Metal-containing molecular sieves as catalysts for the selective oxidation of cyclohexane by oxygen

2674

Studies in Surface Science and Catalysis, volume 154 E. van Steen, L.H. Callanan and M. Claeys (Editors) 9 2004 Elsevier B.V. All fights reserved.

M E T A L - C O N T A I N I N G M O L E C U L A R SIEVES AS CATALYSTS FOR THE SELECTIVE O X I D A T I O N OF C Y C L O H E X A N E BY OXYGEN Tian, P., Liu, Z.*, Wu, Z., Xu, L. and He, Y. Natural Gas Utilization and Applied Catalysis Laboratory, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, P. R. China. Fax: 86- 411-4691570. *E-mail: [email protected]

ABSTRACT Three types of metal-containing molecular sieves with AFI, AEL and CHA structures were hydrothermally synthesized. Metals (Co, Mn, Cr, V) were incorporated into the molecular sieve framework and induced the appearance of charge centres, as evidenced by TG and NH3-TPD. TPR studies revealed that cobalt and manganese in the framework were not reducible before the collapse of molecular sieve structure. The catalytic behaviours of these catalysts in cyclohexane oxidation at 403K and 1.1MPa 02 were investigated. CoAPO-I 1 exhibited best activity and good selectivity of monofunctional oxidation products (88.5%). Cyclohexanol was the preferable product on most catalysts, whereas for CrAPO-5, -11, high selectivity of cyclohexanone was observed. Experimental results indicated that the decomposition of cyclohexyl hydroperoxide (CHHP), the intermediate in cyclohexane oxidation, followed the pathway: cyclohexanone~--CHHP---,cyclohexanol (then cyclohexanol--~cyclohexanone). K e y w o r d s : metal-containing aluminophosphate, characterization, oxidation, cyclohexane INTRODUCTION The partial oxidation of hydrocarbons using oxygen as oxidant is significant and preferred to the chemical industry [1,2]. Among various alkanes oxidation, the selective oxidation of cyclohexane is attractive because its product is the intermediate for the production of Nylon-6 and Nylon-6-6 [3]. Soluble cobalt salts (such as cobalt naphthenate) are often used as oxidation catalysts in industry. However, conversion is typically low in order to maintain high selectivity (conversion<5%, selectivity>80%) in the oxidation processes. Metal-containing molecular sieves have attracted great interests as catalysts since their appearance due to their acidity, redox, shape-selectivity and recyclable properties [4,5]. Up to now, many porous metal-containing molecular sieves with various structures have been synthesized and used as catalysts such as MeAPO, MeS-1 and MeMCM-41 (Me=Co, Mn, V, Cr, Fe, Ti etc.) [6]. Weng et al. first reported the use of CoAPO-5 as catalysts for the cyclohexane oxidation with acetic acid as solvent [7]. Unfortunately, acetic acid as solvent led to considerable leaching of cobalt from the framework. Tomas et al. published interesting works upon the MeAPOs as effective catalysts for the selective cyclohexane oxidation [8,9]. They found that molecular sieves structure influenced the metal valence in the calcined samples and caused different alkane oxidation activity. In this paper, three types of metal-containing molecular sieves were synthesized. TG, TPD of NH3 and TPR were used to investigate the physicochemical properties of MeAPOs. Their catalytic behaviours in cyclohexane oxidation were studied using oxygen as oxidant.

EXPERIMENTAL Synthesis MeAPO-5, MeAPO-I1 and MeAPSO-34 were hydrothermally synthesized according to the literatures [10,11]. The aluminum and phosphorus sources were pseudoboehmite and H3PO4 (85%). Cobalt (2+) acetate, manganese (2+) acetate, chromium (3+) nitride, vanadyl acetate, iron (2+) sulphate and titanium (4+) sulphate were used as metal sources, respectively. The templating agents were triethylamine for MeAPO-5 and MeAPSO-34 and n-dipropylamine for MeAPO-11. The synthesis conditions were given in Table 1. As-synthesized molecular sieves were calcined in air at 823 K for 5h to remove organic template.

2675

Table 1. Gel compositions and metal content in the molecular sieves. Sample A1PO-5 CoAPO-5 MnAPO-5 CrAPO-5 VAPO-5(1) VAPO- 5(2) FeAPO-5 TiAPO-5 A1PO-11 CoAPO-11 MnAPO-11 CrAPO-11 VAPO- 11 CoAPSO-5 CoAPSO-11 CoAPSO-34 MnAPSO-34

Gel compositions R A1203 P205 SiO2 MeO 1.2 1 1 0 0 1.2 1 1 0 0.05 1.2 1 1 0 0.05 1.2 1 1 0 0.05 1.2 1 1 0 0.05 1.2 1 1 0 0.25 1.2 1 1 0 0.05 1.2 1 1 0 0.05 1.0 1 1 0 0 1.0 1 1 0 0.05 1.0 1 1 0 0.05 1.0 1 1 0 0.05 1.0 1 1 0 0.05 1.2 1 1 0.4 0.055 1.1 1 1 0.4 0.055 3 1 1 0.2 0.08 3 1 1 0.2 0.05

H20 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50

Crystallization Time (h) T (K) 40 448 40 448 40 448 40 448 40 448 40 448 40 448 40 448 40 473 40 473 40 473 40 473 40 473 24 473 72 473 48 473 48 473

Me (wt%) 1.62 1.76 1.01 0.37 1.07 1.82 0.49 1.14 1.25 1.08 0.66 1.58 1.62 2.2 1.32

Characterization XRD measurement was carried out using Rigaku X-ray diffractometer (Model D/Max-~/B) with Cu-K~ radiation. The metal content in the molecular sieves was measured by XRF (Philips Magix). TG analysis was recorded on Perkin Elmer Pyriss TGA with a heating rate of 5K/min from 323k to 973K under air (35ml/min). TPD of NH3 and TPR experiments were carried out with Autowin 2910 equipment (Micromeritics). All of the samples were first activated at 873 K (10 K/min) for 30rains under He (20mL/min). For NH3-TPD, ammonia was injected to saturate the sample at 383K. The measurement of desorbed NH3 was performed from 383 K to 823 K (10 K/rain) under He (40mL/min). For TPR, H2 (10%)/Ar (40mL/min) mixture was used and the reactor was heated from room temperature to 1023K (10K/min).

Catalytic reaction Cyclohexane oxidation was carried out in a 100-ml autoclave lined with polytetrafluoroethylene (PTFE). Typically, 18g cyclohexane, about 0.7g catalyst (0.195mmol metal) and 0.04g tetrabutyl hydroperoxide (TBHP, used as initiator) were introduced into the reactor. After charged 1.1MPa of O2, the reactor was heated to 403K under stirring for 3h. When the reaction was finished, 20ml ethyl alcohol was added into the cold reactor to dissolve carboxylic acids. The catalyst was separated by centrifugation. The reaction products were analyzed on a Varian CP-3800 gas chromatograph equipped with an FFAP capillary column (0.53mm*25m) and an FID detector. The conversion was calculated based on the starting cyclohexane. As cyclohexyl hydroperoxide (CHHP) partially decomposed upon injection, its concentration was estimated by double analysis, before and after reducing CHHP to cyclohexanol with Ph3P [12]. RESULTS AND DISCUSSION

Characterization XRD studies XRD was used to check the samples purity. All samples exhibited high and pure crystallinity and no extra peak related to transition metal oxides appeared in the XRD patterns. Figure 1 shows the typical XRD patterns of three types of molecular sieves. MeAPO-5 a n d - 1 1 with AFI and A E L topologies have one-dimensional channels with pore openings of 0.73nm and 0.65"0.4nm. MeAPSO-34 with three-dimensional channels has smaller pore (0.38nm) and a structure similar to the one of natural chabazite (CI-~).

2676 The synthesis conditions and metal amounts in the molecular sieves are displayed in Table 1. Most molecular sieves had metal contents in the products corresponding to those in starting gel despite some variation. But for V-, Ti-A1PO-5 and VAPO-11, the metal contents in the products were unexpectedly low. The low Ti content maybe was due to the use of inorganic titanium precursor in the gel. The V content in VAPO-5 could be enhanced from 0.37% to 1.17 wt% by increasing the V/P ratio in the gel from 0.025 to 0.125 (see Table 1), in agreement with the work ofLopez Nieto [13].

C

tl '

, i

10

9

i

1 '

:20

i

'

~10

i

"0

'

I

50

2 X heta

Figure 1. XRD patterns of (a) CoAPO-5 (b) MnAPO-5 (c) CoAPO-11 (d) MnAPO-11 (e) CoAPSO-34.

Thermal analysis The thermogravimetric curves of MeAPO-5, 11 in air are given in Figure 2. Three-step weight losses could be observed in the 323---373, 373---523 and 523~-893K temperature ranges for MeAPO-5 and in the 323~373,373.--543 and 543-~893K ranges for MeAPO-11. The initial step corresponded to water desorption and the second one to template desorption. The third-step weight loss was ascribed to combustion of some organic species occluded in MeAPOs channels. The results of weight losses for each step are collected in Table 2. Table 2. TG results of MeAPO-5, 11. Sample ALP0-5 ....... CoAPO-5 MnAPO-5 CrAPO-5 VAPO-5(2) A1PO- 11 CoAPO- 11 MnAPO- 11 CrAPO-11 VAPO- 11

Weight loss (wt%) !-step ....2-step 3-step 3.3 6.2 2.1 3.2 4 4.2 3.0 3.9 4.6 3.1 4.5 4.6 4.7 3.6 4.3 0.3 6.5 0.7 0.5 4.7 3.9 0.6 4.6 4.3 1.2 5.3 3.3 1.2 5.2 2.5

.

Template weight Total weight loss (wt%) loss (wt%) . . . . . . . . . . . . . . 8.3 11.6 8.2 11.2 8.5 11.5 9.1 12.2 7.9 12.6 7.2 7.4 8.6 9.1 8.9 9.5 8.6 9.8 7.7 8.9

It can be seen that the water content (-3.2wt%) and organic template content (-8.5wt%) remained nearly the same in all MeAPO-5 and A1PO-5 samples. The template amount corresponded to--1.3 molecules of TEA per unit cell. But the proportion of the template removal in 2- and 3-step was very different between AIPO-5 and MeAPO-5s. About 3/4 of template was removed during 2-step in AIPO-5, whereas only 1/2 of template in this step for all MeA1PO-5s. As the charge of A1PO-5 framework is neutral and the organic template is filled physically in the pores, it is expected that most of the template could be eliminated at lower temperatures. The increased weight loss of metal-containing A1PO-5s in 3-step clearly revealed that there were charge centres in the framework, which made TEA molecules bonded in MeAPO-5s strongly, in agreement with published works [13,14]. For A1PO-11 and MeAPO-11s, more template molecules were

2677 occluded in MeAPO-1 ls than in A1PO-11 (see Table 2). However, the difference of template removal in 2- and 3-step between AIPO-11 and MeAPO-1 ls was similar to that between AIPO-5 and MeAPO-5s. It can be inferred that Co, Mn, V, Cr incorporated into the framework during hydrothermal synthesis and induced the appearance of the charge deficiency.

I

..--..

~

4 % ,.--..

v 03

or) or)

~PO-5

_o

4 9

.r

0-1 1

O')

o

..........................CoAP ........O-1 ......1. ........

(13

O)

I

-~.,.~

......................

........................... MnAP 0-1 1

MnAP 0-5

CrAPO-1 1

=

,

400

I

.

500

~-.

CrAPO-5

....

VAP 0-5

I

6OO

,

I

700

,

I

,

BOO

~ _

I

I

,

,

400

QOO

I

,

I

5 0 0 . 600

VAP O-1 1

,

I

700

Temperature

T em perature (K)

,

I

800

,

|

900

(k)

Figure 2. TG curves of as-synthesized metal-containing aluminophosphates.

TPD of ammonia NH3-TPD was used to characterize the acidity of Co, Mn-containing AIPOs. The results are shown in Figure 3. Comparison of the TPD curves of MeAPOs with that of A1PO-5 indicated a increase of the acid strength and amount. The peaks with maxima at about 473K, found for all samples, were due to weakly bond ammonia (physisorbed ammonia, ammonia adsorbed on weak Lewis sites and terminal P-OH groups) [15]. The high temperature peaks centred at 653K, which was not detectable on A1PO-5, could be ascribed to ammonia desorbed from strong acid sites in MeAPOs. This could be taken as proof that charge centres (Brensted and/or Lewis acid sites) existed in the calcined Co, Mn-containing A1PO-5, 11. The strength of acid sites had the following order: CoAPO-11 > MnAPO-11 ~ CoAPO-5 > MnAPO-5 > A1PO-5.

r =:#, 3=... s162 t_

e "

w

,

L_

,

100

I

200

,

I

300

Tern perature

,

I

~

400

.

I

I

500

(degree)

Figure 3. NH3-TPD of (a) AIPO-5 (b) CoAPO-11 (c) CoAPO-5 (d) MnAPO-11 (e) MnAPO-5.

Temperature programmed reduction The TPR profiles of Co- and Mn-containing samples are given in Figure 4. The two peaks centred at 633K and 833K on Co/A1PO-11 (1.6wt% Co) were due to the reduction of Co3Oa and CoO (or Co species interacting with support strongly), respectively [16,17,18]. For CoAPO-5, 11 and CoAPSO-34, no peak was

2678 detected at low temperature and the 1-I2consumption only occurred after 973K, which could be caused by the reduction of non-framework cobalt and aluminium species, formed after the collapse of molecular sieves structures [19]. This demonstrates that cobalt species existed in the framework of all Co-containing molecular sieves and were not reducible, in agreement with the work of Berndt, who studied TPR of CoAPO-5 and CoAPO-44 and found the similar results [ 18].

CotAJPO-6

.... j-

r-

;=., .1-,

MnAPO-5

L_ .s r

<

i_.

......L...; L T ;

J

............................................................................................ ........................... /

MnAPO-11

MnAPSO-34

oP~P$0-34

i I

i

I

,

I

,

I

i

i

,

i

,

I

,

i

400 5130 600 700 800 900 1000 Tern pe rature(K)

,

|

I

400

,

i

500

.

i

800

,

i

700

,

i

800

,

I

,

I

go0

T ern p e r a t u r e (K)

Figure 4. TPR results of MeAPO-5, 11, MeAPSO-34 and Me/A1PO-11. The supported sample Mn/AIPO-11 (1.6wt% Mn) presented one reduction peak with maximum at 633K, whereas no H2 consumption appeared on MnAPO-5, 11 and MnAPSO-34 before 973K. This implies that manganese species existed in the framework and were not easily reduced, with the same behaviour as CoAPO-5, 11 and CoAPSO-34.

Cyclohexane oxidation The catalytic activity of metal-containing AIPO-5, 11 and SAPO-34 had been studied in cyclohexane oxidation using molecular oxygen as oxidant. Table 3 presents the oxidation results. All of Co, Mn, Cr and V-containing molecular sieves showed catalytic activity, whereas TiAPO-5 and FeAPO-5 gave as low reactivity as A1PO-5 catalyst. CoAPO-11 and MnAPO-ll exhibited best activity and all MeAPO-11 catalysts showed better activity than corresponding MeAPO-5s and MeAPSO-34s. Compared with MeAPO-5 and-11, MeAPSO-34 only possesses smaller pore openings so that cyclohexane molecule can't access its inside. The transition metals on the external surface of MeAPSO-34 maybe contributed to the observed activities. Acidic CoAPSO-5 and CoAPSO-I 1 were also tested as catalysts and gave much lower conversion than corresponding CoAPO-5 and CoAPO-11 (see Table 3). This was possibly due to the increased interaction between the acid channels and the polar oxidation products, which caused the slow diffusion and the poor catalytic behaviour. Good selectivity of monofunctionai oxidation products (cyclohexanol, cyclohexanone and cyclohexyl hydroperoxide) was observed on CoAPO-I 1, although it had highest cyclohexane conversion of 8%. Cyclohexanol was the preferable product on most catalysts except for CrAPO-5 and-11. High selectivity of cyclohexanone was observed on Cr-containing catalysts, in agreement with the work of Sheldon [20]. They investigated the cyclohexane oxidation on CrAPO-5 and found low ratios of cyclohexanol to cyclohexanone regardless of reaction temperature and time. Both VAPO-5 and-11 showed the ability of deep oxidation due to the low selectivity of monofunctional oxidation products. Therefore, both metal types and molecular sieve structures had influence on reaction activity and selectivity.

2679 Table 3. Cyclohexane oxidation on MeAPO-5, -11 and MeAPSO-34 at 403K for 3ha. Conversion TON Product distribution (mol%)* Catalyst (%) (mOlsubstratemOlmeta1-1 h ~) A B CHHP others B/A CoAPO-5 5.0 18.3 38.1 44.6 4.2 13.1 1.2 MnAPO-5 3.1 11.4 26.2 46.5 16.4 10.9 1.8 CrAPO-5 5.9 21.6 61.0 19.9 5.0 14.1 0.3 VAPO-5 (2) 2.9 10.6 24.5 31.8 0 43.7 1.3 FeAPO-5 0.7 2.6 12.4 31.3 51.4 4.9 2.5 TiAPO-5 0.5 1.8 19.1 36.8 17.6 26.5 1.9 CoAPO- 11 7.8 28.6 43.6 43.1 1.8 11.5 1.0 MnAPO-11 7.6 27.8 33.6 43.9 3.8 18.7 1.3 CrAPO- 11 6.1 22.3 58.5 17.1 8.4 16.0 0.3 VAPO- 11 4.3 15.8 42.0 35.5 0 22.5 0.8 CoAPSO-34 2.9 10.6 25.0 43.8 22.2 9.0 1.8 MnAPSO-34 2.8 10.3 24.4 45.2 23.7 6.7 1.9 CoAPSO-5 2.5 9.2 18.0 46.4 27.4 8.2 2.6 CoAPSO- 11 2.7 9.9 17.0 39.2 34.6 9.2 2.3 A1PO-5 0.7 10.9 46.4 25.0 17.7 .3 a: catalyst(0.195mmol metal), 18g cyclohexane, 0.04g TBHP, 1.1MPa 02, 403K, 3h. *A=cyclohexanone, B=cyclohexanol and CHHP=cyclohexyl hydroperoxide The kinetic studies were performed on CoAPO-5 (Figure 5). Because small amount TBHP was added into the reaction mixture, no induction time was observed. Cyclohexane conversion and selectivity of cyclohexanone increased with time during 3hs. The ratio of cyclohexanol to cyclohexanone, the selectivity of cyclohexanol, CHHP and the monofunctional oxidation products decreased following the reaction time. It was reported that CHHP is the intermediate of the cyclohexane oxidation and its formation is the rate-determining step [21]. Two mechanisms exist for CHHP rapid decomposition: heterolytic (CHHP---~cyclohexanone) and homolytic (CHHP--*cyclohexanol). Moreover, cyclohexnol is more active than cyclohexane and can be converted to cyclohexanone easily. So there are two possible pathways for the products formation: (1) CHHP--~cyclohexanol--~cyclohexanone (2) (2) cyclohexanone~--CHHP---~cyclohexanol (then cyclohexanol-~cyclohexanone) In order to distinguish the above pathways, the oxidation of cyclohexanol was performed on CoAPO-11(CH3CN as solvent: 18g, cyclohexanol: 0.36g, TBHP: 0.03g, 02: 1.1MPa, CoAPO11:0.195mmolCo, 403K, 3h) and 10.5% yield of cyclohexanone based on cyclohexanol was obtained. At the same conditions (CH3CN: 10g and cyclohexane: 10g), cyclohexane gave 8.4% conversion and cyclohexanol/cyclohexanone ratio of 1.2. It is inferred that CHHP decomposition in the oxidation reaction involved the latter pathway. Both heterolytic and homolytic mechanisms existed, although homolytic mechanism occupied a large proportion. CONCLUSION The present work demonstrated that TG, NH3-TPD and TPR could be used as effective methods to characterize the metal-containing molecular sieves. The characterization results confirmed that metals existed in the molecular sieve framework and Co, Mn in the lattice position were not reducible before 973K. In cyclohexane oxidation, all MeAPO-11s had better activity than corresponding MeAPO-5s and MeAPSO-34s. CoAPO-I 1 exhibited the best activity and good selectivity of monofunctional oxidation products. Acidic molecular sieves had negative effect on the oxidation reaction. Both heterolytic and homolytic mechanisms existed in the decomposition of CHHP on CoAPO-11 and homolytic mechanism occupied a large proportion.

2680 100

r o

80

//" o

|

|

/ 4

">

~=

80

.........

I ...._---I 09

,-

40

I-----~-t ~ .... " ~I .J/ //" :/

~___ "-I 2

A ............A............I;::=.~L ... ............

...."................... "..........-A ................. A. 20

////H

0

.......'.....~.... 0

~

I 1

,

-

I

2

,

I

0

3

T i m e (h)

Figure 5. Conversion(m) and selectivity of cyclohexane oxidation with the reaction time on CoAPO-5. ratio of cyclohexanol to cyclohexanone (A); selectivity of cyclohexanol (e), cyclohexanone (A), CHHP (o), total selectivity of cyclohexanone, cyclohexanol and CHHP (V).

REFERENCES 1. Sheldon, R. A., Kochi, J. K., Metal-catalyzed oxidations of organic compounds, Academic Press, New York, 1981. 2. Suresh, K. A., Sharma, M. M., Sridhar, T., Ind. Eng. Chem. Res., 39 (2000), 3958-3997. 3. Schuchardt, U., Cardoso, D., Sercheli, R., Pereira, R., Da Cruz, R. S., Guerreiro, M. C., Mandelli, D., Spinace, E. V., Pires, E. L., Appl. Catal., A: 21 l(2001),l-17. 4. Sheldon, R. A., Arends, I. W. C. E., Lempers, H. E. B., Catal. Today, 41 (1998), 387-407. 5. Thomas, J.M., Raja, R., Chem. Commun., 2001, 675-687. 6. Hartmann, M., Kevan, L., Chem. Rev., 99 (1999), 635-663. 7. Lin, S. S., Weng, H. S., Appl. Catal., 105 (1993), 289-308. 8. Sankar, G., Raja, R., Thomas, J. M., Catal. Today, 55 (1998), 15-23. 9. Dugal, M., Sankar, G., Raja, R., Thomas, J. M., Angew. Chem. Int. Ed., 39 (2000), 2310-2313. 10. Wilson, S. T., Flanigen, E. M., U.S. Pat., 4,567,029, 1986. l l. Tan, J., Liu, Z. M., Bao, X. H., Liu, X. C., Han, X. W., He, C. Q., Zhai, R. S., Micropor. Mesopor. Mater., 53 (2002), 97-108. 12. Shulpin, G.B., Attanasio, D., Suber, L., J. Catal., 142 (1993), 147. 13. Blasco, T., Concepcion, P., Lopez Nieto, J. M., Perez-Pariente, J., J. Catal., 152 (1995), 1-17. 14. Yeom, Y. H., Shim, S. C., J. Mol. Catal., A: 180 (2002), 133-140. 15. Hochtl, M., Jentys, A., Vinek, H., Micropor. Mesopor. Mater., 31 (1999), 271-285. 16. Zhang, Y. L., Wei, D. G., Hammache, S., Goodwin, Jr. J. G., J. Catal., 188 (1999), 281-290. 17. Panpranot, J., Goodwin, Jr. J. G., Sayari, A., Catal. Today, 77 (2002), 269-284. 18. Berndt, H., Martin, A., Zhang, Y., Micropor. Mater., 6 (1996), l-12. 19. Roque-Malherbe, R., Lopez-Cordero, R., Gonzales-Morles, J. A., Onate-Martinez, J., Carreras-Gracial, M., Zeolites, 13 (1993), 481-484. 20. Chen, J, D., Sheldon, R. A., J. Catal., 153 (1995), 1-8. 21. Pohorecki, R., Baldyga, J., Moniak, W., Podgorska, W., Zdrojkowski, A., Wierzchowski, P. T., Chem. Eng. Sci., 56 (2001), 1285-1291.