159 Studies in Surface Science and Catalysis, volume 159 Hyun-Ku Rhee, In-Sik Nam and Jong Moon Park (Editors) © 2006 Elsevier B.V. All rights reserved
Regeneration of high silica zeolite used for Beckmann rearrangement Takeshige Takahashi, M ay umi Nakanishi and Takami Kai Dept of Appl. Chem. and Chem. Eng., KagosMma Univ., Kagoshima, 890-0065 JAPAN [email protected]
,kagoshima-u.ae,jp Abstract Beckmann rearrangement of cyclohexanone over the regenerated TS-1 and SSZ-31 was carried out to clarify the effect of regeneration number and regeneration atmosphere on e-caprolactam selectivity and catalyst deactivation rate. The rearrangement rate constant for the both catalysts decreased with the number of regeneration, but the e-caprolactam selectivity and deactivatian factor observed from TS-1 was almost constant. On the other hand, e-caprolactam from SSZ-31 regenerated in Ar atmosphere decreased with the number of regeneration. These behaviors indicated that the coke precursor deposited on the SSZ-21 surface changed to hard coke in the pore, whereas the coke precursor produced on the pore mouth for TS-1 was easily removed in Ar atmosphere. 1. Introduction Recently, vapor phase Beckmann rearrangement of cyclohexanone oxime was carried out over many kinds of solid catalysts, such as high silica MFI zeolite , p type zeolite  or meso-porous materials . Sato and Ichihashi [1,4] reported that high silica MFI zeolite without acidity was the most suitable for the Beckmann rearrangement, but the catalyst deactivation rate was too large to use the catalyst for a fixed bed reactor. Takahashi et al.[5,6] found mat the catalyst deactivation of the Beckmann rearrangement was the adsorption of coke precursor, such as oligomer of the e-caprolactam (CL). In the present study, the regeneration was carried out in argon or air atmosphere. The rearrangement was performed over the regenerated high silica TS-1 zeolite and SSZ-31 zeolites to elucidate the effect of the regeneration number of on CL selectivity and catalyst deactivation rate. The characterization of the used catalyst was also carried out to clarify fee deactivation mechanism for Beckmann rearrangement 2. Experimental High silica SSZ-31 zeolite (Si/Al=962) and TS-1 (Si/Ti= 45~300) was prepared by authentic methods . The pore size of SSZ-31 (0.57*0.86 mn) is larger than that of TS-1 (0.54*0.56 nm). The zeolites changed to proton type immersing in ammonium nitrate solution and followed by calcination at 773K for 5h. The physical properties and acidity of the zeolites is listed in Table 1 The zeolite was compressed to small pellet and sieved to 32 to 48 mesh. The rearrangement was carried out in a feed bed reactor operated at atmospheric pressure.
Conversion [% ]
3. Results and discussion Figure 1 shows the relationship between CHO conversion, CL selectivity and process time (time on stream) over TS-ls wife different Si/Ti ratio and SSZ-41»The result over ZSM-5 (Si/Al ratio=90) is also represented in Figure 1. The CHO conversion decreases with process time, whereas the CL selectivity is almost constant during the process time. The deactivation of SSZ-31 is largest among the zeolites. The CL selectivity over SSZ-31 is lowest among the zeolites. The catalyst deactivation of TS-1(45) is larger than that of TS-l(200). These results suggest that the acidity and micro pore size of the zeolite simultaneously affected the catalyst deactivation.
Selectivity [% ]
Table 1 Properties of zeolites Pore size [nm] Catalyst Surf.area[m2/g] Si/Ti ratio [-] Acidity [mmol/g] TS-1 (45) 0.055 48 TS-1 (90) 0.016 96 0.54*0.56 250-300 TS-1 (200) 221 0.0092 TS-1 (300) 305 0.0040 330 0.57*0.86 SSZ-31 Si/Al = 962 0.026 The zeolite fixed into the reactor was dried by carbon dioxide at 623K for lh. The CHO dissolved into methanol (50 mass%) was supplied by a micro feeder at constant flow rate to evaporator. The vapor diluted by carbon dioxide (20mol%) was fed to the reactor. The reactor effluent was collected at prescribed time intervals up to 2fa and analyzed by a gas chromatograph equipped gkss with column (40m). The used zeolite was regenerated by air atmosphere or argon atmosphere at 723K ftr 15h. The rearrangement was repeated over the regenerated zeolite. When the zeolite was regenerated in argon atmosphere, the volatile coke precursor was removed from the catalyst surface. On the other hand, when the regeneration was carried out in air atmosphere, the coke precursor and non-volatile hard coke were simultaneously removed. The acidity and surface area of the used zeolite were measured to obtain the information for the catalyst deactivation. Furthermore, coke content of the used zeolite was measured by a thermal gravimetric balance. The preciously weighed used zeolite was placed into the cell of the gravimetric balance. After the cell was dried at 773K for 2h in argon atmosphere, air was fed into the balance at the same temperature. The coke content was calculated from the weight loss and the remained zeolite.
-»-TS-l(45) TS-1(45) -•-TS-K200) TS-1(200) -•-ZSM-5 ZSM-5 -A-SSZ-31 SSZ-31
Process time [mini Figure 1 Relationship between CHO conversion, CL selectivity and process time (<673K)
Conversion, Selectivity [% ]
195 195 2
100 80 60 Conversion : Ar Selectivity : Ar Conversion : Air Selectivity : Air
40 20 0 0
120 60 60 120 120 60 60 120 120 60 60 120 120 60 60 120 120 [min] Process time [min] Figure 2 Relationship between CHO conversion, CL selectivity and regeneration Figure 2 shows the effect of regeneration on CHO conversion and CL selectivity over TS-1 zeolite (200) at 623K. The used catalyst was regenerated wife air atmosphere or argon atmosphere at 673K for 17h. The regeneration was continued 8 or 10 times for the rearrangement. The thermal gravimetric analysis revealed tot the most of coke or coke precursor deposited on the zeolite was removed under the conditions as described above. Although the conversion decreases with process time, the selectivity is almost constant for the both atmospheres. The initial conversion for each regeneration decreases with the repetition number. This result mdicates that the concentration of active sites decreases by the regeneration. The catalyst deactivation for the regeneration is known to be represented by Equation (1) . 120
120 60 60 120
: 10 10 40
AAr: Ar: Average Selectivity AAr: Ar: Deactivation Factor OAir: Air: Average Selectivity • AAir: i r : Deactivation Factor
22 33 44 55 6 6 7 7 8 8 9 9 Number of regeneration [-] [-]
Average selectivity [%]
6 6 a A
Average selectivity [%]
Deactivation eactivation factor × 10 [s ]
Deactivation factor × 10 [s ]
60 100 40 20 0
10 0 0
Number Number of of regeneration regeneration [-] [-] Figure 4 Relationiliip between average selectivity, deactivation factor and number of regeneration (SSZ-31).
figure 3 Relationship between average selectivity, deactivation factor and number of regeneration (TS-1), * = A«-exp (-*•*) (1) k and kg are the rate constant at any process time and initial state, respectively, b is deactivation constant and t is process time. Figure 3 shows the effect of regeneration number on average CL selectivity and deactivation constant over TS-1 zeolite regenerated with air and argon. The average selectivity and catalyst deactivation constant did not change with the regeneration atmosphere. Figure 4 demonstrates the
196 196 effect of regeneration on CL selectivity and deactivatioii factor over SSZ-31. The CL selectivity monotonously decreased with the regeneration number over SSZ-31, whereas the deactivation factor was constant over the regeneration number. This result indicates the active sites for Beekmarm rearrangement decreased with the regeneration in argon atmosphere.
6 Air Ar
Rate constant × 102 [m3/kg · s]
Rate constant × 102 [m3/kg · s]
•Air Air Ar AAr
2 1 0 0
Number Number of of regeneration regeneration [-] [-]
Figure S Effects of regeneration gases on the relationship between rate constant and number of regeneration (TS-1)
[-] Number of regeneration [-]
Figure 6 Effects of regeneration gates on the relationihip between rate eonitant and number of regeneration (SSZ-31)
Figures 5 and 6 show the relationship between rate constant and regeneration number over TS-1 and SSZ-31, respectively. The rate constant regenerated by argon was almost constant for all regeneration number, whereas the constant for TS-1 regenerated by air gradually decreased. On the other hand, the constant for SSZ-3 1 constantly decreased with the regeneration number. When the coke was removed from the catalyst surface, the activity almost recovered to the initial activity. However, the catalytic activity gradually decreased with the regeneration by the destruction of active sites called silanol nest. Especially, the catalyst was significantly deactivated by the regeneration of air. Since the micro pore of TS-1 and SSZ-31 was smaller than that of CL or CHO, the CHO could not penetrate in to micro pore of TS-1. Then the rearrangement proceeded at the outer surface of the zeolite. On the other hand, since the pore size of SSZ-31 was larger than that of TS-1, the reactant penetrated into the inside of pore and converted to CL, but as the molecular size of product was larger than that of reactant, the product remained into the pore. As a result, the product changed to hard coke. When the zeolite had strong acid sites, the coking rate should increase. On the other hand, the rearrangement occurred on the weak acid sites located at the outer surface over TS-1. The produced coke was easily removed in argon atmosphere. These results suggest that the zeolite with medium pore size and weak acidity is suitable for Beckmann rearrangement Literature cited  H. Sato, KHirose and MKitamura, Nippon Kagaku Kaishi, 198°, 548(1989)  T.Tatsumi and LXDai, Proceedings of 12th Inter.Zeolite Conference, VoLII, 1455(1998)  D.Shouro, Y.Moriaya, T.Nakajima and SMishima, ApplCataL, A General, 198,275(2000)  HJchihashi and M.Kitamura, CatalToday, 73,23(2002)  T.Takahashi, M.N.A.Nasutionand "LKai, Appl.Catal.,A, General, 210,339(2001)  Takahashi,T. and T.Kai, JJpn.Petml.Inst, 47,190 (2004)