TiO2-ZrO2

TiO2-ZrO2

Applied Catalysis A: General 244 (2003) 273–282 Vapor-phase Beckmann rearrangement of cyclohexanone oxime over B2 O3/TiO2 -ZrO2 Dongsen Mao a,∗ , Qin...

161KB Sizes 1 Downloads 7 Views

Applied Catalysis A: General 244 (2003) 273–282

Vapor-phase Beckmann rearrangement of cyclohexanone oxime over B2 O3/TiO2 -ZrO2 Dongsen Mao a,∗ , Qingling Chen a,b , Guanzhong Lu b a

Shanghai Research Institute of Petrochemical Technology, 1658 Pudong Beilu Pudong, SINOPEC, Shanghai 201208, PR China b Research Institute of Industrial Catalysis, East China University of Science and Technology, Shanghai 200237, PR China Received 11 May 2002; received in revised form 21 September 2002; accepted 6 November 2002

Abstract Pure TiO2 , ZrO2 and TiO2 -ZrO2 mixed oxides with various compositions were prepared by the co-precipitation method. Catalysts containing 12 wt.% boria were prepared using these oxides and their catalytic performance for vapor-phase Beckmann rearrangement of cyclohexanone oxime to ε-caprolactam was determined. With increasing zirconia content in the mixed oxide, the specific surface area and the temperature of the phase transformation increased; maximum values were obtained for a mixed oxide having a Ti and Zr molar ratio of 1/1. XRD results indicated that TiO2 -ZrO2 became amorphous after calcination at 500 ◦ C if neither the content of TiO2 nor that of ZrO2 was less than 25%. The yield of lactam increased with increasing zirconia content in mixed oxide and reached a maximum value for the B2 O3 /TiO2 -ZrO2 (1/1) catalyst. With further increase in zirconia content, the lactam yield decreased; the lowest value was observed for the pure ZrO2 -supported catalyst. The acidic property and the pore characteristics of the catalyst were key factors that affected its performance. The influences of the operating parameters on the performance of B2 O3 /TiO2 -ZrO2 (1/1) catalyst were also investigated. © 2002 Elsevier Science B.V. All rights reserved. Keywords: Beckmann rearrangement; Cyclohexanone oxime; ε-Caprolactam; Titanium oxide; Zirconium oxide; TiO2 -ZrO2 ; Boria; B2 O3 /TiO2 -ZrO2

1. Introduction The Beckmann rearrangement of cyclohexanone oxime to ε-caprolactam, which is the precursor to nylon-6, is now being carried out using fuming sulfuric acid as a catalyst. However, this process yields more undesirable ammonium sulfate than the desired lactam. Moreover, the use of corrosive sulfuric acid should be avoided for environmental and economic reasons. Since ‘vapor-phase’ Beckmann rearrangement on solid catalysts would eliminate these prob∗ Corresponding author. Tel.: +86-21-68462197-6404; fax: +86-21-68462283. E-mail address: [email protected] (D. Mao).

lems, a wide variety of solid acid catalysts have been studied in the search for an alternative clean process. Among them, boria catalysts supported on various carriers have been studied most extensively [1–23]. The performance of supported boria catalysts for the vapor-phase Beckmann rearrangement is known to be sensitive to the dispersion of boria and to the nature of the carrier. For example, Sato and co-workers [10,11] reported that the boria supported on silica, prepared by vapor deposition, was more active and selective than the boria on silica catalyst prepared by impregnation using boric acid, since the vapor deposition technique resulted in a uniform spread of boria on the silica support. Recently, Xu and co-workers [18–20] reported that the catalytic performance of

0926-860X/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 8 6 0 X ( 0 2 ) 0 0 5 9 9 - 9

274

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

boria supported on zirconia, a characteristic carrier that possesses both acidic and basic properties, was much better than that of other boria catalysts supported on Al2 O3 , SiO2 , TiO2 and MgO. Previous researches have highlighted the use of single component oxides such as alumina, silica, thoria, titania and zirconia as supports for boria-based catalysts for the Beckmann rearrangement reaction. However, in comparison with the single component supports, the composite oxide supports are found to exhibit higher surface area, surface acidity, and thermal and mechanical strength [24]. A higher surface area means that it is possible to prepare a highly dispersed boria catalyst. The higher thermal resistance is advantageous since it allows the regeneration of deactivated catalysts by high temperature calcination. It is well-established in the literature that catalysts supported on mixed metal oxides often produce superior catalysis to the ones supported on single oxides for a number of reactions [25,26]. In the present paper, we describe new results for the vapor-phase Beckmann rearrangement of cyclohexanone oxime over B2 O3 catalysts supported on various binary oxides based on SiO2 , Al2 O3 , TiO2 and ZrO2 . Among these catalysts, B2 O3 /TiO2 -ZrO2 , with the highest selectivity and yield of lactam, was studied as a function of the oxide composition. The acidic properties and pore characteristics of a series of B2 O3 /TiO2 -ZrO2 catalysts with different molar ratios of TiO2 to ZrO2 were measured in order to clarify the relationships between the physiochemical properties and the catalytic behavior of B2 O3 /TiO2 -ZrO2 catalysts. Furthermore, the influences of the main process variables, i.e. reaction temperature, oxime space velocity, carrier gas and its flow rate, and diluent solvent, on the catalytic performance of the most efficient B2 O3 /TiO2 -ZrO2 catalyst were also investigated.

2. Experimental 2.1. Catalyst preparation The coprecipitation procedure was used to prepare mixed oxide supports. A series of titania and zirconia mixed oxides with varying molar ratios were prepared by this method using aqueous ammonia as precipitation reagent. In a typical experiment, in order to obtain

the desired composition of the mixed oxide support, the requisite quantities of titanium tetrachloride and zirconium oxychloride were dissolved separately in deionized water and then mixed together. Cold TiCl4 was first digested in cold concentrated HCl and subsequently diluted with deionized water. The mixture was added to an excess amount of aqueous ammonia slowly and this combination was mixed thoroughly by control of the final pH at 9. The precipitate was allowed to stand at room temperature for 24 h, then filtered and washed with deionized water to remove chloride ions. The solid obtained was dried at 110 ◦ C overnight and then calcined at 500 ◦ C in air for 6 h. Other mixed oxide supports were prepared from Na2 SiO3 ·9H2 O and Al(NO3 )3 ·9H2 O. Pure TiO2 and ZrO2 were also prepared using the same method. All the supports were pressed into wafers, crushed and sieved to 40–60 mesh before use for impregnation of B2 O3 . The incipient wetness technique was used to load 12 wt.% B2 O3 on these supports. An appropriate amount of boric acid (AR grade, Shanghai Chemical Reagent Corporation) was dissolved in a predetermined volume of distilled water, based on the pore volume of the support. The impregnated catalysts were dried at 110 ◦ C for 12 h and then calcined in air at 600 ◦ C for 12 h. 2.2. Reaction testing The vapor-phase Beckmann rearrangement reaction of cyclohexanone oxime was conducted under atmospheric pressure in a conventional continuous flow fixed bed-type reactor made of stainless steel (i.d. = 6 mm). A mixture of oxime and solvent (5 wt.%) in the liquid phase supplied from a micro feeder at a constant flow rate was mixed with diluent gas and fed through the evaporator into the reactor. The catalyst, after reactor loading, was pretreated at 350 ◦ C for 1 h in a diluent gas stream (50 ml min−1 ). Unless specified otherwise, benzene was used as a solvent and nitrogen was used as a carrier gas with a flow rate of 30 ml min−1 . The standard reaction conditions are −1 300 ◦ C, 0.1 MPa and a WHSV of 0.33 greactant g−1 cat h . Products and unreacted oxime collected in a trap at 0–10 ◦ C were analyzed off-line with a gas chromatograph (column: capillary, HP-1, crosslinked methyl siloxane, 30 m × 0.32 mm × 0.25 ␮m; FID detector; carrier gas: N2 ; G.C. model: Hewlett Packard 4890

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

275

D). The selectivity to lactam was calculated based on the amount of reacted oxime.

Table 1 Performance of supported boria catalysts for Beckmann rearrangement of cyclohexanone oxime

2.3. Catalyst characterization

Catalysts

Conversion of oxime (wt.%)

Selectivity of lactam (wt.%)

Yield of lactam (wt.%)

B2 O3 /TiO2 -ZrO2 B2 O3 /SiO2 -Al2 O3 B2 O3 /SiO2 -TiO2 B2 O3 /SiO2 -ZrO2 B2 O3 /Al2 O3 -TiO2 B2 O3 /Al2 O3 -ZrO2

100 100 17.5 22.6 97.4 35.2

97.4 60.0 35.1 31.1 36.2 40.0

97.4 60.0 6.14 7.03 35.2 14.1

BET surface area and pore size distribution were measured on a Micromeritics Digisorb 2600 system at liquid N2 temperature using N2 as adsorbate. Before measurements, the samples were dried in situ at 450 ◦ C for 3 h under vacuum. X-ray powder diffraction (XRD) measurements were performed on a Rigaku D/MAX-1400 instrument using Cu K␣ radiation with 15◦ min−1 scan speed and 4–60◦ scan range at 40 kV and 40 mA. Differential thermal analysis (DTA) was conducted on a TA Instruments S.D.T. 2960. Samples were heated in air at a rate of 10 ◦ C min−1 from room temperature to 800 ◦ C. Acidity measurements were performed by temperature-programmed desorption of ammonia (NH3 -TPD) using a conventional flow apparatus equipped with a thermal conductivity detector (TCD). A given amount of the samples, 0.1 g, was pretreated in flowing helium at 500 ◦ C for 1 h, cooled to 150 ◦ C and then exposed to NH3 (20 ml min−1 ) for 30 min. The samples adsorbed by NH3 were subsequently purged with He at the same temperature for 1 h to remove the physisorbed NH3 . The TPD measurements were conducted in flowing He (30 ml min−1 ) from 150 to 550 ◦ C at a heating rate of 10 ◦ C min−1 . The coke content of the catalysts was determined by combustion in a thermogravimetric analyzer. Samples were first heated from 20 to 150 ◦ C in a flow of 30 ml min−1 of nitrogen until no more weight loss occurred. Then a stream of 50 ml min−1 of air was passed through the samples and the temperature was raised to 700 ◦ C at 10 ◦ C min−1 . The weight loss between 350 and 700 ◦ C was attributed to coke.

3. Results and discussion 3.1. Performance of various B2 O3 -supported catalysts Vapor-phase Beckmann rearrangement of cyclohexanone oxime was tested over 12 wt.% B2 O3 cata-

Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent/carrier: benzene/N2 ; N2 flow rate = 30 ml min−1 ; time-on-stream = 2 h.

lysts supported on various binary oxides. The major product was lactam, and the side products included nitriles (cyanopentane and 5-cyano-1-pentene), cyclic ketones (cyclohexanone and 2-cyclohexenone) and aniline, in accordance with the literature [4,20]. The amount of unidentified high-boiling products did not exceed 1% and these can be neglected. The conversion of oxime and the selectivity and yield to lactam obtained at time-on-stream of 2 h are shown in Table 1. The data in Table 1 indicate that the performance of B2 O3 catalysts was strongly affected by the oxide supports, i.e. SiO2 -ZrO2 , SiO2 -TiO2 and Al2 O3 -ZrO2 exhibited both low oxime conversion and low lactam selectivity. By contrast, boria catalyst supported on TiO2 -ZrO2 was both highly active and selective for the lactam synthesis. On the other hand, although very high oxime conversions (>95%) were obtained on Al2 O3 -TiO2 and SiO2 -Al2 O3 supported catalysts, the selectivity to lactam was low. The lactam yield decreased with the support in the order: TiO2 -ZrO2 (lactam yield: 97.4%) > SiO2 Al2 O3 (60%) > Al2 O3 -TiO2 (35.2%) > Al2 O3 -ZrO2 (14.1%) > SiO2 -ZrO2 (7.03%) > SiO2 -TiO2 (6.14%). Evidently, B2 O3 /TiO2 -ZrO2 showed the best performance for the production of lactam at the 12 wt.% level of boria loading. The results indicate that, among the binary oxides examined here, TiO2 -ZrO2 exhibited higher activity for the conversion of oxime and favored the formation of desired lactam. Detailed catalytic properties were accordingly studied for the B2 O3 /TiO2 -ZrO2 catalyst.

276

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

3.2. Structure and catalytic performance of B2 O3 /TiO2 -ZrO2 The BET surface area was measured for TiO2 -ZrO2 and pure TiO2 and ZrO2 after calcination at 500 ◦ C in air. The results are given in Table 2. The surface areas of zirconia and titania were found to be 35.0 and 51.8 m2 g−1 , respectively. With the addition of zirconia to titania, the surface area of the supports increased sharply and reached a maximum value for a mixed oxide having a Ti and Zr molar ratio of 1/1. The surface area of this mixed oxide support (having Ti/Zr = 1/1) was around 171 m2 g−1 . The mixed oxide was reported to exhibit much higher surface area than the single oxide components and to attain a maximum at equal components of TiO2 and ZrO2 [27]. With the addition of further amounts of zirconia to titania, the surface areas of mixed oxides decreased slowly. Fig. 1 shows the powder X-ray diffraction patterns of pure TiO2 , ZrO2 and their mixed oxide supports. The diffractograms show the anatase phase of titania in TiO2 and TiO2 -ZrO2 mixed oxide prepared with Ti/Zr = 9/1. Peaks due to zirconia were not found in these oxides. Similarly, pure ZrO2 and TiO2 -ZrO2 with Ti/Zr = 1/9 were composed of monoclinic/tetragonal phases of zirconia, and peaks due to titania were not found. The mixed oxides having Ti/Zr molar ratios of 3/1, 1/1 and 1/3 were observed as X-ray amorphous. These results were consistent with those previously reported by Wang et al. [27]. The DTA curves of Ti(OH)4 , Zr(OH)4 and mixed titanium–zirconium hydroxides are shown in Fig. 2. The strong exothermal peak of Zr(OH)4 at around 460 ◦ C can be assigned to the crystallization of the amorphous zirconia into monoclinic crystalline phase [28]. The temperature of the phase transformation

Fig. 1. XRD patterns of TiO2 -ZrO2 with various Ti/Zr atomic ratios: (a) ZrO2 ; (b) 1/9; (c) 1/1; (d) 9/1; (e) TiO2 . (䊏) TiO2 (anatase); (䊉) ZrO2 (tetragonal); (䉱) ZrO2 (monoclinic). All samples were calcined at 500 ◦ C.

Table 2 BET surface area of TiO2 -ZrO2 with various Ti/Zr atomic ratios calcined at 500 ◦ C Ti/Zr (mol mol−1 )

SBET (m2 g−1 )

0/100 10/90 25/75 50/50 90/10 100/0

35.0 45.8 133.7 170.9 90.4 51.8

Fig. 2. DTA curves of titania–zirconia co-precipitates with various Ti/Zr atomic ratios: (1) Zr(OH)4 ; (2) 1/9; (3) 1/3; (4) 1/1; (5) 3/1; (6) Ti(OH)4 .

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

was increased with the addition of titania, and this change was observed up to the atomic ratio of Ti/Zr = 1/1. Similarly, the exothermal peak at about 450 ◦ C observed with Ti(OH)4 , which is attributed to the crystallization of amorphous titania, moved to higher temperatures up to the atomic ratio of Ti/Zr = 1/1. This observation was at variance with that of Lónyi and Valyon [28], who found that the phase transformation of titania was only slightly influenced by the presence of zirconia. The above results from XRD and DTA experiments reveal that the crystallization of mixed oxides was inhibited due to the mutual chemical interaction, which induced the increased surface area. The surface acidic properties of the B2 O3 /TiO2 , B2 O3 /ZrO2 and B2 O3 /TiO2 -ZrO2 catalysts were determined by NH3 -TPD. The profiles (Fig. 3) show that the acid strength of the catalysts increased with zirconia content in the mixed oxide, as was indicated by a continuous shift of the desorption peak of maximum height to higher desorption temperatures. When the zirconia content was no more than 50 mol%, the peak maximum occurred at about 250 ◦ C. At higher content of 75 and 90% zirconia, the desorption maximum increased to 285 ◦ C. The peak maximum had

Fig. 3. NH3 -TPD profiles of the B2 O3 /TiO2 -ZrO2 catalysts with various Ti/Zr atomic ratios: (1) ZrO2 ; (2) 1/9; (3) 1/3; (4) 1/1; (5) 3/1; (6) 9/1; (7) TiO2 .

277

Table 3 Acid strength distributions of B2 O3 /TiO2 -ZrO2 catalysts with various Ti/Zr atomic ratios Catalysts

Distribution of peak area (%) <200 ◦ C 200–350 ◦ C >350 ◦ C

B2 O3 /ZrO2 B2 O3 /TiO2 -ZrO2 B2 O3 /TiO2 -ZrO2 B2 O3 /TiO2 -ZrO2 B2 O3 /TiO2 -ZrO2 B2 O3 /TiO2 -ZrO2 B2 O3 /TiO2

(Ti/Zr = 1/9) (Ti/Z = 1/3) (Ti/Z = 1/1) (Ti/Zr = 9/1) (Ti/Zr = 3/1)

5.8 6.0 4.2 4.2 7.5 9.4 10.4

83.0 85.2 88.5 91.4 88.3 86.6 85.5

11.2 8.8 7.3 4.4 4.2 4.0 4.1

increased to 300 ◦ C with the B2 O3 /ZrO2 catalyst. The acid strength distributions were obtained by analyzing the TPD profiles, following the procedure proposed by Curtin et al. [4,5]. The sites of weak, intermediate and strong acid strength were characterized by desorption of adsorbed ammonia <200, 200–300 and >350 ◦ C, respectively. The strength distributions of acid sites over boria-supported catalysts are summarized in Table 3. The pore size distributions of the boria catalysts were measured and the results are depicted in Fig. 4. B2 O3 /ZrO2 and B2 O3 /TiO2 -ZrO2 with titania content of 10% had a sharp peak at about 9 nm. With the increasing of titania content in the mixed oxide, the peak became broad. Meanwhile, the most probable pore size diameter was shifted to a higher value and reached a maximum of around 36 nm for B2 O3 /TiO2 catalyst. The boria catalysts supported on TiO2 , ZrO2 and TiO2 -ZrO2 were tested at 300 ◦ C; the variation in the conversion of oxime with time-on-stream is presented in Fig. 5. Initially, all the boria catalysts showed 100% oxime conversion regardless of the supports. However, they deactivated at different rates with reaction time-on-stream. The conversion of oxime already began to decrease at a time-on-stream of ≤3 h in the cases of the B2 O3 /ZrO2 , B2 O3 /TiO2 and B2 O3 /TiO2 -ZrO2 with Ti/Zr = 1/9 and 1/3. On the other hand, the complete oxime conversion was maintained until 4 h over the B2 O3 /TiO2 -ZrO2 with Ti/Zr = 1/1, 3/1 and 9/1. It is well-known that the catalyst deactivation for the rearrangement reaction is mainly due to coke formation on the catalyst surface. In the case of amorphous materials, the loss of the impregnated metal oxide, such as B2 O3 , may be another

278

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

Fig. 4. Pore size distributions of B2 O3 /TiO2 -ZrO2 catalysts with various Ti/Zr atomic ratios: ( ) ZrO2 ; (䉱) 1/9; (䊏) 1/3; ( ) 1/1; (䊉) 3/1; (䊊) TiO2 .

factor that can account for the catalyst deactivation. Quantitative measurements of boron in the samples revealed that no significant loss of boron occurred during the reaction. Yin et al. [23] also reported that the difference in boria loading between the fresh and corresponding deactivated samples was less than 1% of the loading in the fresh samples. Hence, we conclude that coke formation is the main reason for the deactivation of the B2 O3 /TiO2 -ZrO2 catalysts. Un-

Fig. 5. Change of conversion of oxime with time-on-stream over B2 O3 /TiO2 -ZrO2 catalysts with various Ti/Zr atomic ratios: (1) ZrO2 ; (2) 1/9; (3) 1/3; (4) 1/1; (5) TiO2 . Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent: benzene; carrier gas: N2 ; N2 flow rate = 30 ml min−1 .

like the conversion of oxime, no significant change in the selectivity to lactam with time-on-stream was observed over all of the boria catalysts; this was not shown in the figure. This phenomenon was also found over other catalysts [4,5,18,20]. The average conversion of oxime and selectivity to lactam during the initial 5 h reaction are illustrated in Fig. 6. The oxime conversion reached a maximum at 50–90 mol% of TiO2 , while the selectivity to lactam increased with increasing ZrO2 content in mixed oxide supports and reached a maximum at titania content of 50%; above this support composition, the selectivity steadily decreased with increasing ZrO2 concentration in mixed oxide. The above results can be attributed to two factors that affect the lactam selectivity and the deactivation rate of catalyst in this reaction. The first factor is the acidity strength distribution of catalysts. It has been reported that the acid sites of intermediate strength are responsible for selective formation of lactam [2,4–6,20,29], while the strong acid sites may accelerate the formation of by-products, which makes the lactam selectivity decrease and the catalyst deactivate quickly due to the formation of coke and polymer on the surface of catalyst [2,30,31]. The second factor is the pore size distribution of catalysts. Catalysts with small pore sizes need a longer time for lactam diffusing from the stronger acid sites of

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

Fig. 6. Variation of oxime conversion and lactam selectivity as a function of Ti/(Ti + Zr): (䊏) oxime conversion; (䊉) lactam selectivity. Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent: benzene; carrier gas: N2 ; N2 flow rate = 30 ml min−1 .

catalyst, which may cause undesirable catalytic side reactions such as polymerization and decomposition of the produced lactam, and consequently the catalysts are often deactivated by rapid carbon deposit. Other researchers [31,32] also proposed that the pore sizes of catalysts also play an important role in determining the performance of catalysts. In our cases, B2 O3 /ZrO2 and B2 O3 /TiO2 -ZrO2 with Ti/Zr = 1/9 deactivated faster and were less selective for the synthesis of lactam than other catalysts because of their lower number of intermediate acid sites and higher number of strong acid sites on the surface, and because of the smaller pore sizes. As explained above, the B2 O3 /TiO2 and B2 O3 /TiO2 -ZrO2 with TiO2 content ≥50% which possess lower number of strong acid sites on the surface and larger pore sizes showed higher activity over time. However, the selectivity to lactam decreased with TiO2 content in the mixed oxide. This result was related to the different percentages of the intermediate acid sites, which decreased with TiO2 content (Table 3). Xue et al. [20] and Shouro et al. [29] reported that the lactam selectivity depends on the percentage of acid sites with intermediate strength. The maximum fraction of intermediate strength acids was observed for B2 O3 /TiO2 -ZrO with

279

Ti/Zr = 1/1, which exhibited the highest selectivity for the production of lactam. Based on these results, we conclude that acid sites of intermediate strength on B2 O3 /TiO2 -ZrO2 catalysts are catalytically active sites for the selective Beckmann rearrangement of cyclohexanone oxime. From this result and others in the literature [2,4–6,11,20], we conclude that the acidity requirement for the vapor-phase rearrangement reaction on supported boria catalysts and zeolites is different. A majority of previous reports have suggested that the very weak [33,34] or almost neutral hydroxyl groups [35–38] of zeolites are favorable for Beckmann rearrangement reaction. The influences of the main process variables, i.e. reaction temperature, oxime space velocity (WHSV), carrier gas and its flow rate, and solvent, on the activity, selectivity and stability of the most efficient B2 O3 /TiO2 -ZrO2 (1/1) catalyst for vapor phase Beckmann rearrangement reaction were studied and are discussed. 3.2.1. Influence of reaction temperature The influence of reaction temperature on the catalytic reaction was investigated in the range of 250– 350 ◦ C. As illustrated in Fig. 7, the catalyst lifetime increased along with the temperature. It is assumed

Fig. 7. Influence of reaction temperature on conversion of oxime over B2 O3 /TiO2 -ZrO2 catalyst: (1) 250 ◦ C; (2) 300 ◦ C; (3) 350 ◦ C. Reaction conditions: P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent: benzene; carrier gas: N2 ; N2 flow rate = 30 ml min−1 .

280

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282 Table 4 Influence of WHSV of oxime WHSV (h−1 )

Conversion of oxime (wt.%)

Selectivity of lactam (wt.%)

Yield of lactam (wt.%)

0.22 0.33 0.44

100 98.3 86.9

94.9 97.0 97.9

94.9 95.4 85.1

Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; solvent/carrier: benzene/N2 ; N2 flow rate = 30 ml min−1 and data were averaged for 6 h on stream.

oxime at these temperatures decreased much more rapidly than at higher temperatures.

Fig. 8. Influence of reaction temperature on selectivity to lactam over B2 O3 /TiO2 -ZrO2 catalyst. Reaction conditions: P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent: benzene; carrier gas: N2 ; N2 flow rate = 30 ml min−1 .

that deactivating compounds desorb more easily at elevated temperatures. At 350 ◦ C there was no drop in oxime conversion after 6 h time-on-stream. At a temperature of 300 ◦ C, however, the conversion of oxime dropped to about 92% in that time period and at 250 ◦ C the decline was much more severe (Fig. 7). Similar results were obtained over the B2 O3 catalysts supported on Al2 O3 [3,4], SiO2 [11] and ZrO2 [16,20], and Ta2 O5 /SiO2 catalyst [31]. The highest selectivity to lactam was obtained at a reaction temperature of 300 ◦ C (Fig. 8). Temperatures above 300 ◦ C led to selectivity reduction due to the decomposition of lactam and other side reactions. Only one by-product (cyclohexanone) of the rearrangement reaction was detected at temperatures not higher than 300 ◦ C. At higher temperatures, formations of cyanopentane and 5-cyano-1-pentene also became detectable. Indeed, at 350 ◦ C, the color of the products’ solution was yellow, and unknown products appeared in the gas chromatogram. Lower temperatures also caused a drop in selectivity. Although this result is somewhat at variance with that over B2 O3 /Al2 O3 catalyst reported by Sato et al. [2], we postulate that the likelihood of re-adsorption of the produced lactam on the catalyst surface was increased at these lower temperatures. This hypothesis was borne out by the observation that conversion of

3.2.2. Influence of space velocity The results of the studies on WHSV of the feed on the conversion of oxime and lactam selectivity are shown in Table 4. Progressive decrease in conversion of oxime and increase in selectivity of lactam was observed. The maximum yield of lactam was obtained at WHSV = 0.33 h−1 . The formation of cyclohexanone and 5-cyano-1-pentene was suppressed at high WHSV. Moreover, there was much more rapid deactivation at higher space velocities. 3.2.3. Influence of carrier gas and its flow rate The influence of carrier gas on oxime conversion and lactam selectivity of Beckmann rearrangement reaction is shown in Table 5. It is clear that nitrogen was the most favorable carrier gas. On the other hand, the use of basic ammonia as the carrier gas resulted in loss of both catalytic activity and the selectivity to the desired lactam. The negative effect resulted from a poisoning of the medium strength acid sites by NH3 molecules. This observation was in contrast to that of Table 5 Influence of carrier gas Carrier gas

Conversion of oxime (wt.%)

Selectivity of lactam (wt.%)

Yield of lactam (wt.%)

He N2 H2 CO2 NH3 (5 vol.% in He)

95.3 98.3 96.8 96.8 78.9

96.8 97.0 97.2 94.3 79.2

92.2 95.4 94.1 91.3 62.5

Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent: benzene; carrier gas flow rate = 30 ml min−1 and data were averaged for 6 h on stream.

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282 Table 6 Influence of carrier gas (N2 ) flow rate

281

Table 7 Influence of solvent

Flow rate (ml min−1 )

Conversion of oxime (wt.%)

Selectivity of lactam (wt.%)

Yield of lactani (wt.%)

15 30 60

99.3 98.3 95.2

93.6 97.0 98.2

92.9 95.4 93.5

Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; WHSV = 0.33 h−1 ; solvent/carrier: benzene/N2 and data were averaged for 6 h on stream.

Ichihashi et al. [39], who reported that the selectivity of lactam was increased without affecting the conversion of oxime when ammonia was fed to a high silica MFI catalyst with oxime. The discrepancy may arise from the different requirements for acidity of the active sites with zeolite and supported boria catalysts as stated above. Table 6 shows the influence of carrier gas flow rate on the rearrangement of cyclohexanone oxime to ε-caprolactam. Increasing the flow rate of nitrogen, thus, reducing the residence time of reactant over the catalyst, we observed an increase in the selectivity of lactam and a decrease in the conversion of oxime. The yield of lactam was maximum when the carrier gas flow rate was 30 ml min−1 . At higher flow rates, the formation of cyclohexanone was suppressed. The effect of N2 flow rate was similar to that reported for other solid acids [40]. 3.2.4. Influence of solvent Different solvents used for the Beckmann rearrangement, as well as their effect on selectivity, conversion and catalyst service time, have been reported [16,22,32,40]. Our investigations indicate that polar solvents, such as tetrahydrofuran or acetonitrile, gave much higher selectivity to lactam and lower deactivation rates than non-polar solvents, such as benzene or cyclohexane (Table 7). The positive effect of polar solvents on increasing the selectivity for lactam was verified in the oxime reaction over WO3 /SiO2 catalyst [32]. The most likely reason for this is that the polar solvents increase the desorption rate of lactam, decreasing its contact time on the catalyst surface, so lowering the likelihood for acid-catalyzed opening of lactam to form polymers and the coke precursors on the catalyst surface. This was confirmed by the fact that there was a decrease in the amount of coke found

Solvent

Conversion of oxime (wt.%)

Selectivity of lactam (wt.%)

Yield of lactam (wt.%)

Benzene Cyclohexane Tetrahydrofuran Acetonitrile

98.3 97.6 100 100

97.0 95.6 97.2 98.6

95.4 93.3 97.2 98.6

Reaction conditions: T = 300 ◦ C; P = 0.1 MPa; WHSV = 0.33 h−1 ; carrier: N2 ; N2 flow rate = 30 ml min−1 and data were averaged for 6 h on stream.

on catalyst to 1.43% with acetonitrile, compared to the 1.97% coke found with benzene. 4. Conclusions The following conclusions can be drawn: (1) B2 O3 /TiO2 -ZrO2 was found to be a better catalyst among the investigated catalysts for the rearrangement reaction to ε-caprolactam. (2) Among TiO2 -ZrO2 mixed oxides, the boria catalysts supported on mixed oxide having a composition of 1/1 showed the highest activity for Beckmann rearrangement reaction of cyclohexanone oxime. (3) Both the distributions of acid strength and the pore sizes of B2 O3 /TiO2 -ZrO2 catalysts play important roles in the selective formation of lactam. (4) Conversion of oxime increased with increase in temperature; however, the highest selectivity to lactam was obtained at a reaction temperature of 300 ◦ C. The B2 O3 /TiO2 -ZrO2 catalyst deactivated quickly at lower temperatures. (5) Polar solvents gave higher selectivity to lactam and lower deactivation rates than non-polar solvents, since the desorption of lactam from the catalyst surface was promoted by the attack of a polar solvent molecule. Acknowledgements Financial support provided by the Shanghai Research Institute of Petrochemical Technology (SRIPT), SINOPEC, is gratefully acknowledged. References [1] H. Sakurai, S. Sato, K. Urabe, Y. Izumi, Chem. Lett. (1985) 1783.

282

D. Mao et al. / Applied Catalysis A: General 244 (2003) 273–282

[2] S. Sato, S. Hasebe, H. Sakurai, K. Urabe, Y. Izumi, Appl. Catal. 29 (1987) 107. [3] T. Curtin, J.B. McMonagle, B.K. Hodnett, Stud. Surf. Sci. Catal. 59 (1991) 531. [4] T. Curtin, J.B. McMonagle, B.K. Hodnett, Appl. Catal. A 93 (1992) 75. [5] T. Curtin, J.B. McMonagle, B.K. Hodnett, Appl. Catal. A 93 (1992) 91. [6] T. Curtin, J.B. McMonagle, B.K. Hodnett, Catal. Lett. 17 (1993) 145. [7] T. Curtin, J.B. McMonagle, B.K. Hodnett, Stud. Surf. Sci. Catal. 75 (1993) 2609. [8] T. Curtin, B.K. Hodnett, Stud. Surf. Sci. Catal. 78 (1993) 535. [9] T. Curtin, J.B. McMonagle, M. Ruwet, B.K. Hodnett, J. Catal. 142 (1993) 172. [10] S. Sato, H. Sakurai, K. Urabe, Y. Izumi, Chem. Lett. (1985) 277. [11] S. Sato, K. Urabe, Y. Izumi, J. Catal. 102 (1986) 99. [12] W.F. Yates, R.O. Downs, J.C. Burleson, US Patent 3 639 391 (1972), to Monsanto Corporation. [13] K. Yoshida, K. Fujiki, T. Harada, Y. Moroi, T. Yamaguchi, Japan Patent 7 310 478 (1973), to Mitsubishi Chemical Industries Co., Ltd. [14] Y. Izumi, S. Sato, K. Urabe, Chem. Lett. (1983) 1649 [15] S.B. Cheng, B.Q. Xu, D.Q. Wang, F. Wang, T.X. Cai, X.S. Wang, Chin. J. Catal. 17 (1996) 281. [16] S.B. Cheng, B.Q. Xu, S.P. Tang, S. Jiang, T.X. Cai, X.S. Wang, Chin. J. Catal. 17 (1996) 330. [17] S.B. Cheng, B.Q. Xu, S. Jiang, T.X. Cai, X.S. Wang, Chin. J. Catal. 17 (1996) 404. [18] S.B. Cheng, B.Q. Xu, S. Jiang, F.P. Tian, T.X. Cai, X.S. Wang, Chin. J. Catal. 17 (1996) 512. [19] S.B. Cheng, B.Q. Xu, F. Wang, D.Q. Wang, T.X. Cai, X.S. Wang, Chin. J. Mol. Catal. 10 (1996) 161.

[20] B.Q. Xu, S.B. Cheng, S. Jiang, Q.M. Zhu, Appl. Catal. A 188 (1999) 361. [21] B.Q. Xu, S.B. Cheng, X. Zhang, S.F. Ying, Q.M. Zhu, Chem. Commum. (2000) 1121. [22] B.Q. Xu, S.B. Cheng, X. Zhang, Q.M. Zhu, Catal. Today 63 (2000) 275. [23] S.F. Yin, B.Q. Xu, Q.M. Zhu, Chin. J. Catal. 22 (2001) 494. [24] J.B. Miller, S. Rankin, E.I. Ko, J. Catal. 148 (1994) 673. [25] H. Aritani, O. Fukuda, T. Yamamoto, T. Tanaka, S. Imamura, Chem. Lett. (2000) 66. [26] H.S. Roh, W.S. Dong, K.W. Jun, S.E. Park, Chem. Lett. (2001) 88. [27] I. Wang, W.F. Chang, R.J. Shiau, J.C. Wu, C.S. Chung, J. Catal. 83 (1983) 428. [28] F. Lónyi, J. Valyon, J. Therm. Anal. 46 (1996) 211. [29] D. Shouro, Y. Moriya, T. Nakajima, S. Mishima, Appl. Catal. A 198 (2000) 275. [30] T. Takahashi, K. Ueno, T. Kai, Can. J. Chem. Eng. 69 (1991) 1096. [31] T. Ushikubo, K. Wada, J. Catal. 148 (1994) 138. [32] N. Kob, R.S. Drago, Catal. Lett. 49 (1997) 229. [33] W.F. Hölderich, J. Röseler, G. Heitmann, A.T. Liebens, Catal. Today 37 (1997) 353. [34] T. Yashima, N. Oka, T. Komatsu, Catal. Today 38 (1997) 249. [35] H. Sato, N. Ishii, K. Hirose, S. Nakamura, Stud. Surf. Sci. Catal. 28 (1986) 755. [36] H. Sato, K. Hirose, M. Kitamura, Y. Nakamura, Stud. Surf. Sci. Catal. 49 (1989) 1213. [37] H. Sato, K. Hirose, Chem. Lett. (1993) 1765. [38] H. Sato, K. Hirose, Y. Nakamura, Chem. Lett. (1993) 1987. [39] H. Ichihashi, M. Kitamura, H. Kajikuri, E. Tasaka, US Patent 5 354 859 (1994), to Sumitomo Chemical Corporation. [40] P.S. Singh, R. Bandyopadhyay, S.G. Hegde, B.S. Rao, Appl. Catal. A 136 (1996) 249.