Jackson S.D., Hargreaves J.S.J. Metal Oxide Catalysis

Подождите немного. Документ загружается.

730 18 Titanium Silicalite-1

produced glycolic acid. Further oxidation yielded the α - dicarbonyl derivatives and,

at prolonged contact time, C

–

C bond oxidative cleavage.

The oxidation of tetrahydrofuran, tetrahydropyran and dihydropyran produced

γ - butyrolactone and δ - valerolactone in fairly good yields. Linear ethers, conversely,

produced exclusively the corresponding acids, presumably through the in situ

hydrolysis of intermediate esters [116] .

The Baeyer – Villiger rearrangement of cyclohexanone and acetophenone with

TS - 1/H

proved to be poorly selective [117] . Notably, Ti - β and Sn - β have different

chemoselectivities in the oxidation of unsaturated ketones, leading selectively to

corresponding epoxides and lactones, respectively [118] . The different oxidation

pathways were attributed to the preferential adsorption of hydrogen peroxide on

Ti - sites and of the carbonyl group on Sn - sites.

18.6

Ammoximation of Carbonyl Compounds

The word ammoximation deﬁ nes the reaction of a carbonyl compound with

ammonia, under oxidative conditions, to yield the corresponding oxime (Equation

18.12 ). Early homogeneous and heterogeneous catalysts, for example heteropoly

compounds and amorphous Ti/SiO

, had too poor selectivity to envisage any

application. The prospects changed when Rofﬁ a and coworkers discovered that

TS - 1 was an excellent ammoximation catalyst with hydrogen peroxide as oxidant

[119] . As Table 18.12 shows, TS - 1 and amorphous Ti/SiO

, albeit with similar Ti

contents, lead to completely different results: to quantitative yields for the former

and to a mixture of products, with only a minor amount of oxime, for the latter.

(18.12)

In contrast to the stability shown in other oxidations, TS - 1 undergoes a slow but

irreversible deactivation. The siliceous matrix of the catalyst, in fact, slowly dis-

solves in the basic medium, while Ti, released from the framework, aggregates to

Table 18.12 Ammoximation of cyclohexanone on different catalysts.

Cat. Ti (wt%) Conv. C

O (%) Sel. on C

O (%) Yield on H

(%)

TS - 1 1.5 99.9 98.2 93.2

Ti/SiO

1.5 49.3 9.3 4.4

S - 1 – 59.4 0.5 0.3

a Solvent, t - butanol/water; T , 80 ° C; molar ratios, cyclohexanone:

: H

1.0 : 2.0 : 1.1; t.o.s. 1.5 h.

form TiO

particles. Accordingly, the activity decreases with time on - stream,

approaching that of conventional catalysts at prolonged times [120] . Major side

reactions consist of the oxidation of ammonia to N

, N

O, nitrites and nitrates and

in the decomposition of H

to molecular O

. Condensation reactions of cyclo-

hexanone in the basic medium and consecutive oxidation of the oxime produce

minor amounts of organic by - products. However, the losses of the oxidant are

generally moderate, for example < 7% under the conditions of Table 18.12 , and the

decay of the catalyst is not so fast as to preclude the development of an industrially

viable process.

The reaction of ammoximation is generally applicable and provides an efﬁ cient

route to many other oximes in addition to the industrially relevant cyclohexanone

oxime. Acetone, butanone, acetophenone, C

–

cyclic ketones and methyl - sub-

stituted cyclohexanones produced the corresponding oximes with high conversion

and selectivity. Even the ammoximation of cyclododecanone and 4 - butylcyclohexa-

none, unable to diffuse in TS - 1, occurred with high yields [121 – 123] . Other ammo-

ximation catalysts are carbon - supported TS - 1, TS - 2, Ti - MOR and Ti - MWW, with

conversions and selectivities close to those of TS - 1 [124 – 128] .

Two ammoximation mechanisms were initially proposed, one envisaging the

formation of cyclohexanone imine (Equation 18.13 ) and the other that of hydrox-

ylamine (Equation 18.14 ) as key intermediates.

(18.13)

(18.14)

The high yields obtained from ketones unable to diffuse inside the pores of TS - 1,

namely cyclododecanone and 4 - butylcyclohexanone, strongly supports the second

mechanism [121, 123] . The imine intermediate, if present, could only play a sec-

ondary role in the overall balance of the ammoximation process. In this regard,

Mantegazza and others showed that TS - 1, under the reaction conditions, catalyzes

the oxidation of ammonia to hydroxylamine with yields greater than 60% [129] .

The subsequent condensation with the carbonyl group is a fast reaction and does

not require catalysis by TS - 1. It occurs in the external medium with bulky ketones,

while with smaller ones the intra - porous volume also could be involved. In the

absence of a carbonyl group acceptor, hydroxylamine undergoes consecutive oxida-

tion to N

, N

O, nitrites and nitrates. On these grounds, the overall ammoximation

process could be better deﬁ ned as an oximation reaction with in situ production

of hydroxylamine. The relative rates of condensation and consecutive oxidation

determine the selectivity of the overall process (Scheme 18.13 ).

Scheme 18.13 Competing reactions of hydroxylamine with cyclohexanone and H

18.6 Ammoximation of Carbonyl Compounds 731

732 18 Titanium Silicalite-1

The nature of active species and the mode of generation of hydroxylamine are

the subject of several studies by various spectroscopic techniques [20, 130] . Accord-

ing to Zecchina and coworkers, the chemisorption of hydrogen peroxide produces

the familiar Ti

–

OOH species, which, however, in the basic medium evolves

towards a side - on - bonded anionic peroxide Ti( η

–

)

−

(Scheme 18.14 ). The adsorp-

tion of water and ammonia completes the coordination sphere of Ti, leading to

various species in which either adsorbate can be present with one or two molecules

at each time. In Scheme 18.14 , the suggestion is made, somewhat arbitrarily, of

the mixed adsorption species. A second issue that spectroscopic data do not help

solve is whether hydroxylamine is formed by intramolecular reaction between

chemisorbed species (in a Languimir – Hinshelwood - type mechanism) or by exter-

nal attack of physisorbed ammonia on peroxy oxygen (in an Eley – Rideal - type

mechanism). A recent kinetic study seems to favor the former pathway [131] . A

third issue concerns a possible role of Ti

–

OOH species in the formation of hydrox-

ylamine. The basic medium, indeed, shifts the acid – base equilibrium towards the

formation of the anionic peroxy species, but the presence of Ti

–

OOH in small

amounts cannot be ruled out completely. The greater electrophilicity of Ti

–

OOH

could compensate for its lower concentration.

18.7

Oxidation of N - Compounds

Primary amines, with at least one α - C

–

H bond, undergo consecutive oxidation to

yield unstable alkylnitroso intermediates. These can isomerize to corresponding

oximes, dimerize to nitroso dimers and produce nitro derivatives by further oxida-

tion (Scheme 18.15 ). The selectivity depends on the oxidant system and the reac-

tion conditions. The size restrictions of TS - 1 allow a greater selectivity to the

Scheme 18.14 Ammoximation mechanism on TS - 1.

(Adapted from Ref. [130] , with permission from Elsevier).

desirable oxime than conventional catalysts, at the expense of the bulkier dimer.

Methanol and t - butanol are suitable solvents, while acetone undergoes side reac-

tions with the amine and the oxidant [132] . The oxidation was broadened to include

several aliphatic and benzylic amines, owing to the versatility of alkyl oximes in

organic synthesis [133] .

The oxidation of secondary amines can produce the corresponding hydroxyl-

amines and nitrones (Equation 18.15 ) [134, 135] . TS - 1 and TS - 2 catalyzed the pro-

duction of the former, at H

/amine molar ratios below unity. The addition of

alkali metal additives improved the selectivity up to 90% [136] . High yields of

nitrones could be obtained at greater H

/amine ratios.

(18.15)

The oxidation of aniline and other arylamines yields a wide spectrum of prod-

ucts, among which azoxybenzenes are the most desirable (Scheme 18.16 ). TS - 1,

Ti - ZSM - 5, TS - 2, Ti,Al - β and Ti - MCM - 41, were studied as catalysts, with hydrogen

peroxide and t - butyl hydroperoxide as the oxidants [137 – 140] . The best yields, up

to 95%, were reported for Ti,Al - β and mesoporous Ti - silicates, with high aniline/

molar ratios to minimize over - oxidation to nitrobenzene. Anilines, with

either electron - donating or - withdrawing groups, and 1 - naphthylamine behaved

similarly. Owing to diffusional restrictions, TS - 1 was less effective than large - pore

and mesoporous catalysts.

The oxidation of aromatic and aliphatic oximes and of tosylhydrazones on TS - 1

with excess H

was reported to regenerate corresponding aldehydes [141, 142] .

These were obtained also by the one - pot oxidation of primary amines under analo-

gous conditions [143] .

Scheme 18.15 Oxidation of primary amines.

Scheme 18.16 Oxidation of aniline.

18.7 Oxidation of N-Compounds 733

734 18 Titanium Silicalite-1

18.8

Oxidation of S - Compounds

The oxidation of thioethers on TS - 1, Ti - β and Ti,Al - β produced corresponding

sulfoxides and, more slowly, sulfones by consecutive oxidation [144 – 149] . Allyl

methyl thioether was oxidized selectively on sulfur, with no reaction of the double

bond [148] . The kinetic law was obtained for the oxidation of dibutylsulfoxide on

Ti,Al - β [147] .

The oxidation of thiophene sulﬁ des combined with a selective separation

process of oxidized products, for example by adsorption on inert materials, might

offer an alternative to deep hydrotreating for desulfurization of oil fractions. On

these grounds, the oxidation of disulﬁ des has gained considerable interest [149] .

Owing to the nature of sulfur impurities, large - pore and even more mesoporous

Ti - silicates would be appropriate catalysts.

18.9

Industrial Processes Catalyzed by TS - 1

Three oxidation processes are already commercial. Other processes could be devel-

oped in the future, for example for the production of epichlorohydrin and phenol,

especially in the event that new production strategies reduced the cost of hydrogen

peroxide (see below). It is unlikely, however, that Ti - zeolites other than TS - 1 could

ﬁ nd application in the foreseeable future.

18.9.1

Hydroxylation of Phenol to Catechol and Hydroquinone

In 1986 the EniChem process replaced the Brichima process, with signiﬁ cant

materials and energy savings (Table 18.5 , entries 1 and 2). It operates with excess

phenol, fed with H

into a slurry of TS - 1 in aqueous acetone, at ca 100 ° C [5] .

While the conversion of the oxidant is driven to completeness, excess phenol is

recovered by distillation and recycled. The catalyst is structurally stable under

operating conditions, being deactivated only by pore plugging and active site

fouling. The regeneration is conveniently carried out thermally under ﬂ owing air.

With an overall capacity of 10 000 t a

− 1

, it is a modest process when compared to

recent applications of TS - 1 in ammoximation and propene epoxidation. The intro-

duction of digital photography, which no longer needs hydroquinone for the devel-

opment of silver emulsions, risks the continuity of the process, unless the selectivity

to catechol is greatly improved or new uses are developed for hydroquinone.

18.9.2

Salt - Free Production of Cyclohexanone Oxime

A demonstration plant, built by EniChem at the industrial site of Porto Marghera,

near Venice, began operation in 1994 (12 000 t a

− 1

). The process was operated con-

tinuously in the liquid phase in a stirred reactor. Cyclohexanone, ammonia and

hydrogen peroxide were fed, at 80 – 90 ° C and slight over - pressure, into a slurry of

TS - 1 in aqueous t - butanol. The efﬂ uent was separated from the catalyst through

a ﬁ lter, and sent to a distillation column. Excess ammonia and the solvent, recov-

ered overhead, were recycled in the reactor, while the aqueous solution of oxime,

taken from the bottom, underwent further puriﬁ cation. Periodic purging and

make - up operations of the catalyst were necessary to compensate for irreversible

decay.

The ﬁ rst commercial application, made possible by an agreement between

EniChem and Sumitomo, went on - stream in 2003 in Japan, within the context of

an integrated process for the production of ε - caprolactam by a new salt - free tech-

nology (ca 60 000 t a

− 1

). Actually, besides the ammoximation step, no major by -

product is produced even in the gas - phase rearrangement carried out on silicalite - 1

as the catalyst. On the whole, the ammonium sulfate is no longer a burden and

the gaseous emissions too are drastically reduced.

The truly innovative nature of the EniChem process over earlier ones is appar-

ent. The preparation of hydroxylamine in the latter case necessitates multi - step

operation, often ending with major co - production of inorganic salts, as in the

Raschig process. The ammoximation of cyclohexanone, with its in situ generation

of the intermediate, reduces signiﬁ cantly the investment and operation costs while

improving the environmental compatibility. The new process represents a good

example of how to combine proﬁ tability and environmental concern.

18.9.3

Propene Oxide Synthesis ( HPPO )

The industrial process for propene oxide manufacture is commonly referred to as

the HPPO ( hydrogen peroxide propene oxide ) process. EniChem set up a prototype

plant in 2002 [150] . BASF/Dow Chemicals and Degussa, in turn, have the con-

struction of commercial plants already in progress or at the planning stage [151] .

In the EniChem process (6 t day

− 1

), a dilute solution of hydrogen peroxide and a

gaseous stream of propene were fed into a stirred reactor containing TS - 1 and

aqueous methanol [150] . The process was operated continuously, at moderate

pressure and temperature. The selectivity could be as high as 98%. Propene, the

epoxide and methanol were recovered by distillation of the efﬂ uent, while glycol

by - products accumulated in the aqueous solution at the bottom of the column.

The latter, depending on the selectivity of epoxidation, was either processed to

recover the glycols or sent directly to the biological plant, for ﬁ nal treatment before

disposal. The stability of the catalyst to deactivation, achieved by the appropriate

control of operating conditions, was very high and did not require frequent regen-

eration. According to Romano, the new process is characterized by lower environ-

mental impact, simpler process design and relatively lower investment and

operating costs than conventional ones [150] .

Less is known of the processes developed by other companies. BASF/Dow

Chemical, in a joint venture, started the construction of a 300 000 t a

− 1

plant at the

BASF site in Antwerp (Belgium) at the end of 2006 [151] . Hydrogen peroxide will

18.9 Industrial Processes Catalyzed by TS-1 735

736 18 Titanium Silicalite-1

be supplied by a 230 000 t a

− 1

plant based on Solvay ’ s anthraquinone technology and

located at the same site in Antwerp. The operative start - up is scheduled for early

2008. Both plants will be the world ’ s largest single train ones in operation. HPPO

technology is also under evaluation for other facilities in the USA and Asia.

Degussa, in turn, has recently announced the commercialization of a HPPO

process, jointly developed with the engineering company Uhde. Parallel to this,

Degussa with Headwaters is also working on the direct synthesis of H

, for

which a demonstration plant was completed in 2006 [151] . According to the news

release, hydrogen peroxide will be obtained in the new process as a dilute metha-

nol solution to be used directly in the epoxidation of propene.

18.10

Problems in the Use of H

and Possible Solutions

Hydrogen peroxide is currently produced by the anthraquinone autoxidation ( AO )

process. It is a relatively expensive chemical, used for bleaching in the pulp/paper

(ca 57%) and textile industries (6%), for waste water and efﬂ uent treatment (5%)

and for laundry products (ca 12%). Only a minor amount (ca 10%) is currently

consumed for the production of ﬁ ne chemicals (the ﬁ gures refer to the early

2000s). The cost and the production technology, while compatible with its present

uses, ﬁ t less well with petrochemical applications. For instance, the cost of com-

mercial hydrogen peroxide (0.69 $ /lb, 100% H

) is comparable, on a weight

basis, to that of propene oxide (0.70 $ /lb) [152] . The annual capacity of the plants,

mainly comprising the ranges of 10 000 – 20 000 and 40 000 – 70 000 tons, with a

very few slightly over 100 000 t a

− 1

, is undersized for average petrochemical pro-

duction. The shipment of large amounts of hydrogen peroxide could represent a

problem from the point of view of logistics and safety. In addition, the anthra-

quinone route involves complex technology, not easy to access (Figure 18.2 ).

Thus, the applications of TS - 1 in the petrochemical industry were, from the

beginning, strong incentives for the investigation of potential alternatives for

commercial hydrogen peroxide production. Direct synthesis from the elements,

Figure 18.2 Simpliﬁ ed block diagram of anthrahydroquinone autoxidation (AO).

in situ production and process integration have been the most studied solutions.

Unless otherwise speciﬁ ed, HPPO is the process of reference in the following

discussion.

18.10.1

Direct Synthesis of Hydrogen Peroxide

The catalyzed reaction of the elements produces hydrogen peroxide directly, in

competition with water (Equation 18.16 ). Most selective catalysts are based on

supported Pd, often modiﬁ ed by other metals. Methanol is increasingly studied as

a reaction medium, from the perspective of the direct use of the solution obtained

in the HPPO process. For high selectivities, the reaction normally necessitates

relatively low temperatures, often room temperature or lower, and the presence

of mineral acids and other additives [153] .

(18.16)

The direct synthesis is in principle a one - step process, which is simpler and

easier to scale - up for large capacity plants than the AO route. However, it also

suffers from major drawbacks, for example continuing low yields and the hazards

of mixing ﬂ ammable gases. Various technical solutions are envisaged in the patent

literature to minimize the risks of explosion and ﬁ re, such as the dilution of

hydrogen and oxygen with an inert ﬂ uid.

18.10.2

In Situ Production of Hydrogen Peroxide

In situ generation removes dependence on external supplies. The availability of

hydrogen at the chemical site is the only requisite for its feasibility. The approach

generally envisaged in the literature is of two types [154 – 158] . The ﬁ rst requires

that TS - 1 is modiﬁ ed by a metal species, generally Pd, able to catalyze the direct

synthesis. H

, O

and propene are fed to the reactor to produce propene oxide and

water directly (Equation 18.17 ). On a microscopic scale, hydrogen peroxide is

formed on metal particles, then diffuses in TS - 1 to react with propene on Ti sites.

The high activity of TS - 1 in very dilute solutions of hydrogen peroxide prevents

the accumulation of the oxidant in contact with metal particles that would other-

wise catalyze its decomposition. A major drawback of the method relates to the

excellent hydrogenation properties of Pd, leading to signiﬁ cant losses of propene

as propane.

(18.17)

In early studies, carried out batch - wise on Pd - Pt/TS - 1, the maximum yield was

11.7%, with 46% selectivity based on propene [156] . By operating in a continuous

ﬁ xed - bed reactor, Jenzer and others showed that the selectivity could initially be

18.10 Problems in the Use of H

and Possible Solutions 737

738 18 Titanium Silicalite-1

as high as 99% at 3.5% conversion. The catalyst, however, underwent rapid deac-

tivation with a parallel decrease of the selectivity. Methyl formate from the oxida-

tion of methanol, and acrolein, acetone and other compounds from that of propene

became the main products after few hours on - stream [157] . The use of supercritical

carbon dioxide, without the addition of methanol/water, allowed the reduction of

side reactions [158] . Other studies envisage in situ generation by an electrochemi-

cal cell [159] and gas - phase epoxidation on Ag/TS - 1 at 150 ° C [160] .

A second approach consists of the autoxidation of an organic carrier, in the

presence of propene and TS - 1 (Equations 18.18 and 18.19 ) [154, 155, 161, 162] .

The solvent is a crucial aspect in the use of alkylanthrahydroquinone compounds

[154] . It should be able to solubilize both the oxidized and the hydrogenated forms

of the carrier and possess a molecular size larger than the pores of TS - 1. On this

basis, the various components of the working solution cannot interfere in the

epoxidation mechanism or undergo oxidative degradation at the Ti sites. Normally,

the conditions are met by the use of a mixture of two or more exotic solvents, to

which methanol is also added to increase the rate of epoxidation. Speciﬁ c tests

proved the feasibility of the method, albeit with somewhat lower yields than when

carried out with ex situ hydrogen peroxide [154] . A subsequent study, while con-

ﬁ rming early results, provided useful information on the reactor and operative

conditions [161] . The use of a gas - lift loop reactor, allowed the process to be carried

out satisfactorily even in the presence of two liquid phases, one richer in metha-

nol/water and the other in the working solution. The best yields were 82%, with

85% selectivity based on alkylanthrahydroquinone.

(18.18)

(18.19)

Major shortcomings also affect this in situ route to propene oxide. The most

obvious one is the even greater complexity than for the AO process, owing to the

modiﬁ cation required to accommodate the epoxidation of propene. This explains

why patents have been ﬁ led on anthraquinone compounds and other organic car-

riers, soluble in aqueous methanol. The aim was a radical simpliﬁ cation of the

working solution, trying to eventually use the simple methanol/water medium for

the whole series of reactions. Examples of soluble carriers are anthraquinone

compounds carrying hydrophilic alkyl ammonium or sulfonic substituent groups

and secondary alcohols.

18.10.3

Process Integration

The integration of propene oxide and hydrogen peroxide processes appears, from

a short - term perspective, a more realistic approach than in situ generation. It is

also favored by the location of both facilities at the same chemical site. Basically,

process integration envisages a simpliﬁ cation of H

recovery that can even

result, at best, in the complete elimination of the puriﬁ cation, concentration and

stabilization steps [154] . In this hypothesis, the aqueous H

extract obtained is

fed, without further processing, into the epoxidation reactor. Impurities from the

working solution, which may have been extracted with the H

, are hindered

from diffusing to Ti sites by their relatively large molecular size. Drawbacks could

ensue, for example for the activity of the catalyst, from inorganic additives that

may be required by the AO process [161] . Beside anthraquinone compounds, old

and new carriers such as isopropanol, once used industrially for the production

of hydrogen peroxide and 1 - phenylethanol, have also been studied from the per-

spective of process integration [29, 163] .

A variant envisages the use of aqueous methanol for both the extraction of

hydrogen peroxide and the epoxidation of propene as shown, in a simpliﬁ ed form,

by Figure 18.3 . Thus, no extra water is added to the epoxidation reactor, which

decreases separation and waste - treatment costs. On the laboratory scale, the quan-

tity of methanol dissolving into the working solution is small and apparently does

not interfere with the AO cycle of reactions [154, 161] .

In principle, process integration applies even more to direct synthesis of hydro-

gen peroxide, further improving the advantages over the anthraquinone route.

Methanol can be used to replace water as the solvent and the dilute methanol

solution obtained fed into the epoxidation reactor. Minimal puriﬁ cation may be

required, for example for the removal of hydrogen bromide and other additives

that may have been needed to increase the selectivity.

18.10.4

Miscellanea

The relatively small capacity of the phenol hydroxylation process does not justify

the development of alternatives to the use of commercial hydrogen peroxide. The

direct synthesis of phenol, for which process integration might be crucial, would

be a different matter.

The ammoximation process is not very compatible with in situ generation. From

one side, the possible reaction of ammonia with the working solution and the

complex recovery of cyclohexanone oxime rule out the use of organic carriers, and,

Figure 18.3 Integrated process scheme for the production of propene oxide.

18.10 Problems in the Use of H

and Possible Solutions 739