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

Подождите немного. Документ загружается.

498 11 Oxidation Reactions over Supported Metal Oxide Catalysts

33 Burcham , L.J. , Briand , L.E. and Wachs ,

I.E. ( 2001 ) Langmuir , 17 , 6164 .

34 Briand , L.E. , Jehng , J. - M. , Cornaqglia , L. ,

Hirt , A.M. and Wachs , I.E. ( 2003 )

Catalysis Today , 78 , 257 .

35 Badlani , M. and Wachs , I.E. ( 2001 )

Catalysis Letters , 75 , 137 .

36 Holstein , W. and Machiels , C. ( 1996 )

Journal of Catalysis , 162 , 118 .

37 Burcham , L.J. , Deo , G. , Gao , X. and

Wachs , I.E. ( 2000 ) Topics in Catalysis ,

11 / 12 , 85 .

38 Khaliullin , R.Z. and Bell , A.T. ( 2002 )

Journal of Physical Chemistry B , 106 , 7832 .

39 Doebler , J. , Pritzsche , M. and Sauer , J.

( 2005 ) Journal of the American Chemical

Society , 127 , 10861 .

40 Calatayud , M. and Minot , C. ( 2004 )

Journal of Physical Chemistry B , 108 ,

15679 .

41 Deo , G. and Wachs , I.E. ( 1994 ) Journal of

Catalysis , 146 , 323 .

42 Dunn , J.P. , Koppula , P.R. , Stenger , H.G.

and Wachs , I.E. ( 1998 ) Applied Catalysis

B Environmental 19 , 103 .

43 Martinez - Huerta , M.V. , Gao , X. , Tian ,

H. , Wachs , I.E. , Fierro , J.L.G. and

Banares , M.A. ( 2006 ) Catalysis Today ,

118 , 279 .

44 Gao , X. , Jehng , J. - M. and Wachs , I.E.

( 2002 ) Journal of Catalysis , 209 , 43 .

45 Zhao , Z. , Gao , X. and Wachs , I.E. ( 2001 )

Journal of Physical Chemistry B , 107 ,

6333 .

46 Zhao , C. and Wachs , I.E. ( 2006 ) Catalysis

Today , 118 , 332 .

47 Wachs , I.E. , Deo , G. , Jehng , J. - M. ,

Weckhuysen , B.M. , Guliants , V.V. ,

Benziger , J.B. and Sundaresan , S. ( 1997 )

Journal of Catalysis , 170 , 75 .

48 Turek , A.M. , Wachs , I.E. and DeCanio , E.

( 1992 ) Journal of Physical Chemistry , 96 ,

5000 .

49 Deo , G. and Wachs , I.E. ( 1994 ) Journal of

Catalysis , 146 , 335 .

50 Mitra , B. , Wachs , I.E. and Deo , G. ( 2006 )

Journal of Catalysis , 240 , 151 .

51 Hu , H. and Wachs , I.E. ( 1995 ) Journal of

Physical Chemistry , 99 , 10911 .

52 Kim , D.S. and Wachs , I.E. ( 1993 ) Journal

of Catalysis , 142 , 166 .

53 Kim , D.S. and Wachs , I.E. ( 1993 ) Journal

of Catalysis , 141 , 419 .

54 Jehng , J. - M. and Wachs , I.E. ( 1990 )

Catalysis Today , 8 , 37 .

55 Chen , Y. , Fierro , J.L.G. , Tanaka , T. and

Wachs , I.E. ( 2003 ) Journal of Physical

Chemistry B , 107 , 5243 .

Vanadium Phosphate Catalysts

Jonathan K. Bartley , Nicholas F. Dummer , and Graham J. Hutchings

499

Metal Oxide Catalysis. Edited by S. David Jackson and Justin S. J. Hargreaves

ISBN: 978-3-527-31815-5

12.1

Introduction

Vanadium phosphates have been of considerable interest since the mid - 1960s,

when Bergman and Frisch [1] found them to be effective catalysts for the oxidation

of n - butane to maleic anhydride (Equation 12.1 ).

(12.1)

Prior to this, maleic anhydride had been industrially manufactured by the oxida-

tion of benzene over supported V

- MoO

catalysts. In the late 1970s, when pol-

lution laws that restricted benzene emissions came into effect, industry began to

use the n - butane route. The reaction is of great importance as it is the only indus-

trial large - scale selective oxidation of alkanes currently in operation. It also involves

the functionalization of an alkane, providing a use for this rather unreactive oil

fraction.

Maleic anhydride is used as a chemical intermediate in the synthesis of fumaric

and tartaric acids, certain agricultural chemicals, resins in numerous products,

dye intermediates and pharmaceuticals [2] . It is primarily used as a co - monomer

for unsaturated polyester resins, which are used in the production of bonding

agents for plywood manufacture and when mixed with glass ﬁ bres for reinforced

plastics. Annual production of maleic anhydride is estimated to be over one

million tonnes [3] .

The selective oxidation of C

hydrocarbons leads to various commercially desir-

able products, which encouraged the development of the route involving the oxida-

tion of n - butane to maleic anhydride. This new route was also less expensive owing

to the lower cost of butane compared with benzene.

The n - butane oxidation process is interesting as it is an extensive oxidation, with

the cleaving of eight C

–

H bonds and the introduction of three oxygen atoms, yet

500 12 Vanadium Phosphate Catalysts

it still manages to occur selectively. The process is a fourteen - electron oxidation,

in comparison to other selective oxidation processes, which only require a

maximum of four electrons.

Many well characterized, crystalline vanadium phosphate phases have been

identiﬁ ed, whose structure and catalytic properties have been well documented.

Some of the most widely studied are the V

vanadyl orthophosphates ( α - , β - ,

γ - , δ - , ε - and ω - VOPO

, and VOPO

· 2H

O), and the V

vanadyl hydrogen phos-

phates (VOHPO

· 4H

O, VOHPO

· ½H

O, VO(H

)

), vanadyl pyrophosphate

((VO)

) and vanadyl metaphosphate (VO(PO

)

). Of these compounds,

VOHPO

· ½H

O (vanadyl hydrogen phosphate hemihydrate) is of particular inter-

est as a catalyst precursor, which after activation gives a catalyst mainly composed

of (VO)

( vanadyl pyrophosphate , hereafter VPP ).

In this chapter we will discuss how the behavior of the catalyst is inﬂ uenced by

a number of factors including: the method of preparation of the precursor, the

oxidation state of the catalyst, the phosphorus/vanadium ratio of the catalyst

and the activation conditions. A variety of techniques have been used to character-

ize the morphology and nature of the active sites of the catalysts and gain insight

into the mechanism. Furthermore, the effect of these preparatory techniques will

be discussed, with emphasis on the partial oxidation of n - butane. We will pay

particular attention to industrially relevant examples where possible and attempt

to describe the current state of the art.

Although there is a great deal of debate about certain aspects of vanadium

phosphate catalysts, Hodnett [4] has laid out a set of statements that most research-

ers generally agree upon:

• The most active and selective catalysts comprise mainly (VO)

• During testing in n - butane lean conditions, the oxidation state of the catalyst is

close to +4.

• The surface has some phosphorus enrichment. Only the surface layers are

directly involved in catalysis.

• The rate determining step is butane activation by hydrogen abstraction

12.2

The Active Catalyst

The structure of vanadium phosphate catalysts is dependent on a number of

factors. The P/V stoichiometry, thermal treatment time, activation temperature

and gas phase composition can all affect catalyst composition. By varying these

factors a variety of crystalline phases can be identiﬁ ed (by high - resolution trans-

mission electron microscopy ( HRTEM ) [5] Figure 12.1 a and X - ray diffraction

Figure 12.1 b) in the freshly activated catalyst [6] . It is widely accepted that VPP

plays an important role in the oxidation of butane to maleic anhydride and most

hypotheses are based on the (100) face (Figure 12.2 ). Additionally, this phase has

been reported to be an efﬁ cient catalyst for the oxyfunctionalization of light paraf-

12.2 The Active Catalyst 501

ﬁ ns: (a) for the oxidation of ethane to acetic acid [6, 7] , (b) for the oxidation and

ammoxidation of propane to acrylic acid [8] and acrylonitrile [9, 10] , respectively,

and (c) for the oxidation of n - pentane to maleic and phthalic anhydrides [10 – 19] .

The catalytic behavior of the different crystal faces has been investigated by

Inumaru and coworkers by exposing individual planes [20] . VPP was deactivated

by the surface deposition of SiO

. The crystallites were then fractured to expose

Figure 12.1 (a) An axial HREM image from the (021)

projection of (VO)

, (b) XRD pattern of an activated

(VO)

catalyst. (Reproduced with permission).

Figure 12.2 The (100) plane of (VO)

. Grey = V octahedra, Purple = P tetrahedra, Red = O.

502 12 Vanadium Phosphate Catalysts

the side faces, for example the (021) and (001) faces. The side faces were found to

be non - selective, with maleic anhydride formed only on the (100) face. There is

debate as to whether VPP is indeed the active catalyst, or if a combination of phases

are responsible for the reaction.

Transient studies by Ballarini and coworkers showed that the active surface of

equilibrated catalysts is different depending on the reaction conditions and the

P/V ratio of the catalyst [21] . At low temperature (320 ° C) an active surface forms

that is selective and probably is more like VOPO

than VPP. However, as the

temperature is increased to 380 ° C this material becomes less selective. The active

phase formed at T > 380 ° C was found to be less active than the low - temperature

phase ( T < 380 ° C) but has increased selectivity at this temperature. At these tem-

peratures the active site is found to hydrolyze and oxidize and Ballarini and

coworkers propose that the active surface is a VO

/polyphosphoric acid mixture.

The authors speculate that the different phase evolutions at different tempera-

tures, which are also dependent on very minor changes in the P/V ratio, could be

the cause for the very differing surfaces observed by both in situ and ex situ studies

of the active catalyst.

12.2.1

The Oxidation State of the Catalyst

The ﬁ nal oxidation state of the activated catalysts varies between +4.00 and +4.40,

depending on the amount of V

present in the catalyst. There is extensive discus-

sion as to whether V

and V

phases are important in the reaction mechanism.

Ebner and Thompson have postulated that the V

phases that are formed during

the activation period, are unimportant and do not contribute to the oxidation

mechanism [22] . They have found that after several hundred hours on stream, the

orthophosphate phases are reduced to pyrophosphate, giving an active catalyst

with a ﬁ nal oxidation state of +4.00 to +4.04. The concentration of oxygen and

butane in the reactant feedstock determines the time needed to equilibrate the

catalyst. Based on this study, they suggest that other researchers (who ﬁ nd V

phases in the active catalyst) have not performed the activation process fully, or

have an unfavorable redox potential in the gas stream.

A series of vanadium phosphate catalysts prepared by different routes and con-

taining different phases were examined by Guliants and coworkers [23] . From this

study it was concluded that the catalytically active phase is an active surface layer

on VPP. Their experimental results showed VOPO

phases to be detrimental to

the performance of the catalyst. This was conﬁ rmed by Cavani and Triﬁ r ò , who

suggested that V

sites are responsible for the over - oxidation of maleic anhydride

to carbon oxides [24] .

A number of groups dispute this one - phase hypothesis. They suggest that V

phases are important in the active catalyst and are formed as the result of a redox

mechanism (Figure 12.3 ) [25 – 27] . Bordes cites the apparent need for two contra-

dictory conditions during the oxidation of butane as evidence for a multi - phase

active catalyst [19] . It was stated that oxygen associated with V

activates butane

12.2 The Active Catalyst 503

while the oxygen associated with V

is incorporated later. V

gives a very active

catalyst that has poor selectivity, while V

gives a very selective catalyst that has a

low activity. This mechanism has been supported by the X - ray Photoelectron

Spectroscopy ( XPS ) studies of Coulston and coworkers [28] and Temporal Analysis

of Products ( TAP ) studies by Rodemerck and coworkers [29, 30] and Lorences and

coworkers [31] . These studies report that in the absence of V

sites no maleic

anhydride was observed and the main reaction product was furan. When V

sites

were present maleic anhydride was produced.

Experimental results on pure vanadium phosphate phases and active catalysts

suggest that the active catalyst is VPP with domains of V

on the (100) face [33]

The lack of selectivity of side faces found by Inumaru and coworkers [34, 35] is

attributed to the difﬁ culty of the re - oxidation to V

of these planes. Hutchings and

coworkers propose a V

couple, which can be present on the surface of a range

of vanadium phosphate phases, as the active site [36] . The active phase is suggested

to comprise a well dispersed micro - crystalline VOPO

on a VPP matrix. Centi and

coworkers proposed that the reaction proceeds via a series of redox couples [37] ,

the activation of butane requiring a V

–

couple, while the subsequent conver-

sion to maleic anhydride requires a V

–

couple.

There has been speculation about the source of the V

in these two - phase

systems. α

- VOPO

, γ - VOPO

and δ - VOPO

are commonly found in (VO)

based catalysts [38, 39] . More recently, ω ′ - VOPO

has been identiﬁ ed as a surface

species [40 – 42] in the ﬁ nal active catalysts, especially those activated in slightly

reducing conditions. Koyano and coworkers [43, 44] propose the X

phase as the

source, although they admit this phase may be δ - VOPO

. It should also be

noted that a recent in situ X - ray Diffraction ( XRD ) study by Conte and coworkers

[45] showed that metastable VOPO

phases can exist at the elevated temperatures

of the catalytic process but that these can be undetectable in the ﬁ nal catalyst as

they disorder into an amorphous material at room temperature.

Agaskar and coworkers [46] proposed that the catalyst surface contains clusters

of four active dimeric sites that can each exist in one of four states. These states

differ in the number of oxygen atoms associated with them and in the oxidation

state of the vanadium ions present (Scheme 12.1 ). The V

site [ S

] acts as an

Figure 12.3 Scheme of the proposed evolution of the VPO catalyst with time [32] .

504 12 Vanadium Phosphate Catalysts

oxygen acceptor. It can either react with gaseous oxygen to give [ S

], or accept an

oxygen atom from [ S

] to give [ S

]. The active site for the butane adsorption and

reaction is thought to be one associated with two vanadyl groups ([ S

] or [ S

]).

Rodemerck and coworkers [29] have investigated the importance of V

using TAP

experiments. They propose that only V

and V

sites are important, and demon-

strate that maleic anhydride was not formed on catalysts with an oxidation state

below +3.96 [29] . However, this is the bulk oxidation state and so does not discount

the possibility of isolated V

sites being present at the catalyst surface.

12.2.2

The Phosphorus - to - Vanadium Ratio of the Catalyst

Triﬁ r ò has proposed that the active catalyst is pure VPP and found that the catalyst

had a slight increase in oxidation state after the equilibrium period [47] . The small

increase from +4.00 to +4.03 was reproducible and attributed to isolated V

surface

sites being formed. The P/V ratio is proposed to be a key factor in the stabilization

of V

within the catalyst, as VOPO

formation becomes very difﬁ cult at P/V ratios

above 2.0. Triﬁ r ò had stated that a very high surface P/V ratio is required for an

active and selective catalyst, and experimentally has found surface P/V ratios of

10 : 1.

The P/V ratio was proposed by Ruiz and coworkers to be an important factor

in determining the activity and selectivity of the catalyst [48] . They acknowledge

that the majority of industrial catalysts have two or more different phases present,

and suggest that the active catalyst is a mixture of two phases with different P/V

ratios. By mixing a low P/V ratio catalyst and a high P/V ratio catalyst, they found

a 5.7 - fold increase in yield and a 3.8 - fold increase in selectivity, compared to the

sum of the individual phases. Ruiz suggests different roles for the two phases.

The high P/V ratio phase acts as an oxygen acceptor, whereas the low P/V ratio

phase acts as an oxygen donor. Experiments with Sb

and BiPO

, which are

Scheme 12.1

12.2 The Active Catalyst 505

known to be good oxygen donors, support this assignment of roles for the different

phases. Ruiz found mechanical mixtures of a high P/V ratio catalyst with Sb

and BiPO

gave a 6.0 - fold increase in yield and a 5.4 - fold increase in selectivity,

over the standard (VO)

catalyst.

Garbassi and coworkers [49] observed that a P/V ratio of 1.05 – 1.1 is necessary

for a high catalytic performance.

This slight excess of phosphorus is proposed to stabilize the V

phase, as V

is only detected in catalysts with a bulk P/V ratio less than 1.0. The ﬁ ndings of

Garbassi and coworkers were conﬁ rmed by Hodnett and coworkers [50, 51] who

suggested that a surface enrichment of phosphorus (P/V = 2.4) is the important

factor for a selective catalyst. They propose that the excess surface phosphorus is

responsible for the selectivity of the catalyst by isolating the V

active site on the

surface.

Morishige and coworkers [52 – 54] found that although the industrial catalyst had

a bulk P/V ratio of approximately 1.0, the surface had a high phosphorus concen-

tration. Their studies showed that the surface P/V ratio was 1.6 – 1.8, when the bulk

P/V ratio was 0.95 – 1.2. They suggested that a phosphorus - rich layer, supported

on the surface of (VO)

, is responsible for the oxidation of butane to maleic

anhydride. For catalysts with a P/V ratio of between 1.0 and 2.0, Satsuma and

coworkers also found catalytic improvements [55] . They proposed that at P/V ratios

less than 1.0 a number of vanadium sites remain inactive but at higher P/V ratios

all surface sites are active.

However, it should be noted that Coulston and coworkers [28] reported contra-

dictory ﬁ ndings. In summary, they found that the surface P/V of a number of pure

vanadium phosphate phases and activated catalysts are all closer to that of the bulk

than has been previously reported, leading to the claim that phosphorus enrich-

ment reported by other groups was based on incorrect calibration of XPS

instrumentation.

12.2.3

The Role of Amorphous Material

There is considerable discussion in the literature as to whether amorphous mate-

rial, commonly found in vanadium phosphate catalysts, plays a role in the catalysis

of n - butane to maleic anhydride. Industrial catalysts undergo a lengthy activation

process (often > 1000 h) before they are equilibrated. During the activation, the

crystallinity of the catalysts increases, leading some researchers to suggest that the

more crystalline the catalyst, the better the performance [25] . Guliants and cowork-

ers observed that although an amorphous surface layer was found to be present

on fresh catalysts this disappeared on - line, leading to an increase in catalytic activ-

ity [56] . Most researchers consider that the (100) plane of crystalline VPP is the

catalytically active plane, and the mechanisms that have been proposed use this

as their active site [46, 57 – 69] . However, other researchers consider that an amor-

phous material supported on a (VO)

matrix plays an important role in

catalysis.

506 12 Vanadium Phosphate Catalysts

As catalysts commonly contain both amorphous and crystalline material, it

cannot be stated with any certainty whether or not the amorphous phase is catalyti-

cally active. However, experimental observations have added weight to the theory

that amorphous material is the catalytically active phase in vanadium phosphate

catalysts.

Research by Ruiz and coworkers [48] into high and low P/V ratio catalysts, led

to the hypothesis that the active catalyst is made up of VPP in conjunction with

an amorphous phase with a high oxidation state near +5.

Morishige and coworkers also suggested that the active catalyst is an amorphous

phase with excess phosphorus on the surface of (VO)

[52] . Evidence was

found by extracting the amorphous phase from the bulk. Catalytic testing yielded

the same activity and selectivity as (VO)

, suggesting that the amorphous

phase is catalytically active. It has been proposed that the amorphous phase is

phosphorus rich, accounting for the high surface P/V ratio that is commonly

observed experimentally [28, 48, 50 – 52, 70] .

In situ Raman spectroscopy has been carried out on precursors prepared in

aqueous solution (referred to as VPA materials see Section 12.3.1 ), as they were

converted to the active catalyst [38] . They found that during the activation, there

is a structural disordering at 370 ° C, which corresponds to the appearance of maleic

anhydride. The disordering was found to occur at a lower temperature (300 ° C),

when maleic anhydride was added to the butane/air reaction mixture. This dem-

onstrates that the presence of the products is important in controlling the struc-

tural transformations, and that a highly disordered structure can be important in

selective butane oxidation.

Further evidence for the catalytic importance of amorphous material comes

from experiments using cobalt - doped catalysts. Hutchings and coworkers found

that doping the catalysts with cobalt improved their performance [36] . Additionally,

Sajip and coworkers [71] found that the cobalt - promoted catalysts are far more

disordered than the undoped catalysts. In the doped catalysts, the promoter is

dispersed in the amorphous phase and cobalt is not found in the VPP crystals. It

is thought that one of the properties of the cobalt promoter is the stabilization of

the disordered phase and V

phases in the ﬁ nal catalysts, which leads to improved

performance. This implies that the disordered material is the catalytically active

vanadium phosphate phase.

Hutchings and coworkers [72 – 74] have prepared vanadium phosphate catalysts

using supercritical precipitation methods. These materials were found to be

amorphous by XRD and electron diffraction, but showed activity comparable to

standard vanadium phosphate catalysts. This demonstrates that an amorphous

surface layer can be the active phase in these catalysts and that the crystalline

VPP that has been so well studied may be nothing more than an elaborate

support.

However, Schimoda and coworkers [75] and others [6, 76] have dismissed the

amorphous phases prepared from the precursor VO(H

)

as less selective

than the crystalline (VO)

catalyst.

12.2 The Active Catalyst 507

12.2.4

The Disordered Plane

The selective oxidation of n - butane and 1 - butene on vanadium phosphate catalysts

prepared via different routes was investigated by Cavani and coworkers [77] Precur-

sors prepared in aqueous medium were found to have greater crystallinity than

those prepared in organic solvents (the activity and selectivity of which was the

same for 1 - butene oxidation). However, for butane activation, the crystalline cata-

lyst was considerably less active than the organically prepared catalyst, which had

an XRD pattern showing some disorder in the (100) plane.

It has been suggested that organic compounds occluded between the vanadium

phosphate layers cause this disorder [78] . The disorder may be derived from a

number of structural modiﬁ cations: a missing oxygen atom, an inversion from a

trans - to a cis - vanadyl position, or from the modiﬁ cation of the V

–

O bond strength.

It is proposed that these defects can all cause the creation of new active centers

for butane activation. Furthermore, Cornaglia and coworkers [79] also report an

increase in selectivity to maleic anhydride as the disorder in the (100) plane

decreases.

The opposite effect has been reported for a series of organically prepared cata-

lysts [25] . The catalytic performance to maleic anhydride was found to improve

with the increase of crystallinity of VPP. A number of studies on catalysts that

have been on - line for several hundred hours have provided supporting evidence

that crystalline compounds are the most active and selective [79, 80] .

12.2.5

Acid – Base Properties

Some research groups consider the acidity of the catalysts to be an important

aspect in controlling the catalyst performance. An infrared study of the acid sites

using NH

, pyridine and acetonitrile as probe molecules showed the existence of

Lewis and Br ø nsted acid sites [24, 65, 79, 81 – 83] . A correlation was observed

between the selectivity to maleic anhydride and the number of strong Lewis acid

sites. It has been suggested that hydrogen abstraction occurs on Lewis acid sites

), and that butane C

–

H bond cleavage results from interactions between Lewis

acid sites and Br ø nsted acid sites (P

–

OH) [83] .

Centi and coworkers [84] have suggested that, in addition to Lewis acidity,

Br ø nsted acidity plays an important role in the selective oxidation of butane to

maleic anhydride. The surface phosphorus enrichment means that a number of

–

OH groups are present on the catalyst surface. Centi offers three hypotheses

for the role of Br ø nsted acidity: the stabilization of reaction intermediates, the

stabilization of an adsorbed oxygen species, or the generation of an organic surface

species that is involved in oxygen activation or transport.

Ai measured the acid site concentration by the adsorption of ammonia [85] . No

correlation was found between the P/V ratio, the acidity and the catalytic activity.