Heterogeneous Hydrocarbon Oxidation - American Chemical Society

---

Chapter 10

The Catalytic Activity of Wells—Dawson and Keggin Heteropolyoxotungstates in the Selective Oxidation of Isobutane to Isobutene 1

2

2

3

Fabrizio Cavani , Clara Comuzzi , Giuliano Dolcetti , Richard G. Finke , Arianna Lucchi , Ferruccio Trifirò , and Alessandro Trovarelli 1

1

2

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

1

Dipartimento di Chimica Industriale e dei Materiali, Viale Risorgimento 4, 40136 Bologna, Italy Dipartimento di Scienze e Tecnologie Chimiche, Via Cotonificio 108, 33100 Udine, Italy Department of Chemistry, Colorado State University, Fort Collins, CO 80523

2

3

The catalytic performance of the Wells-Dawson-type polyoxotungstate, K P W O , in the oxydehydrogenation of isobutane to isobutene has been studied, and compared with the activity of the corresponding Keggin compound, K PW O . The Wells-Dawson compound is characterized by a lower activity than the Keggin salt, but by a remarkably higher selectivity to the olefin. The activity of the WellsDawson catalyst is strongly dependent on the hydrocarbon content in the feed, and increases with increased partial pressure of isobutane. This unexpected autocatalytic behavior is explained by a proposed mechanism where the olefin product induces a modification of the catalyst surface, with the creation of more active sites. In addition, at the highest hydrocarbon content in the feed, the contribution of heterogeneously-initiated, homogeneous gas-phase reactions become important, thus favoring high selectivity to isobutene. 6

2

18

62

3

12

40

Interest in heteropolyanions as catalysts for the liquid-phase and gas-phase oxidation of hydrocarbons has been growing rapidly in the last years (7,2). The unique features of these materials, their high acidity and tunable redox properties, make them ideal for a variety of applications. Heteropolycompounds have found industrial applications for many years as catalysts for the oxidation of methacrolein to methacrylic acid, and are well studied as catalysts for acid-catalyzed reactions (alkylation of isobutane, etherification) and as catalysts for the homogeneous and heterogeneous oxidation of olefins and paraffins (7,2). Their reactivity and selectivity can be fine-tuned, for example by modifying their composition and structure through incorporation of transition metal ions or alkali metal counterions. Among the different classes of heteropolyanions (Keggin, Anderson, WellsDawson and Waugh), Keggin-type complexes account for the majority of studies and applications in homogeneous and heterogeneous reactions, probably because of their 0097-6156/96/0638-0140$15.00/0 © 1996 American Chemical Society In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

10.

CAVANI ET AL.

Selective Oxidation of Isobutane to Isobutene

141

higher thermal stability under reaction conditions, their easy preparation, and the fact that the Keggin-anion structure has been known since 1934. However, the use of Wells-Dawson type polyoxoanions as oxidation catalysts has been limited to homogeneous, liquid-phase oxidation of olefins (3,4). More specifically, no example or application of a gas-phase, heterogeneous catalyst using a Wells-Dawson type polyoxoanion has previously been reported in the scientific literature. Among the organic substrates which are studied as raw materials for oxifunctionalization reactions, paraffins deserve particular attention. The oxidative activation of paraffins is a current challenge in catalysis: new low cost, widely available, raw materials for the petrochemical industry is the goal here (5). Among light paraffins, isobutane is one of the currently most interesting targets due to the increasing market for isobutene, which in turn is a raw material for the synthesis of oxidation products such as methacrylic acid (currently produced via the environmentally unfriendly acetone-cyanohydrin route), or for the synthesis of methyl t-butyl ether, an octane booster for reformulated gasolines. Heteropolyoxomolybdates have, for example, been investigated as catalysts for the one-step transformation of isobutane to mixtures of methacrolein and methacrylic acid (6,7). Herein we report the catalytic reactivity of Wells-Dawson-type and Keggintype heteropolyoxotungstates for the oxidation of isobutane. Potassium salts of (PW12O40) - and of ( P W 0 ) - were prepared and characterized, since it is known that potassium salts exhibit a considerably enhanced structural stability with respect to the parent acid form, or in comparison to salts containing smaller cations CO. 3

6

2

1 8

6 2

Preparation of Catalysts The Wells-Dawson compound of composition K ^ ^ W j g O ^ ' l O ^ O was prepared as detailed elsewhere (#); its purity and composition were checked by FT-IR, solution 31p NMR, and elemental analysis. K3PW12O40 prepared by precipitation from an aqueous solution containing Na2WC>4 and H3PO4. H N O 3 was added to the solution until a pH lower than 1.0 was reached, and then K N O 3 was added for precipitation of the Keggin salt. The resultant K3PW12O40 purity was checked by X R D and IR. Before catalytic tests, all compounds were calcined at 450°C for 3 h in air. Surface area analyses were carried out by measuring N 2 adsorption isotherms with a Carlo Erba Sorptomatic 1900 Instrument interfaced with a computer. The infrared spectra were collected as KBr pellets with a Digilab FTS-40 instrument. w

a

s

Thermal Evolution of K^WigOftj-lOE^O and K3PW12O40 The surface areas of the "as synthesized" powders were respectively 1.5 and 15 m^/g for K 6 P 2 W 1 8 O 6 2 I O H 2 O and for K3PW12O40. Treatment in air at 450°C does not affect the textural properties; in the case of the Wells-Dawson compound, K6P2WigO62'10H2O, a slight modification of surface area (from 1.5 to 3 rrfi/g) is observed, which is probably related to the loss of crystallization water from the material, which occurs at around 150°C, as detected by thermogravimetric analysis. Figure 1 shows for comparison the FT-IR spectra of the catalysts before and after thermal treatment in air at 450°C for 2h. The vibrational bands typical of P-0 (ca.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

142

HETEROGENEOUS HYDROCARBON OXIDATION

b d

I Transmittance 10%

1200

1100

1000 900

800

Wavenumber

700

600

500

400

(cm~1)

Figure 1. FT-IR spectra of the K3PW12O40 and of the K6P2W18O62 catalysts

as synthesized (a and b, respectively), and after calcination at 450°C (c and d, respectively).

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

10. CAVANI ET AL.

Selective Oxidation of Isobutane to Isobutene

143

1090 and 1020 cm~l in the Wells-Dawson and in the Keggin compound, respectively), W=0 (ca. 780-800 c n r ) and W-O-W (ca. 910 and 950 cm" for the Wells-Dawson compound, and ca. 980-1000 c m for the Keggin) are observed. No major modification of the JR pattern is induced by treatment in an oxidizing atmosphere up to 500°C, indicating that the polyoxoanion primary structure remains intact. After reaction (at least 300 h time-on-stream), the IR patterns of the unloaded catalysts did not differ from those of the starting materials, indicating that the catalysts were stable in the reaction environment. This was also confirmed by checking the catalytic activity which was constant with time on stream. 1

1

- 1

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

Catalytic Activity of the Wells-Dawson and Keggin Compounds in Isobutane Oxidation The catalytic activity for the gas-phase oxidative dehydrogenation of isobutane was examined using a stainless steel flow reactor operating at atmospheric pressure, all as previously described (6). Three grams of granulated catalyst (particle size 0.3-0-5 mm) were loaded for tests. Results were collected after approximately 50 h time-on-stream, and no effect of catalyst deactivation was observed even after 300 h time-on-stream. Analysis of the products was done as described previously (6). Carbon balance was always in the range 95-110%. The effect of the reaction temperature on the catalytic performance of K3PW12O40 and K6P2Wig062> both calcined at 450°C, is shown in Figures 2 and 3, respectively. The main products in both cases were isobutene and carbon oxides; minor amounts of propylene also formed, as well as acetic acid, methacrolein and methacrylic acid. Significantly, the main products obtained were different from those obtained under the same conditions with Keggin polyoxomolybdates; in this case, in fact, the K3PM012O40 catalyst yielded methacrolein and methacrylic acid as the main products of selective oxidation, and isobutene was obtained only in minor amounts (6). This fact points out the different reactivity of molybdenum and tungsten in polyoxometalates; Mo VI is able to i) abstract hydrogen and ii) insert oxygen onto the activated hydrocarbon, while W ^ I can only perform the first type of oxidative reaction. The MoVI vs. W ^ I greater electronegativity and lower M-O bond strength account for the different reducibility and reoxidizability (9) of the two metal ions. Results show that the Keggin, K3PW12O40, and the Wells-Dawson, K 6 P 2 W l 062> compounds are characterized by a different catalytic performance. The Keggin compound was clearly more active, activating isobutane already at 350°C (total oxygen conversion was reached at approximately 420-430°C), while the WellsDawson type catalyst was almost inactive at this temperature. This difference might be explained by the different values of surface area of the two compounds. However, the different apparent activation energies for isobutene formation in the two cases (30 kcal/mol for the Wells-Dawson compound, 15 kcal/mol for the Keggin compound) also indicates a different nature or reactivity of the active site. The Wells-Dawson compound was, however, much more selective to isobutene; in addition, remarkably, this selectivity to isobutene was roughly constant over the entire examined range of temperature, at a residence time of 3.6 s. This might indicate that isobutene is kinetically stable under the reaction conditions, and does not undergo consecutive oxidative degradation reactions. The Wells-Dawson compound is more selective even 8

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

144

HETEROGENEOUS HYDROCARBON OXIDATION

100 6^

> o CD CO

> c o o c Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

0)

O) >* X

o

360

380

400

temperature, °C Figure 2. Isobutane and oxygen conversions, and selectivity to the products as a function of the reaction temperature for the K3PW12O40 catalyst. Symbols: isobutane conversion (0); oxygen conversion ( A ) , selectivity to isobutene ( • ) , selectivity to carbon oxides ( V ) , selectivity to the other products (M). Other conditions: residence time 3.6 s, feedstock composition: 26% isobutane; 13% oxygen; 12% water; balance helium.

100

> o 0)

0 CO

> c o o c

0 D)

o

370

390

410

430

450

temperature, °C Figure 3. Isobutane and oxygen conversions, and selectivity to the products as a function of the temperature for the K6P2W18062- Symbols and conditions as in Figure 2. In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

10.

CAVANI ET AL.

Selective Oxidation of Isobutane to Isobutene

145

if we compare the results obtained on the two catalysts at comparable isobutane conversion, despite the higher temperature at which the Wells-Dawson catalyst is active; the higher selectivity accounts for a lower consumption of dioxygen, thus allowing a higher maximum isobutane conversion in comparison to the Keggin compound.

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

The autocatalytic effect at high isobutane concentration The effect of an increasing concentration of isobutane in the feed on the Keggin and Wells-Dawson catalysts are shown in Figures 4 and 5, respectively. The higher activity of the Keggin compound is confirmed; the isobutane conversion remains approximately constant (8-9%) over the complete range of isobutane concentrations examined, as expected for the case of a reaction order close to 1. The oxygen conversion correspondingly increased when the hydrocarbon concentration was increased, because of the higher hydrocarbon-to-oxygen ratio in the feedstock, and the selectivity to isobutene was significantly improved. On the contrary, the behavior of the Wells-Dawson compound proved rather unexpected. Under the conditions examined, we can distinguish two regions: a first region, for hydrocarbon concentrations lower than approximately 15-20%, in which the catalyst exhibited a very low activity, and the conversion was approximately constant, implying a reaction order close to 1 (i.e., a behavior similar to that observed with the Keggin catalyst, apart from the lower activity), and a second region, for isobutane concentrations higher than 20%, where the conversion steadily increased, a trend which clearly indicates a reaction order higher than 1. The effect was completely reversible, and therefore it cannot be explained as due to a structural irreversible transformation, but rather might be due to a temporary and reversible surface modification. This behavior clearly points to an autocatalytic effect at high isobutane partial pressure, possibly due either to i) an in-situ modification of the active centers when the gas-phase composition is varied, or to ii) one of the products acting itself as a reactant or as catalyst for the reaction, or acting as a modifier of the catalyst surface. Several different possibilities could explain this effect, so additional tests were carried out in order to further define the phenomenon: 1) A first possibility is that an increase in the isobutane concentration might favour the reduction of the catalytic surface (for instance, of the W ^ I ions), with formation of more active W^* centers. In favour of this hypothesis is the observation of an induction period when changing the feed composition, a phenomenon which might well be explained by a progressive modification of the catalyst surface. This induction period is shown in Figure 6, which illustrates the progressive variation in catalytic performance after varying the feed composition from 42% to 8% isobutane content in the feed (i.e., from a gas phase composition under which the catalyst is active to a situation under which it is inactive), and vice versa from 26% to 42% isobutane. The reduction of tungsten should be disfavoured under more oxidizing conditions, that is at higher oxygen concentration in the feedstock. On the contrary, Figure 7 reports data showing a remarkable increase in the catalytic performance when the oxygen content in the feed is increased. The oxygen effect is similar to that obtained by increasing the isobutane concentration. Since an increase in either

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

146

HETEROGENEOUS HYDROCARBON OXIDATION

100 6^

> o 0 0

CO

> c o o c Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

0 X

o

isobutane content in feed, mol.% Figure 4. Isobutane and oxygen conversions, and selectivity to the products as a function of the isobutane content in the feed for the K3PW12O40 catalyst. Symbols are the same as in Figure 2. Other conditions: T 470°C, residence time 1 s, feedstock composition: 13% oxygen; 12% water; balance helium.

0

10

20

30

40

50

isobutane content in feed, mol.% Figure 5. Isobutane and oxygen conversions, and selectivity to the products as a function of the isobutane content in the feed for the K6P2W18O62 catalyst. Symbols and conditions are the same as in Figure 4.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

10. CAVANI ET AL.

Selective Oxidation of Isobutane to Isobutene

147

180

150

time of reaction, min Figure 6. Isobutene ( A ) and carbon oxides (O) yields as a function of the time of reaction after changing the isobutane content in the feedstock from 26% to 42%, and from 42% to 8%. Time = 0 min corresponds to the moment at which the change is made. Other conditions: T 470°C, residence time 1 s, feedstock composition: 13% oxygen; 12% water; balance helium.

0

5

10

15

20

25

30

oxygen content in feed, mol.% Figure 7. Isobutane and oxygen conversions and selectivities to the products as functions of the oxygen content in the feed for the K6P2W18062 catalyst. Reaction conditions: T 470°C, residence time 1 s, feedstock composition: 26% isobutane; 12% water; balance helium. Symbols are the same as in Figure 2. In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

148

HETEROGENEOUS HYDROCARBON OXIDATION

isobutane or oxygen clearly increases the concentration of the products, it is possible that the products (rather than isobutane) may change the catalyst surface to make it more active, thus yielding the observed autocatalytic effect. Water, a product of the reaction, was shown to not be playing this role, since water was also fed at low isobutane concentration, but a low activity was seen under these conditions. In order to study the possible role played by the isobutene product, tests were done by co-feeding 8% isobutane and 2% isobutene, together with oxygen and water, at 470°C and a residence time 1 s. The results reported in Table I indicate that, under these conditions and in the presence of isobutene, the isobutane was more reactive than it was in the absence of added isobutene, but that primarily carbon oxides were formed (for comparison, the results obtained when oxidizing only 8% isobutane and only 2% isobutene are also given). This effect might be due either to a modification of the surface induced by the presence of the gas-phase olefin, with the formation of sites active in isobutane oxidation, or to the ignition of gas-phase homogeneous reactions due to the presence of the more labile, allylic C-H bond in isobutene. An analogous activation effect was obtained when co-feeding 2% propylene and 8% isobutane; in this latter test isobutane was more reactive, but with the formation of only carbon oxides, and no isobutene.

Table I. Results obtained by co-feeding isobutane and isobutene. Hydrocarbon in feed, 1C4H8 in the outlet, CO in the outlet, mol.% mol.% mol.% 2.4 1.8 8 % i C H + 2%iC H8 x

4

1 0

4

8% 1C4H10

0.05

0.02

2% i C H * 1.85 0.6 Reaction conditions: T 470°C, residence time 1 s, O2 13 mol.%, H 2 O 12 mol.%, balance He. 4

In order to define more clearly the role of the catalyst redox level, the calcined catalyst was first reduced by treatment in a diluted H 2 flow at 400°C, and then subjected to the reaction environment. The effect of the isobutane content in feed on the catalytic activity is displayed in Table II.

Table II. Reactivity of the KfiPiWiftOfil catalyst prereduced in H^/He / C 4 / / / q in feed, iC4H j q conv./ Sel. i- C4H8, Sel. CO , Sel. others, mol.% O2 conv., mol.% mol.% mol.% mol.% 13 1.9/10.0 0 88 13 x

a

26 2.3/15.2 39 57 4 Reaction conditions: T 470°C, residence time 1 s, O 2 13 mol.%, H 2 O 12% mol.%, balance He. Others are propylene, and traces of methacrolein and methacrylic acid. a

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

10. CAVANI ET AL.

Selective Oxidation of Isobutane to Isobutene

149

The reduced catalyst was more active than the untreated one at low hydrocarbon concentration, but was also characterized by a very low selectivity to the isobutene. Even though the pre-reduced catalyst soon re-equilibrates in the reaction environment, the results obtained point out the role of the surface degree of oxidation. This is also in agreement with the results obtained by co-feeding the isobutane and an olefin (either isobutene or propylene), which led to an increase in the conversion of isobutane, but with the exclusive formation of carbon oxides. Figure 8 shows the UV-Vis Diffuse Reflectance spectra of: (a) the K6P2W18O62 sample after calcination, (b) the Keggin K3PW12O40 sample after calcination, (c) the 1^6^2^18062 sample after reaction under conditions at which the catalyst showed very poor activity (8% isobutane in feed), (d) the K6P2W18O62 sample after reaction under conditions at which the catalyst was active (42% isobutane in feed), and (e) the K6P2W18O62 sample after pre-reduction treatment in diluted hydrogen. Also shown is the UV-Vis spectra if) of the Keggin catalyst after reaction. In all cases the catalyst was unloaded after cooling in helium, in order to avoid possible modifications in the surface average redox level during cooling. Samples (a) and (c) were pale yellow, while sample (d) was light grey. The spectra indicate that the main difference between the spent "active" WellsDawson catalyst (d) and the "inactive" catalyst (c) is an increase of the absorption in the range 450-900 nm which is observed in the former sample, and which disappears by reoxidation of the catalyst. This absorption is very intense in the Keggin compound after reaction (/), which is more active than the Wells-Dawson catalyst. Octahedral is known to exhibit an absorption centered at around 700 nm (9); this absorption is seen, for instance, in spectrum (e) relative to the sample reduced in hydrogen, while it is not present in spent samples. In addition, catalysts after reaction did not exhibit the blue color typical of reduced heteropolyblues. However, when the oxidized catalyst was subjected to a reducing treatment in a diluted H 2 stream with increasing temperature, and the corresponding UV-Vis spectra were recorded, it was found that in correspondence with the progressive reduction of tungsten, the spectrum first exhibited an increase in the absorption at wavelengths higher than 450 nm, then only at temperatures higher than 350°C the absorption centered at 700 nm appeared. In addition, a change in the relative intensities of the bands in the 250 to 450 nm range occurred. During heating, the sample turned from the original pale yellow colour to a grey-yellow colour (after reduction at temperatures higher than 150°C but lower than 300°C), to a pale green and, only after prolonged reduction at 400°C, the compound was definitely blue. This behavior is likely due to the fact that the first electron(s) furnished to the compound are delocalized throughout the structure; therefore, entities are only formed when an energetic barrier to electron derealization is established (9), that is for high reduction degrees. These data support the hypothesis that one difference between the "active" and the "inactive" catalyst is the extent of surface reduction which develops under the different operating conditions. Another aspect concerns the significant differences in the charge-transfer spectral region between the calcined Keggin and Wells-Dawson samples (samples (a) and (&)), which can be tentatively assigned to differences in the coordination sphere of tungsten at the surface of the crystallites.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

150

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

HETEROGENEOUS HYDROCARBON OXIDATION

200

400

600

800 nm

Figure 8. UV-Vis Diffuse Reflectance spectra of fresh and used catalysts. See the text for further explanation.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

10.

Selective Oxidation of Isobutane to Isobutene

CAVANI ET AL.

151

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

2) A second possibility to explain the observed autocatalytic effect is that an increase in isobutane concentration might favor the formation of carbonaceous substances, which may act themselves as active sites. However, this hypothesis can be ruled out due to the fact that an increase in activity is also observed on increasing the oxygen partial pressure (Figure 7), conditions which should greatly hinder formation of active surface carbon. 3) A third possible explanation is that surface-initiated, homogeneous, radical reactions may occur at high isobutane and high oxygen concentration, analogously to what has been reported in the case of methane oxidative coupling and of ethane and propane oxydehydrogenation to the corresponding olefins (10-13). In this hypothesis, after the activation of the hydrocarbon by the catalyst, the radical intermediate can be ejected into the gas phase and be converted to the olefin via a radical chain-reaction, following well-known mechanisms (14): ( C H ) C H + cat --> [ ( C H 3 ) 0 ] i s + cat-H [(CH ) 0] -> ( C H ) 0 ( C H ) 0 + 0 --> ( C H ) C = C H + H 0 * H 0 # + ( C H ) C H --> HOOH + ( C H ) 0 HOOH - > 2 OH* ( C H ) C H + OH* --> ( C H ) 0 + H 0 etc. 3

3

3

3

3

3

a d s

3

2

2

3

3

3

3

3

ac

3

2

2

2

3

3

3

3

3

2

These reactions sometimes exhibit an order higher than 1 with respect to the hydrocarbon. In this explanation for the observed autocatalysis, at low hydrocarbon concentration there is insufficient isobutane in the gas phase to sustain the homogeneous chain reaction, and the mechanism is mainly heterogeneous; under these particular conditions, the catalyst shows, therefore, a poor activity. Moreover, surfaceinitiated, homogeneous radical-type mechanisms in light paraffins oxydehydrogenation are known to be more selective than completely heterogeneous mechanisms (10,13). The autocatalytic effect observed might also be due to a contribution of isobutene in the chain-propagation, due to the enhanced reactivity of the weaker allylic C-H bond (86 kcal/mol) with respect to the stronger C-H bond in isobutane (91 kcal/mol). In summary, the differences observed between the Keggin and the WellsDawson catalyst may be due not only to the different surface area, but also to differences in the reactivity of tungsten. The Keggin compound is characterized by an higher activity and lower selectivity, its catalytic performance is less sensitive to variations in the feedstock composition, and the operating mechanism is also less dependant upon the reaction conditions. On the other hand, different surface properties may lead to a more selective mechanism, as in the case of the Wells-Dawson compound. The proposed working mechanism of isobutane oxidation on ^^^\%0^2 K3PW12O40 catalysts

an(

*

The data obtained indicate that the operating mechanism is a sensitive function of the reaction conditions; in addition, it is likely that more than one of the possibilities

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

152

HETEROGENEOUS HYDROCARBON OXIDATION

O n the oxidized surface H Q

/

3

H C

H

3

H

C

C - C ^

3

H

~

H C-C—C^-H 3

'3~ /

H

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

H

+1/2

O,

O n p a r t i a l l y r e d u c e d s u r f a c e , at h i g h r e a c t a n t

H C H H C-C —C—H H H 3

H

C

3

0

°

3

3

H 0 2

3

2

2

—

3

2 OH-;

3

o ^ o

A ?

_ l 2) - H 0 -

2

2

3

—• H 0 2

2

3

2

2

+ H0 2

+ (CH ) C3

(CH ) CH + HO 3

3

H

—% ( C H ) C O O - — * ( C H ) C = C H 3

C

3 \

H C-C—CH

H

\

t

(CH ) CH + H0 3

H

?

1) + 0

(CH ) C-

c

H

H

/ r

3

3

"

H C _ V /

concentration

3

• H 0 + (CH ) C. 2

3

3

etc.

Figure 9. Proposed working mechanism for the oxydehydrogenation isobutane to isobutene.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

of

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

10.

CAVANI ET AL.

Selective Oxidation of Isobutane to Isobutene

153

above reported could contribute to the observed behavior. Figure 9 shows the proposed reaction mechanism, one that is summarized in the discussion which follows. This mechanism is proposed with insights from solution spectroscopic studies of reduced, oxygen-deficient forms of heteropolyoxomolybdates and tungstates (the so-called "heteropolyblues") (75-77), which generally show adjacent M - O - M sites (M=W, Mo, V and others) in close electronic communication (15-19). We propose that isobutane is activated on the surface through the formation of an adsorbed species which, according to literature indications (20), may be an alkoxy species. The extent of catalyst reduction affects the activity of the surface sites responsible for isobutane activation: a partially reduced surface leads to more active sites, but the reaction is less selective, possibly due to a lower availability of surface oxygen, which in turn does not allow the selective transformation of the adsorbed intermediate into the olefin. Under these conditions the intermediate evolves towards the formation of carbon oxides, possibly by oxidative attack from gaseous oxygen. This is confirmed by the performance of the pre-reduced compound, as well as by the results obtained by co-feeding isobutane and an olefin: paraffin conversion was increased but with formation of carbon oxides only. On the contrary, an oxidized surface is less active, but favours the heterogeneous transformation of the adsorbed intermediate species to isobutene. At isobutane concentration lower than the 15-20%, the Wells-Dawson compound shows poor activity, much lower than that of the Keggin catalyst. This is due to the lower surface area and, likely, also to the different reducibility of the tungsten ion. Under these conditions the surface of the Wells-Dawson catalyst is oxidized (as it was after the calcination treatment); this oxidized surface is less active, but it is selective. When the paraffin concentration (or the oxygen concentration) in feed is increased, the concentration of isobutene in the gas phase is also increased. Under these conditions, two main effects are likely obtained: i) on one hand, the higher reducing power of the isobutene makes the tungsten more reduced, thus yielding a more active catalyst; ii) on the other hand, the high hydrocarbon concentration in the gas phase makes the contribution of the homogeneous selective mechanism of isobutene formation the prevailing one. A reduced catalyst may initiate more radicals, increasing the amount of gas-phase homogeneous reactions. In other words, under increased paraffin concentration the overall activity (i.e., the step of isobutane activation) may be governed by the higher catalyst reduction degree, while the selectivity is mainly governed by the homogeneous gas-phase mechanism. The fundamental role of isobutene (which may also participate in the gas phase mechanism, due to its more labile C-H bond) yields the observed autocatalytic effect. Acknowledgements. MURST and CNR (Progetto Strategico Tecnologie Chimiche Innovative) are gratefully acknowledged for financial support. Literature Cited

1 2

Misono, M.Catal.Rev. Sci. Eng. 1987, 29, 269. Ono, Y. In Perspectives in Catalysis; Thomas, J.M., Zamaraev, K.I., Eds.; Blackwell Scientific Publ.: Oxford, 1992; p. 431.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

154

3 4 5 6 7 8

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

9 10 11 12 13 14 15 16 17 18 19 20

HETEROGENEOUS HYDROCARBON OXIDATION

Mizuno, N.; Lyon, D. K.; Finke, R. G. J.Catal.1991, 128, 84. Mansuy, D.; Bartoli, J. F.; Battioni, P.; Lyon, D. K.; Finke, R. G. J. Am. Chem. Soc. 1991, 113, 7222. Kung, H. H. Adv.Catal.1994, 40, 1. Cavani, F.; Etienne, E.; Favaro, M.; Galli, A.; Trifirò, F.; Hecquet, G.Catal.Lett. 1995, 32, 215. Yamamatsu, S.; Yamaguchi, T. Eur. Patent 425,666 A1, 1989, assigned to Asahi Chem. Co. Lyon, D. K.; Miller, W. K.; Novet, T.; Domaille, P. J.; Evitt, E.; Johnson, D. C.; Finke, R. G. J. Am. Chem. Soc. 1991, 113, 7209. Pope, M. T. Heteropoly and Isopoly Oxometalates; Springer Verlag; New York; 1983. Cavani, F.; Trifirò, F.;Catal.Today 1995, 24, 307. Burch, R.; Crabb, E. M. Appl.Catal.A: General 1993, 100, 111. Burch, R.; Chalker, S.; Loader, P.; Thomas, J.; Ueda, W.Appl.Catal., A: General 1992, 82, 77. Kennedy, E. M.; Cant, N. W. Appl.Catal.1991, 75, 321. Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic Compounds; Academic Press; New York; 1981. Pope, M. T., In Mixed-valence Compounds; Brown, D. B., Ed.; D. Reidel Publishing: Dordrecht, 1980; p. 365. Prados, R. A.; Pope, M. T. Inorg. Chem. 1976, 15, 2547. Kazansky, L. P.; Fedotov, M. A. J. Chem. Soc. Chem. Commun. 1980, 644. Piepgrass, K.; Pope, M.T. J. Am. Chem. Soc. 1987, 109, 1586. Kawafune, I. Chem. Lett. 1989, 185. Finocchio, E.; Busca, G.; Lorenzelli, V.; Willey, R. J. J.Catal.1995, 151, 204.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

154

3 4 5 6 7 8

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

9 10 11 12 13 14 15 16 17 18 19 20

HETEROGENEOUS HYDROCARBON OXIDATION

Mizuno, N.; Lyon, D. K.; Finke, R. G. J.Catal.1991, 128, 84. Mansuy, D.; Bartoli, J. F.; Battioni, P.; Lyon, D. K.; Finke, R. G. J. Am. Chem. Soc. 1991, 113, 7222. Kung, H. H. Adv.Catal.1994, 40, 1. Cavani, F.; Etienne, E.; Favaro, M.; Galli, A.; Trifirò, F.; Hecquet, G.Catal.Lett. 1995, 32, 215. Yamamatsu, S.; Yamaguchi, T. Eur. Patent 425,666 A1, 1989, assigned to Asahi Chem. Co. Lyon, D. K.; Miller, W. K.; Novet, T.; Domaille, P. J.; Evitt, E.; Johnson, D. C.; Finke, R. G. J. Am. Chem. Soc. 1991, 113, 7209. Pope, M. T. Heteropoly and Isopoly Oxometalates; Springer Verlag; New York; 1983. Cavani, F.; Trifirò, F.;Catal.Today 1995, 24, 307. Burch, R.; Crabb, E. M. Appl.Catal.A: General 1993, 100, 111. Burch, R.; Chalker, S.; Loader, P.; Thomas, J.; Ueda, W.Appl.Catal., A: General 1992, 82, 77. Kennedy, E. M.; Cant, N. W. Appl.Catal.1991, 75, 321. Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic Compounds; Academic Press; New York; 1981. Pope, M. T., In Mixed-valence Compounds; Brown, D. B., Ed.; D. Reidel Publishing: Dordrecht, 1980; p. 365. Prados, R. A.; Pope, M. T. Inorg. Chem. 1976, 15, 2547. Kazansky, L. P.; Fedotov, M. A. J. Chem. Soc. Chem. Commun. 1980, 644. Piepgrass, K.; Pope, M.T. J. Am. Chem. Soc. 1987, 109, 1586. Kawafune, I. Chem. Lett. 1989, 185. Finocchio, E.; Busca, G.; Lorenzelli, V.; Willey, R. J. J.Catal.1995, 151, 204.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

154

3 4 5 6 7 8

Downloaded by UNIV LAVAL on July 3, 2014 | http://pubs.acs.org Publication Date: August 13, 1996 | doi: 10.1021/bk-1996-0638.ch010

9 10 11 12 13 14 15 16 17 18 19 20

HETEROGENEOUS HYDROCARBON OXIDATION

Mizuno, N.; Lyon, D. K.; Finke, R. G. J.Catal.1991, 128, 84. Mansuy, D.; Bartoli, J. F.; Battioni, P.; Lyon, D. K.; Finke, R. G. J. Am. Chem. Soc. 1991, 113, 7222. Kung, H. H. Adv.Catal.1994, 40, 1. Cavani, F.; Etienne, E.; Favaro, M.; Galli, A.; Trifirò, F.; Hecquet, G.Catal.Lett. 1995, 32, 215. Yamamatsu, S.; Yamaguchi, T. Eur. Patent 425,666 Al, 1989, assigned to Asahi Chem. Co. Lyon, D. K.; Miller, W. K.; Novet, T.; Domaille, P. J.; Evitt, E.; Johnson, D. C.; Finke, R. G. J. Am. Chem. Soc. 1991, 113, 7209. Pope, M. T. Heteropoly and Isopoly Oxometalates; Springer Verlag; New York; 1983. Cavani, F.; Trifirò, F.;Catal.Today 1995, 24, 307. Burch, R.; Crabb, E. M. Appl.Catal.A: General 1993, 100, 111. Burch, R.; Chalker, S.; Loader, P.; Thomas, J.; Ueda, W.Appl.Catal., A: General 1992, 82, 77. Kennedy, E. M.; Cant, N. W. Appl.Catal.1991, 75, 321. Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic Compounds; Academic Press; New York; 1981. Pope, M. T., In Mixed-valence Compounds', Brown, D. B., Ed.; D. Reidel Publishing: Dordrecht, 1980; p. 365. Prados, R. A.; Pope, M. T. Inorg. Chem. 1976, 15, 2547. Kazansky, L. P.; Fedotov, M. A. J. Chem. Soc. Chem. Commun. 1980, 644. Piepgrass, K.; Pope, M.T. J. Am. Chem. Soc. 1987, 109, 1586. Kawafune, I. Chem. Lett. 1989, 185. Finocchio, E.; Busca, G.; Lorenzelli, V.; Willey, R. J. J.Catal.1995, 151, 204.

In Heterogeneous Hydrocarbon Oxidation; Warren, B., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.