Catalytic combustion of methane over transition metal oxides.

Catalytic combustion of methane over transition metal oxides.

NATURAL GAS CONVERSIONV Studies in Surface Science and Catalysis, Vol. 119 A. Parmaliana et al. (Editors) o 1998 Elsevier Science B.V. All rights rese...

463KB Sizes 0 Downloads 4 Views

Recommend Documents

No documents
NATURAL GAS CONVERSIONV Studies in Surface Science and Catalysis, Vol. 119 A. Parmaliana et al. (Editors) o 1998 Elsevier Science B.V. All rights reserved.

65

Catalytic c o m b u s t i o n o f m e t h a n e over transition metal oxides. S. Arnone a, G. Bagnasco ~, G. Busca b, L. Lisi~, G. Russo ~, M. Turco ~. aDipartimento di Ingegneria Chimica, Universitb. "Federico II", Napoli, Italy. bIstituto di Chimica, Facolt~. di Ingegneria, Universit/l di Genova, Genoa, Italy. ~Istituto di Ricerche sulla Combustione, CNR, Napoli, Italy.

Simple and mixed metal oxides containing Co, Mn, Cr and Fe have been investigated as catalysts for the combustion of methane in the temperature range 300-600~ under diluted conditions. The effect of the catalyst composition on the catalytic performances and on the redox properties has been evaluated. Single metal oxides containing Cr, Co and Mn show comparable activities and were found more active than Fe/O3. Mixing of Co with Cr oxide and Fe oxide with Mg and Zn to give spinels, improves the catalytic activity with respect to pure compounds. Temperature programmed reduction (TPR) shows that redox properties are strongly dependent on the catalyst composition. Fe based mixed oxides are more hardly reducible than the other catalysts, this effect being related to the dilution of Fe with bivalent cations. The comparison of the kinetic parameters, evaluated on the base of a first order rate equation, gave evidence of a correlation between the activation energy values and the ease of the reduction showing that the oxides reducible at lower temperature give rise to a reaction mechanism with a lower activation energy.

1. I N T R O D U C T I O N Catalytic combustion has been proposed as an alternative technique to the homogeneous combustion for several applications like as gas turbines, boilers, aircrafts, afterburners, domestic heaters, VOC removal. The process can be carried out in a wide range of fuel/air ratio at low temperature thus leading to a marked reduction of NOx emission levels [1]. Different systems are reported in literature as active catalysts in methane combustion such as noble metals [2] and perovskite oxides [3, 4]. Nevertheless, less attention has been paid to transition metal oxides with different structures such as spinels, mostly because some of them are instable at high temperature. Nevertheless, some spinels are very active and the study of the factors affecting their activity can give light on the mechanism involved in total combustion catalysis [5]. In this paper different transition metal oxides have been studied as methane combustion catalysts. The effect of the partial substitution of the metal cation was also investigated. TPR technique was employed for the characterization of the catalysts with the aim to define a correlation between redox properties and catalytic activity.

66 2. E X P E R I M E N T A L Single and mixed transition metal oxides were prepared by precipitation, except Zn and Mg ferrites that were prepared as aerogels and Co304 and y-Fe203 that were commercial materials. XRD analysis was performed using a Philips PW 1710 diffractometer. Specific surface area of catalysts was determined by N2 adsorption at 77K according to the BET method using a Carlo Erba 1900 Sorptomatic apparatus. Temperature programmed reduction (TPR) experiments were carried out in a Micromeritics 2900 TPD/TPR flow system equipped with a TCD. After treatment in air flow at 600~ samples were reduced with a 2% H2/Ar mixture (25 cm 3 min"1) at heating rate of 10~ min1 up to 600~ Catalytic combustion of methane was studied in a fixed bed quartz micro-reactor. The catalyst (particle size = 300-4001am), diluted 1:10 in quartz powder, was placed on a frit disk. Quartz pellets upside the catalytic bed and a narrowing of the reactor section both in the post-catalytic and pre-catalytic zone reduced the homogeneous volume. A thermocouple placed in the catalytic bed, allowing the monitoring of the temperature during the reaction, showed that the maximum temperature gradient was lower than 5~ The feed composition was 0.4% CH4 and 10% 02 in a balance of N2. A constant space velocity of 40000 cm 3 hx g~ was ensured by Brooks 5850 TR Series mass flow controllers. Catalytic tests were carried out in the temperature range 300-600~ The concentration of reactants and products was measured using a Hewlett Packard 6890 gas-chromatograph equipped with two capillary columns (a poraplot Q and a molecular sieve 5A) and thermal conductivity and flame ionization detectors. Carbon balance was verified within +_5 %. 3. RESULTS AND DISCUSSION

3.1. Physico-chemical characterization The crystalline phases and the values of the specific surface area of simple and mixed oxides are reported in Table 1. Table 1 XRD phases, specific surface areas and thermal behaviour of the catalysts. Catalyst XRD phase Specific surface Thermal behaviuor area (m 2 8 "1) Co304 normal spinel 15 to CoO at T>450~ 24 Mn304 random spinel to a-Mn203 at T=400~ to Mn304 at T=970~ Cr203 corundum 18 thermodynamically stable y-Fe203 non stoichiometric spinel 22 to a-Fe203 at T>650~ thermodynamically stable ct-Fe203 corundum 102 CoCr204 normal spinel 110 thermodynamically stable ZnFe204 normal spinel 27 thermodynamically stable Mg0.sZn0.sFe204 random spinel 37 thermodynamically stable MgFe204 inverted spinel 56 thermodynamically stable

67 All samples show the spinel structure except Cr203 and one of the ferric oxides that crystallize in the corumdum structure. Simples oxides have a surface area of about 20 m2 g-I, except ot-Fe203, whereas higher values of surface area are shown by the mixed oxides. For Fe based oxides, the dilution with Mg leads to a marked enhancement of the surface area of the simple oxide, a lower increase is related to the introduction of the Zn cation. TG/DTA experiments show that all samples are structurally stable up to 1000~ except C0304, Mn304 and y-Fe203 (Table 1). TPR profiles of simple oxides are reported in Figure 1. TPR curve of CrzO3 sample has not been reported due to the very low H2 uptake compared to that of the other simple oxides. /

2.0 -I " " " I

"-"

l

A

........ Mn30,

% J -'= 1 . 5 _

F:

% 1.0

700

C0304

!"

(z_Fe203

_

/

: 7 -_

I

600

'

!

!

,

I

.,.;

700

Mgo.sZno.sFe204 ] .........

600

ZnFe20'/'~

3.O

I

::>

500 5" F~

o._., 400 e -~

400

300

~

,-

300 ~. E

s 0.5 -

t

200 ~.

.....,i

co ~_1.0

200 ~.

loo

-10.0 0

500 5`

~:) 2.0

x

~

%0)

~>

;/,

7-Fe203

4.0 .--.

I

I

1

I

20

40

60

80

0 100

100 0.0

~" 0

Time (min)

20

40

60

80

100 120 140

Time (min)

Figure 1. TPR profiles of simple and mixed oxides. In Table 2 the H2 consumed in the TPR experiments, the onset temperature (Tonset) and the temperature corresponding to the maximum uptake (Tmax) are reported for all samples. Table 2 Results of TPR experiments. Catalyst H2 uptake Tonset (mol H2 mol 1 M*) (~ Co304 1.2 264 Mn304 0.5 187 Cr203 0.025 182 y-Fe203 0.3 373 ct-Fe203 0.3 336 CoCr204 0.023 185 ZnFe204 1.0 284 Mgo.sZno.sFez04 0.6 236 M~Fe204 0.8 248 * M in mixed oxides refers to the total metal content.

Tmax

(~ 382, 470 385, 520 295, 470 448, 535 460, 520 255 -

The reduction occurs in two or more steps for all oxides starting at quite low temperature for Co304, Mn304 and Cr203 simple oxides and is complete within 600~ For the two Fe203 samples the shift of the baseline at 600~ suggests that the the reduction is still continuing isothermally at this temperature. The extent of the reduction is markedly affected by the metal cation only cobalt undergoing a deep reduction. The values of H2/M ratio suggest that Co cations in Co304 undergoes the complete reduction to Co ~ The presence of metallic Co was

68 confirmed by XRD analysis carried out after the TPR experiment. Thus, the first peak could correspond to the reduction from the average oxidation state 2.7+ to 2+ and the second one to the reduction from 2+ to metallic oxidation state. The XRD spectra taken on the M n 3 0 4 after TPR experiment show the signal of MnO phase, suggesting that manganese is reduced to 2+ oxidation state. Taking into account this result an average Mn initial oxidation state of 3+ can be evaluated from HE consumption, then higher than that expected from the stoichiometry of the compound. This suggests that the oxidation to M n 2 0 3 can occur during the pretreatment, in agreement with literature data reporting that M n 3 0 4 undergoes the transition to M n 2 0 3 in oxidizing atmosphere at about 600~ [6]. A further confirmation was also obtained by XRD analysis effected on the sample after the first TPR peak, showing the signals of M n 3 0 4 phase. For Cr203 sample the very low extent of the reduction make uncertain the determination of the exact stoichiometry of the final compound. Finally, both Fe203 samples are reduced to Fe304 as suggested by the value of H2/M ratio. In CoCr204 sample the H2 uptake is strongly reduced with respect to C0304 sample and is very close to that observed for Cr203. Moreover, a shift of Tmaxin compared to the that of pure compound was observed. A different behaviour was shown by Fe based mixed oxides that need higher temperatures to activate the reduction. As shown by TPR profiles reported in Figure 1 the reduction shows the maximum rate at temperatures approaching 600~ therefore higher than the other catalysts, and continues isothermally at this temperature. This suggests that the dilution of Fe with lower valence cations makes the mixed oxides more hardly reducible even if the extent of the Fe reduction increases compared to the simple Fe oxides. It can be supposed that Mg and Zn, being stable in 2§ oxidation state, do not undergo reduction therefore the H2 uptake can be due to the reduction of Fe cation only. After TPR experiments the samples were treated in air flow at 600~ and reduced again under the same conditions of the first experiments. The reduction-oxidation process was found reversible for all oxides except for Fe based sample. In this case a shift of Ton~ot and Tmax and a modification of the intensity of the signals were observed. 3.2.

Catalytic

activity

tests

Preliminary tests, performed under the same conditions of the catalytic tests, but without catalyst, showed that homogeneous reactions are negligible under the experimental conditions investigated. The results of the catalytic activity tests are reported in Figure 2. 100

~-.

100

-

v

to

-

80

-

60

-

40

-

20

-

0

-~ v

60~

I--

(D > to ~-)

V../? t

t

80-

k,,.

40-

12 200 I

I

I

I

I

I

I

I

300

400

500

600

300

400

500

600

Temperature ( ~

tO

~>

tO 0

Temperature (~

Figure 2. CH4 conversion as a function of temperature for y-Fe203 (A), ot-Fe203 (Y), MgFe204 (o), ZnFe204 (O), Mgo.sZno.sFe204 (O), MnaO4 (@), Cr203 (!"!),C0304 (ll), CoCr204 (~).

69 All catalysts, except y-Fe203, give complete conversion of methane within 600~ with 100% selectivity to CO2. The catalysts are able to activate the reaction in a temperature range lower than that of perovskite oxides [3] and comparable to that of noble metals [7]. Cr203, C0304 and Mn304 show a comparable activity. Fe203 is the less active when it crystallizes in a non stoichiometric spinel structure. The substitution of the metal cation enhances the activity of both Fe and Cr based catalysts. In the Fe based oxides, the mixing with Mg oxide gives rise to a larger effect compared to that due to mixing with Zn. The ternary system, obtained by the partial substitution of Zn with Mg, has an activity higher than ZnFe204 but comparable to that of MgFe204. After a first cycle of tests all catalysts were cooled down to room temperature and a new cycle of experiments was performed. The results of the second cycle were the same of the first one for all catalysts except for Fe based samples that gave rise to some loss of activity suggesting that these oxides undergo a deactivation under the reaction conditions. Catalytic activity data were elaborated assuming a methane first order rate equation [8] and a plug flow integral reactor. CH4 conversions ranging from 10 to 90% were used to evaluate the values of activation energy and preexponential factor reported in Table 3. The activation energy is about 20 Kcal mol 1 for Mn, Co, Cr and Fe single oxides and for CoCr204, however, a higher value of the activation energy was evaluated for Fe mixed oxides. This result suggests that the dilution of Fe 3+ with a bivalent cation can modify the mechanism of methane activation. The higher activity of CoCr204 catalyst (Figure 2) can thus be due to the greater value of the surface area shown by this sample as can be demonstrated by the value of preexponential factor referred to the catalyst specific surface comparable to that of Mn, Co and Cr simple oxides. The comparison of the preexponential factors of simple oxides suggests that Fe203 oxides exhibit the lowest surface sites concentration. Likewise, the best catalytic performances of ct-Fe203 in respect with y-Fe203 could be associated to the higher surface area of our corundum type sample more than to an effect of the different structure of this oxide. Fe mixed oxides show the highest activation energy value despite of their catalytic activity is comparable to that of other catalysts, and significantly higher than that of y-Fe203. This effect is due to the higher values of preexponential factors referred to surface area induced by the Fe dilution with Mg or Zn.. Table 3 Activation energy (Ea) and preexponential factor (A). Catalyst Ea A x 10-s (1 h"l g-l) (Kcal molq ) Co304 20 0.4 Mn3Oa 20 O4 Cr203 20 O5 0.07 y-Fe203 20 0.2 ot-Fe203 20 CoCr204 20 2.9 ZnFe204 30 8O Mg0.sZn0.sFe204 30 395 M~Fe204 30 350

A x 10"s (1 h"1 m"E) 0.029 0.017 0.027 0 003 0 002 0 027 3 11 6

On the base of the above results a correlation between the catalytic activity and the redox behaviour can be drawn. If a relationship between the extent of the reduction seems to be excluded, a correlation between the ease of reducibility and the activation energy appears quite reasonable. Catalysts which are reduced within 600~ show the same value of

70 activation energy. By contrast, mixed Fe based catalysts whose reduction is delayed, as the maximum H2 uptake occurs at temperatures approaching 600~ and the process continues isothermally, show the same activation energy value, higher then that of the previous materials. This suggests that the availability of the surface lattice oxygen significantly affects the catalytic properties in activating methane oxidation. It is reported that the catalytic activity in the total oxidation of methane is strongly related to the oxidation properties of the catalysts, the surface oxygen being involved in the reaction mechanism [1]. Moreover, the disactivation observed in both reduction with H2 and CI-h oxidation processes for Fe based oxides gives a further confirmation of the correlation between redox and catalytic properties, suggesting that the reversibility of the reduction process is an important feature for catalysts that could be employed in the catalytic combustion of methane. 4. CONCLUSIONS Simple and mixed oxides activate the oxidation of methane in a temperature range comparable to that of noble metals and lower than of perovskite oxides ensuring a 100% selectivity towards the total oxidation products. All simple oxides catalyse the methane oxidation activating the same reaction mechanism not depending on the nature of the transition metal. They show a comparable density of active sites except FezO3 oxides which have a lower concentration of surface sites. The partial substitution of Cr 3+ with Co 3+ leads to an increase of catalytic activity attributed to the enhancement of the specific surface area. On the contrary, the dilution of Fe with bivalent cations results in a different reaction mechamism and, at the same time, in an increase of surface sites concentration. The evaluation of the redox properties by TPR analysis showed a close correlation between the range of temperature in which the reduction occurs and the activation energy of methane oxidation estimated for the metal oxides catalysts. REFERENCES 1. M. F. M. Zwinkels, S. G. Jaras, P. G. Menon and T. A. Griffin, Catal. Rev. Sci. Eng., 35 (1993) 319. 2. R.Prasad, L.A. Kennedy and E. Ruckenstein, Catal. Rev. Sci. Eng., 26(1) (1984) 1. 3. L.G. Tejuca, J.L.G. Fierro and J.M.D. Tascon, Adv. Catal., 36 (1989) 37. 4. P. Ciambelli, L. Lisi, G. Minelli, I. Pettiti, P. Porta, G. Russo and M. Turco, Proceedings of 3rd World Congress on Oxidation Catalysis, San Diego, 1997. 5. R.Prasad, L.A. Kennedy and E. Ruckenstein, Comb. Sci. Tec., 22 (1980) 271. 6. M. Baldi, E. Finocchio, F. Milella and G. Busca, Appl. Catal.: B. Environ., in press. 7. R. Burch and P. K. Loader, Appl. Catal. B: Environ., 5 (1994) 149. 8. H. Arai, T. Yamada, K. Eguchi and T. Seiyama, Appl. Catai., 26 (1986) 265.