being regenerated ready for use. Continuous cata-

lyst regeneration, similar to the method used in the platinum-catalyzed reform-

ing of naphtha, could be used providing that the plant size is large enough to

make the additional capital cost economic.

Metathesis is, of course, an equilib-

rium process, so that depending on the ratio of the reacting molecules and the

temperature, the nature of the products can be selected.

Following experience at an Arco plant in the United States, it was shown

that propylene production from a naphtha steam cracker can be increased by

using a metathesis unit. This has the advantage of maintaining maximum eth-

ylene production at high cracking severity. The largest naphtha steam cracker

built by the year 2000 was designed to incorporate an isomerization unit to pro-

vide butene-2 from the mixed C

stream that could produce propylene when

reacted with an equivalent proportion of ethylene. While the steam cracker on its

own was designed to produce 950,000 tonnes.year

−1

of ethylene and 540,000

metric tons/year of propylene, the use of the metathesis unit increased the level

of production to 860,000 tonnes/year of both propylene and ethylene.

The rhenium/alumina catalyst operating at 20°–50°C can produce 63% con-

version at 35°C, whereas a tungsten oxide/silica catalyst operating at 300°–

400°C gave 55% conversion at 330°C.

It is important to remove all oxygen, dienes and acetylenes from the feed to

the metathesis reactor. Furthermore, the C

stream needs to contain the mini-

mum practical level of butene-1 and isobutene to minimize the metathesis reac-

tion of the C

hydrocarbons, which results in the formation of C

and C

olefins.

The higher olefins lead to polymer formation and catalyst deactivation. An ex-

cess of ethylene normally suppresses the C

reactions. A typical steam cracker

stream, which has been subjected to selective hydrogenation to remove impu-

rities and fractionation to provide a suitable butene-2 rich raffinate-2, or the bu-

tene-2 rich effluent from an MTBE unit, can provide a suitable feed for the me-

tathesis reactor. The catalyst operating cycle with a rhenium catalyst is usually

fairly long. Regular catalyst regeneration may be necessary and the catalyst can

last for several years.

7.5.2. The Shell Higher-Olefins Process

The Shell Higher-Olefins Process (SHOP), which includes a commercial appli-

cation of metathesis, was developed to convert ethylene into a range (C

–C

)

306 Chapter 7

of straight-chain detergent alkenes containing an even number of carbon at-

oms.

Three catalytic steps are involved.

In the first stage, ethylene is oligomerized to give a Schultz-Flory distribu-

tion of linear olefins in the range C

to C

. The catalyst is a soluble nickel com-

plex formed by the reaction of bis(cyclo-octadienyl) nickel with a phenyl or

cyclohexyl-substituted phosphine derivative of the type R

PCH

COO

−

, in bu-

tanediol solution. When the ethylene is removed, the products separate as a hy-

drocarbon phase, and the catalyst solution can then be recycled.

About 30% of the α-olefin products are in the C

–C

range. These are

separated by distillation and the residual C

–C

and C

20+

fractions pass into an

isomerization reactor containing a suitable heterogeneous catalyst. An almost

equilibrium distribution of straight-chain internal olefins forms at 80°–140°C

and 3.5–17 bar. Hardly any α-olefins remain. The bottom C

fraction from the

distillation column can be changed to any boiling range depending on the prod-

ucts required. The final metathesis step uses a typical molybdena/alumina cata-

lyst to converted the isomerized internal olefins into a new range of mainly line-

ar internal olefins from which about 10–15% of C

–C

olefins can be separated

by distillation. The fraction of olefins with fewer than ten carbon atoms is recy-

cled to the metathesis reactor and the fraction with more than fifteen carbon at-

oms is recycled first to the isomerization reactor. It is also possible to add eth-

ylene to the internal olefin feed entering the metathesis reactor to increase the

yield of α-olefins. The olefins produced are converted into detergent alcohols by

hydroformylation to aldehydes, which are then hydrogenated to alcohols.

REFERENCES

1. R. F. Goldstein and A. L. Waddams, The Petroleum Chemical Industry (3rd

Edition), E. and F. N. Spon., London, 1967; P. H. Spitz, Petrochemicals:

The Rise of an Industry, Wiley, New York, 1988.

2. A. L. Waddams, Chemicals from Petroleum (An Introductory Survey), John

Murray, London, 1963/1968; R. Long, Production of Polymer/Plastic In-

termediates from Petroleum, Butterworths, London, 1967; H. Steiner, In-

troduction to Petroleum Chemicals, Pergamon, London/NewYork, 1961.

3. P. H. Spitz, Petrochemicals: The Rise of an Industry, Wiley-Interscience,

New York, 1988; K. Weissermel and H-J. Arpe, Industrial Organic Chem-

istry, VCH, Weinheim, 1993; S. Matar and L. F. Hatch, Chemistry of Petro-

leum Processes, Gulf., Houston, 1994; C. N. Satterfield, Heterogeneous Ca-

talysis in Industrial Practice (2nd Edition), Krieger, Malabar, Florida, 1996;

C. Masters, Homogeneous Transition Metal Catalysts, Chapman & Hall,

London, 1981; R. Pearce and W. R. Patterson, Catalysis and Chemical Pro-

cesses, Leonard Hill, Glasgow, 1981; Catalysis Science Technology, Ed. by

Petrochemical Catalysts 307

J. R. Anderson and M. Boudart, Vols. 1–8, Springer-Verlag, Berlin, 1981–

1987; Preparation of Catalysts—A Series, Ed. by B. Delmon, G. Poncelet,

P. A. Jacobs, and Grange, Elsevier, Amsterdam, 1976–present; G. M.

Wells, Handbook of Petrochemicals and Processes (2nd Edition), Ashgate,

Oxford, 1999.

4. H. Hock and S. Lang, Berichte 77 (1944) 257.

5. Chem. Eng. News, April 4 (1994) 23.

6. H. V. Regnier, Ann. Chim. (Phys.) 58 (1835) 307.

7. H. Kolbe, Ausfs Lehrb Organ Chem. 1 (1854) 346.

8. E. Baumann, Liebigs Ann. 163 (1872) 649.

9. Griesheim Electron German Patent 278249 (1912) (Gas phase).

10. Griesheim Electron German Patent 281584 (1913) (Liquid phase).

11. Griesheim Electron German Patent 281887 (1913).

12. Consortium für Electrochemische Industrie, British Patent 339093 (1928).

13. B. F. Goodrich US Patent 2225635 (1938).

14. M Kaufman, The History of PVC, Ch. 3, MacLaren, London (1969).

15. D. P. K. Keane, R. E. S. Stobaugh, and P. L. T. Miller Townsend, Hydro-

carbon Processing, Feb (1973) 100.

16. Burke, Miller, Chemical Week, (August 22, 1964) 93.

17. US Patent 4206180 (1980).

18. J. S. Naworski and E. S. Velez, Oxychlorination of Ethylene in Applied In-

dustrial Catalysts, Vol. 1, Ch. 9, Ed. by B. E. Leach, Academic, New York,

1983, p. 239.

19. J. W. Harpring, A. E. Van Antwerp, R. F. Sterbenz, and T. L. Kang, US

Patent 3488398 (1970).

20. N. V. EVC International, Hydrocarbon Processing, (March 1, 1997) 164.

21. German Patent 570980 (1929).

22. P. Morris, The American Synthetic Rubber Research Program, University

of Pennsylvania Press, Philadelphia, 1989, p. 7.

23. P. Morris, The American Synthetic Rubber Research Program, University

of Pennsylvania Press, Philadelphia, 1989, p. 13.

24. Rubber Reserve Co. Report on Rubber Program 1940–45, Schedule II, p.

61 (1945) Supplement No.1, p. 41/45.

25. S. V. Lebedev, British Patent 331402.

26. G. Egloff and G. Hulla, Chem. Rev. 36 (1945) 63.

27. BIOS Report 1060.

28. BIOS Report 356.

29. F. E. Frey and W. F. Huppke, Ind. Eng. Chem. 25 (1933) 54; US Patent

2098959 (1937).

30. Report on Investigations of Fuels and Lubricants Teams: US Bureau of

Mines Information Circulars (Potash), 7370, 7375 (1946); B. B. Corston,

US Patent 2375405 (1942) (Magnesia).