612
Ind. Eng. Chem. Fundam. 1986, 2 5 , 612-620
(60) Levashov, A. V.; Klyachko, N. L.; Martinek, K. Bioorg. Khim. 1981, 7 , 670-679. (61) Levashov, A. V.; Klyachko, N. L.; Pantin, V. I.; Khmel'nitski, Y. L.; Martinek, K. Bioorg. Khim. 1980, 6 , 929-943. (62) Levashov, A. V.; Khmel'nitski. Y. L.; Klyachko. N. L.; Chernyak. V. Y.; Berezin. I. V.; Martinek, K. Doki. Akad. Nauk SSSR 1981, 259, 485-488. (63) Levashov, A. V.; Khmel'nitski, Y. L.; Klyachko, N. L.; Martinek, K. I n Surfactants in Solution; Mittal, K. L., Lindman, B., Eds.; Plenum: New York. 1984; pp 1069-1091. (64) Levashov, A. V.; Pantin, V. I.; Martinek, K.; Berezin, I . V. Doki. Akad. Nauk SSSR 1980, 252, 133-136. (65) Levashov, A. V.; Khmel'nitski, Y. L.; Klyachko. N. L.; Chernyak, V. Y.; Martinek, K. Anal. Biochem. 1981. 118, 42-46. (66) Levashov, A. V.; Khmel'nitski, Y. L.; Klyachko, N. L.; Chernyak, V. Y.; Martinek, K. J . Colloid Interface Sci. 1982, 88, 444-457. (67) Luisi, P. L.; Lljthi, P.; Tomka, I,: Prenosil, J.; Pande, A. I n Enzyme Engineering 7 ; Lasion, A. I., Tsao, G. T., Wingard, L. 5.. Jr., Eds.; New York Academy of Sciences: New York, 1984; pp 549-557. (68) Luisi, P. L.; Bonner, F. J.: Pellegrini, A . ; Wignet, P.; Wolf, R. Heiv. Chim. Acta 1979, 6 2 , 740-753. (69) Luisi, P. L.; Henninger, F.; Joppich. M.; Dossena, A,; Casnati, G. Biochem. Biophys. Res. Commun. 1977, 74, 1384-1389. (70) Luisi, P. L. Angew. Chem., Int. Ed. Engi. 1985, 2 4 , 439-450. (71) Luthi, P.; Luisi, P. L. J . Am. Chem. SOC. 1984, 106, 7285-7286. (72) Malakhova, E. A.; Kurganov, B. I.; Levashov, A. V.; Berezin. I . V,; Martinek, K. Doki. Akad. Nauk SSSR 1983, 270, 474-477. (73) Martinek, K.; Levashov, A. V.; Klyachko, N. L.; Berezin, I . V. Doki. Akad. Nauk SSSR 1977, 236, 920-923. (74) Martinek, K.; Khmel'nitski, Y. L.; Levashov. A. V.; Berezin. I.V. Doki. Akad. Nauk SSSR 1981, 2 6 3 , 737-741. (75) Martinek, K.; Levashov, A. V.; Klyachko, N. L.; Pantin. V. I.; Berezin, I . V. Biochim. Biophys. Acta 1981. 657, 277-294. (76) Martinek, K.; Levashov, A. V.;Khmel'nitski, Y. L.; Klyachko. N. L.; Berezin, 1. V. Science 1982, 218, 889-891. (77) Martinek, K.: Semenov. A. N. J . Appi. Biochem. 1981, 3 , 93-126. (78) Martinek. K.; Semenov, A. N. Biochim. Biophys. Acta 1981. 658, 90-101. (79) Martinek, K.; Levashov, A. V.; Kiyachko, N.; Khmelnitski, Y. L.; Berezin, I.V. Eur. J . Biochem. 1986, 155, 453-468. (80) Meier, P.; Luisi, P. L. J . Solid-Phase Biochem. 1980, 5 , 269-262. ( 8 1 ) Menger, F. M.: Yamada, K. J . Am. Chem. SOC. 1979, 101, 6731-6734. (82) Misiorowski, R. L.; Wells, M. A. Biochemistry 1974, 73, 4921-4827. (83) Morihara. K.; Oka, T. Biochem. J . 1977, 163, 531-542. (84) Morita, S.:Narita, H.: Matoba, T.: Kito. M. J . A m . Oil. Chem. SOC. 1984, 61, 1571-1574. (85) Nilsson, K.; Mosbach, K. Biotechnoi. Bioeng. 1984, 2 6 , 1146-1154. (86) Ohshima, A.: Narita, H.; Kito. M. J Biochem. (Tokyo) 1983, 9 3 . 1421-1425. (87) Omata, T.; Iwamoto, N.; Kimura, T.: Tanaka, A,; Fukui, S.Eur. J . Appl. Microbioi. Biotechnol. 1981, 1 7 , 199-204
(88) Oyama, K.; Kihara, K. CHEMTECH 1984, 16, 100-105. (89) Pellegrini, A.; Luisi, P. L. Biopolymers 1978, 17, 2573-2580. (90) Pelsy. G.; Klibanov. A. M. Biochim. Biophys. Acta 1983, 742, 352-357. (91) Pileni, M. P. Chem. Phys. Lett. 1981, 81, 603-605. (92) Pileni, M. P.; Zemb. T.: Petit, C. Chem. Phys. Lett. 1985, 118, 414-420. (93) Porter, R.; Clark, S., Eds. Enzymes in Organic Synthesis; Pitman: London, 1985. (94) Ramakrishnan. V. R.; Darzon, A.; Montal, M. J . Bioi. Chem. 1983, 258, 4857-4860. (95) Schindler, J . ; Schmid, R. D. Process Biochem. 1982, 17, 2-6. (96) Schonfeld. M.; Montal, M.; Feher, G. Biochemistry 1980, 19, 1535- 1542. (97) Schutt, H.; Schmidt-Kastner, G.; Arens, A,; Preiss, M. Biotechnoi. Bioeng. 1985, 2 7 , 420-433. (98) Schwartz. R. D.: McCoy, C. J. Appi. Environ. Microbioi. 1977, 3 4 , 47-49. (99) Scopes, R. Protein Purification Principles and Practice; Springer-Verlag: New York, 1982; p 199. (100) Semenov. A. N.; Berezin, I . V.; Martinek, K. Biotechnoi. Bioeng. 1981. 2 3 , 355-360. (1 01) Shield, J., Massachusetts Institute of Technology, unpublished results, 1986. (102) Svedes, V.; Galaev, I . U. Russ. Chem. Rev. (Engl. Transi.) 1983, 5 2 , 1184-1202. (103) Tamamushi, B.; Watanabe, N. Colloid Polym. Sci. 1980, 258, 174- 178. (104) Tramper, J. Trends Biotechnoi. 1985, 3 , 45-50. (105) van't Riet, K.; Dekker, M. Preprint from Proceedings of the 3rd European Congress on Biotechnology, Mnchen, Sept. 10- 14, 1984, (106) Vieth, W. R.; Venkatasubramanian, K. CHEMTECH 1973, 677-684. (107) Visser, A. J. W. G.; Fendler. Janos H. J . Phys. Chem. 1982, 8 6 , 947-950. (106) Whitesides, G. M.; Wong, C. Aldrichimica Acta 1983, 16. 27-34. (109) Whitesides, G. M. I n Techniques of Chemistry; Jones, J. B., Sih, C. J., Perlman, D., Eds.; Wiley: New York, 1976; Vol. IO, pp 901-927. (1 10) Wolf, R.; Luisi, P. L. Biochem. Biophys. Res. Commun. 1979, 8 9 , 209-217. (1 11) Wong, M.; Thomas, J. K.; Nowak, T. J . Am. Chem. SOC.1977, 9 9 , 4730-4735. (112) Yokozeki, K.; Yamanaka. S.; Takinami, K.; Hirose, Y.; Tanaka, A,: Sonomoto, K.; Fukui, S. Eur. J . Appi. Microbioi. Biotechnoi. 1982, 1 4 . 1-5. (113) Zaks. A.; Klibanov, A. M. Science 1984, 224, 1249-1251. (114) Zaks, A.; Klibanov, A. M. Proc. Nati. Acad. Sci. U . S . A . 1985. 8 2 , 3 192-3 196.
Received f o r review June 16, 1986 Revised manuscript received July 17, 1986 Accepted August 11, 1986
Vanadium-Phosphorus-Oxygen Industrial Catalysts for n -Butane Oxidation: Characterization and Kinetic Measurements Robert W. Wenig and Glenn L. Schrader' Department of Chemical Engineering and Ames Laboratory-US. Department of Energy, Iowa State University, Ames, Iowa 50011
Vanadium-phosphorus-oxygen (V-P-0) catalysts, prepared in alcohol solution with P-to-V ratios from 0.9 to 1.2, were characterized by X-ray diffraction, infrared spectroscopy, laser Raman spectroscopy, scanning electron microscopy, and X-ray photoelectron spectroscopy. A strong effect of P-to-V synthesis ratio on catalyst structure, catalyst morphology, vanadium oxidation state, and reactivity in n-butane selective oxidation to maleic anhydride was observed. A slight "excess" of catalyst phosphorus (P/V = 1.1 catalyst) was found to stabilize a (VO)zPzO, phase which was active and selective in n-butane oxidation; a larger excess of phosphorus was found to enhance selectivity of the (VO),P,O, phase at the expense of catalyst activity. Used V-P-0 catalysts without "excess" phosphorus contained the active but nonselective a-VOPO, phase.
Introduction Vanadium-phosphorus mixed-oxide catalysts are used industrially for the selective oxidation of C4 hydrocarbons to maleic anhydride. The development of V-P-0 catalysts 0196-4313/86/1025-0612$01.50/0
which are capable of activating paraffin feedstocks is an important recent advance. Catalysts which are prepared in organic media often possess large surface areas and reduced vanadiurn(1V) (E 1986
American Chemical Society
Ind. Eng. Chem. Fundam., Vol. 25, No. 4, 1986
phases which are active and selective in n-butane oxidation (Katsumoto and Marquis, 1979; Cavani et al., 1984; Centi et al., 1984). The effects of the reducing agent, solvent, activation conditions, and phosphorus-to-vanadium (Pto-V) ratio on phase structure and reactivity are poorly understood (Schneider, 1975; Hodnett, 1985). This paper presents a systematic study of the effects of nominal (preparative) P-to-V ratio on phase composition, vanadium oxidation state, catalyst morphology, and catalyst reactivity for n-butane selective oxidation by v-P-0 catalysts prepared in organic solution.
Experimental Section Catalyst Preparation. Four vanadium-phosphorus mixed-oxide catalysts were prepared according to the preparation method reported by Katsumoto and Marquis (1979). P-to-V ratios of 0.9, 1.0, 1.1and 1.2 were selected for the catalyst preparations. The precipitation equipment included a heated (100-mL) three-neck, round-bottom flask, a cooled reflux condenser, and a water trap. Thirty grams of V205was suspended by rapid stirring in 60 mL of isobutyl alcohol and 40 mL of benzyl alcohol. The vanadium oxide-alcohol mixture was refluxed for 3 h at 120 "C. During this period the solution changed in color from brown to black, and approximately 2 mL of water was collected and removed. The apparently reduced vanadium suspension was cooled to 40 "C prior to the addition of orthophosphoric acid. The suspension was again heated to 120 "C and refluxed for an additional 3 h. During this time a reduced vanadium phosphate (probably V4+)was formed, as indicated by a change in color from black to blue. The resulting catalyst was separated by filtration, dried for 12 h at 125 "C, and activated in air for 3 h at 380 "C. Surface Area Measurements. Surface areas were measured by using a Micromeretics 2100E Accusorb instrument with nitrogen as the adsorbate. X-ray Diffraction. X-ray powder diffraction patterns were obtained with a Siemans D500 diffractometer using Cu K a radiation. The samples were rotated in the X-ray beam. Fourier-Transform Infrared Spectroscopy. Transmission infrared spectra were recorded on a Nicolet 60-SX Fourier-transform infrared spectrometer with single-beam optics. Each spectrum represents the accumulation of 100 scans at 1-cm-' resolution. Laser Raman Spectroscopy. Raman spectra were obtained in the backscattering mode with a Spex 1403 laser Raman spectrometer. The 514.3-nm line of a Spectra Physics argon ion laser operating at 200-mW output was the primary excitation source. A spectral resolution of 5 cm-' and a central slit setting of 60 pm were used. Spectral accumulation was provided by a Nicolet 1180E computer system. All spectra reported were obtained with 50 scans a t scan drives of 3.125 cm-'/s. Scanning Electron Microscopy. Scanning electron microscopy studies were conducted using a Joel Model JSM-U3 scanning electron microscope. Samples were prepared by depositing catalyst powder onto a thin film of carbon black on a graphite sample mount. A gold thickness of 300 A was then sputtered onto the sample mount. Micrographs were obtained at a potential difference of 25 kV. A Tracor/Northern 2000 microanalyzer was used for the energy dispersive X-ray analysis. X-ray Photoelectron Spectroscopy. X-ray photoelectron spectra were obtained with an AEI 200B spectrometer using A1 K a radiation. Signal averaging was performed using a Nicolet 1180 computer system; all samples were referenced to the carbon 1s bending energy
613
Table I. BET Surface Areas (m*/g) of Precipitated and Air-Activated V-P-0 Catalysts PIV ratio 0.9 1.0 1.1 1.2 14.3 4.5 1.6 precipitated catalyst 3.2 air-activated catalyst 12.5 16.8 33.0 60.9
of 284.6 eV. Catalyst samples were sealed in evacuated Pyrex tubes which were opened in a helium drybox attached directly to the photoelectron spectrometer.
Catalyst Activity Measurements The selective oxidation of n-butane to maleic anhydride was performed using a continuous-flow, fixed-bed reactor and a gas chromatograph analytical system reported by Ozkan and Schrader (1985). The reactor was packed with 2 g of catalyst homogeneously mixed with 40 g of Sic. Both the catalyst and the S i c particles were in the 2-mm size range. Reaction temperatures from 350 to 450 "C were investigated at a space velocity of 750 h-l. The following reactor conditions were established for each experimental run: oxygen partial pressure, 0.22 atm; n-butane partial pressure, 0.015 atm; nitrogen partial pressure, 0.765 atm; volumetric flow rate of n-butane, 540 cm3 (STP)/h; total molar flow rate, 1.61 g-mol/h. For these studies the percent conversion is defined as moles of n-butane consumed x 100% moles of n-butane in feed The selectivity to product A is defined as moles of A produced 1 x - x 100% moles of n-butane consumed y where y is the ratio of the number of C atoms in the reactant to the number of C atoms in the product. Rates of n-butane disappearance, maleic anhydride formation, and COz formation are based on the surface area of the used catalyst. Characterization of Precipitated Catalysts Surface Area Measurements. The BET surface area measurements for the precipitated and air-activated catalysts are presented in Table I. In general, an increase in nominal P-to-V ratio was found to result in somewhat lower surface areas for the precipitated catalyst and significantly higher surface areas for the air-activated catalysts. Surface areas in excess of 60 m2/g were observed for the air-activated P / V = 1.2 catalyst. X-ray Diffraction. The X-ray powder diffraction patterns of the precipitated catalysts having compositions P / V = 0.9, 1.0, 1.1,and 1.2 are given in Figure 1, a, b, c, and d, respectively. Each of the precipitated catalysts was found to possess the X-ray diffraction lines at 5.72, 4.53, 3.68, 3.29, 3.11, 2.94, 2.79, and 2.66 A reported by Stefani and Fontana (1978) and by Cavani et al. (1984) for the B-phase precursor. Additional X-ray diffraction lines were observed at 2.40, 1.91, 1.85, and 1.76 A. The current literature describes the precatalyst as a VO(HP04).0.5H20 phase (Johnson et al., 1984). Following 3 h of activation in air, an extensive transformation of the VO(HPO4).0.5H2O phase was observed for the P / V = 0.9 catalyst; only small changes were observed in the XRD pattern of the P / V = 1.0, 1.1, and 1.2 catalysts. The air-activated P / V = 0.9 catalyst (Figure 2a) was found to have principal XRD lines at 4.72, 4.57, 4.04, 3.67, 3.13, 2.96, and 2.61 A. These lines are assigned to an oxidized form of the VO(HP04).0.5H20phase referred to as the B' phase by Hutchings and Higgins (1980). The activated P/V = 1.0, 1.1,and 1.2 catalysts (Figure 2, b, c, and d) were found to have very weak lines at these posi-
614
Ind. Eng. Chem. Fundam., Vol. 25, No. 4, 1986
ALA -1 1 , (0)
5.835
4.436
2.976
2.252
-
-
*
:.a23
~
~
1.54
(A,
a Spacing
I
Figure 1. X-ray diffraction patterns of precipitated catalysts: (a) P ' V = 0.9,tb) P ' V = 1.0. (d) P,'V = 1.1. and (d) P,'V = 1.2.
I
200
1
1
I
8.838
400
600
800
1000
1200
Wavenumbers ( c r n - 1 )
Figure 3. Raman spectra of air-activated catalysts having composition (a) P/V = 0.9 and (b) P/V = 1.0.
/+Jb (a)
4.436
2.976 d spacing
2.252
!.S23
1.541
(X!
Figure 2. X-ray diffraction patterns of air-activated catalysts: ta) I'!V = tJ.:j. l b ) P,'V = 1.0. i c ) P '1-= 1.1. and (d) P/\' = 1.2.
tione as well as stronger VO(HP04).0.5H20 diffraction lines. The P/V = 0.9 catalyst was completely oxidized to the B' phase, whereas the catalysts having greater phosphorus loadings exhibited only trace anlounts of this oxidized V5+phase. Laser Rainan Spectroscopy. Characterization of the air-calcined catalysts having P-to-V ratios of 0.9 and 1.0 was obtained by laser R m a n spectroscopy (Figure 3, a and ti. respectively). Other air-calcined and precipitated catalyst samples did not yield informat,ive Raman spectra due to fluorescence. Raman bands were observed for the aircalcined P / V = 0.9 catalyst at 287, 305, 406, 486, 529. 702, and 994 cm-'. These bands were assigned to semicrystalline V 2 0 5present in a nearly amorphous B' phase (see Sanchez et al., 1982). Very weak V,05 bands were also observed for the air-calcined P/V = 1.0 catalyst; in addition, a band at 923 cm-.' was assigned to a P-0 stretching vibration of a disordered VO(HP0,).0.5H20 phase. Infrared Spectroscopy. Infrared spectroscopy was used to identify vibrational bands for the precipitated catalysts having P-to-V ratios from 0.9 to 1.2 (Figure 4a-d). Each precatalyst possessed infrared bands at 416,485, 529, ,547. 64.5, 686, 929, 9'75, 1048, 1102, 1133, and 1198 ern-.'. Catalysts prepared with "excess" phosphorus (P/V = 1.1 and P/V = 1.2) displayed additional infrared bands at 696 and 735 cm I. Infrared bands a t 416,485,529,547, and 645 cm-' were assigned to P-0 bending vibrations; the bands at 696 and 7 3 5 cm ' were assigned to P-0 stretching vibrations for catalysts incorporating excess phosphorus. The 686-cm-' band was assigned to H,O bending vibrations, and the 929-.cIil ' infrared hand was associat,ed with a P -(OH) ~
I 1350
I
I200
750 600 Wovenumbers ( c r n - l l
1050
900
450
3
Figure 4. Infrared spectra of precipitated catalysts having compo,ition (a) P'V =- 0 9. (h) P ' V = 1 0 , (c) P I V = 1 1 and (d) P/V = 12
stretch of the VO(HP04).0.5H20phase (Busca et al., 1986). An intense band at 975 cm-' was assigned to a V=O stretching vibration by analogy to the 1018-cm-' V=O vibration observed for V205(Sanchez et al., 1982). A symmetric PO., stretching vibration was assigned t o the 1048-cm-' band. Asymmetric PO, stretching vibrations were observed at 1102, 1133, and 1198 cm These assignments are consistent with those reported for the P20-4ion by Greenfield and Clift (1975) and are made in analogy to model compound assignments reported by Bhargava and Condrate (1977) and Bordes and Courtine (1979). Infrared spectroscopy rvas also used to characterize the air-calcined catalysts having P-to-V ratios from 0.9 to 1.2. The P/V = 0.9 (Figure 5a) catalyst was found to undergo a phase transformation during air activation. Infrared bands were observed at 423,601,642, 932, 993, 1020, 1033. and 1205 cm-'. Coupled V-0 and P-0 bending vibrations were assigned to the 423-cm-I band. Infrared bands at 601 and 642 cm-I were attributed to P-0 bending vibrations of the B' phase identified by X-ray diffraction. The intense band at 993 cm was assigned to a V=O stretching vibration (see Sanchez et al., 1982). A symmetric P-0 stretching vibration was assigned to the 932-cm hand; asymmetric P-0 stretching vibrations were assigned tu infrared bands at 1020, 1033, and 1205 cui '. These aqvgnments are in agreement with those reported hy hloser
'.
Ind. Eng. Chem. Fundam., Vol. 25, No. 4, 1986 615
I 1
1350
\ wi
1200
1050
M I 1
900 750 600 Wavenumbers ( c r n - l i
1
450
300
Figure 5. Infrared spectra of air-activated catalysts having composition (a) P / V = 0.9, (b) P / V = 1.0, (c) P / V = 1.1, and (d) P / V = 1.2. Table 11. V 2p3,2Photoelectron Spectra Binding Energies (eV) for Precipitated and Air-Activated V-P-0 Catalysts P / V ratio 1.1 1.2 0.9 1.0 precipitated catalyst 517.0 517.0 517.2 517.2 air-activated catalyst 517.9 517.6 517.4 517.3
and Schrader (1985) and Bhargava and Condrate (1977) for oxidized V-P-0 phases. Only subtle changes were observed in the infrared spectra of the P / V = 1.0, 1.1, and 1.2 catalysts following air activation (Figure 5, b, c, and d, respectively). An additional P-0 bending vibration was observed at 578 cm-’. The P-0 stretching band observed for precatalysts with “excess” phosphorus at 696 cm-l was no longer apparent. The P-0 bending vibrations at 485,529, and 547 cm-’ were observed to broaden and to decrease in intensity. No changes were observed in the remaining infrared bands. These subtle changes in structure, which occur upon activation in air, seem to indicate mild restructuring and possible distortion of the P2074ion during calcination. Scanning Electron Microscopy. Scanning electron micrographs of the precipitated and air-calcined catalysts were obtained for catalysts having P-to-V ratios of 0.9 and 1.2. The VO(HPO4).0.5H20 precatalyst was found to consist of 3 X 5 X 0.5 pm (approximate) hexagonal platelets. Johnson et al. (1984) has described these catalysts as lamellar platelets whose crystallographic orientation changes upon activation. Representative micrographs of the precipitated and air-activated catalysts are given in Figure 6. X-ray energy dispersive spectroscopy (Figure 7) established that vanadium and phosphorus were welldispersed in a homogeneous mixed-oxide catalyst. X-ray Photoelectron Spectroscopy. X-ray photoelectron spectra of the precipitated and air-calcined catalysts are shown in Figures 8 and 9; Table I1 provides the observed binding energies. Each of the precipitated catalysts (Figure 8a-d) was characterized by a band at 517.C-517.2 cm-l, which was assigned to the 2p3,, binding
energy of vanadium(1V) (see Moser and Schrader, 1986). Multiple oxygen oxidation states are indicated by one or more shoulder bands in the oxygen 1s photoelectron spectrum of the precipitated catalysts. The XPS spectra for the air-calcined catalysts (Figure 9a-d) yielded V 2p3,, binding energies of 517.9, 517.6, 517.4, and 517.3 eV for the P / V = 0.9, 1.0, 1.1, and 1.2 catalysts. These results indicate that the surfaces of the catalysts were oxidized during air activation, especially for the catalysts not having “excess” phosphorus. Kinetic Measurements for n -Butane Oxidation Conversion and Selectivity of V-P-0 Catalysts. Reactivity measurements were obtained at 350-450 “C for vanadium-phosphorus-oxygen catalysts having P-to-V ratios of 0.9, 1.0, 1.1, and 1.2. The percent selectivity for maleic anhydride and percent n-butane conversion are displayed in Figures 10 and 11, respectively. Since only maleic anhydride and carbon dioxide reaction products were observed, the selectivity for carbon dioxide has not been shown (SMA+ SCoB= 100%). The selectivity for maleic anhydride was found to decrease with an increase in reaction temperature and to increase with catalyst phosphorus loading. Consequently, the selectivity for carbon dioxide formation was found to increase with increasing reaction temperature and to decrease with additional catalyst phosphorus content. In general, the n-butane conversion was found to decrease with catalyst P-to-V ratio; however, the P/V = 1.1catalyst displayed an unusually high conversion of n-butane. Product Yields. Maleic anhydride and carbon dioxide yields were obtained as a function of reaction temperature and catalyst phosphorus loading (Table 111). The highest yields of maleic anhydride were obtained with the catalyst having a P-to-V ratio of 1.1. Catalysts having P-to-V ratios of 0.9 and 1.0 displayed intermediate maleic anhydride yields, whereas the catalyst having a P-to-V ratio of 1.2 demonstrated a relatively low yield of maleic anhydride. The highest C 0 2 yields were observed with the catalyst having a P-to-V ratio of 0.9. Catalysts having P-to-V ratios of 1.0 and 1.1displayed intermediate C 0 2yields; very low C 0 2 yields were observed for the P / V = 1.2 catalyst. Specific Rates of Oxidation. Catalyst phosphorus loading and reaction temperature were shown to strongly affect catalyst surface area, activity, and selectivity. Rates of n-butane disappearance, maleic anhydride formation, and carbon dioxide formation were normalized to catalyst surface area (see Figures 12-14). Each of these specific rates was observed to decrease with an increase in catalyst P-to-V ratio. Consequently, an increase in catalyst phosphorus content was found to decrease the specific activity of both selective and nonselective oxidation reactions. Furthermore, since “excess” phosphorus more strongly depressed the specific rate of carbon oxide formation, catalyst selectivity was favored by increased P-to-V ratios. Characterization of Used Catalysts Following n-butane selective oxidation, the used V-P-0 catalysts were cooled to room temperature in the presence
Table 111. Product Yields as a Function of Catalyst P-to-V Ratio and Reaction Temperature carbon dioxide yield, ‘70, maleic anhydride yield, ‘70, for P / V ratios of for P / V ratios of temm ‘C 0.9 1.0 1.1 1.2 0.9 1.0 1.1 0.8 1.3 3.0 5.4 2.1 1.2 350 3.4 3.1 2.0 1.2 1.8 375 4.2 3.7 6.6 1.6 2.0 4.5 7.1 3.7 2.4 400 5.0 5.4 3.7 2.3 2.9 6.0 8.7 425 6.6 7.1 6.0 3.6 4.3 450 8.7 8.1 12.0
1.2 0.4 0.8 1.0 1.6 2.1
616
Ind. Eng. Chem. Fundam., Vol. 25, No. 4. 1986
Figure 6. Scanning electron micrographs of precipitated and air-activated catalysts: (a) precipitated 1'/V = 0.9 catalyst; ill, P ) air-activated PIV = 1.0 catalyst: (d) air-activated P/V = 1.2 catalyst.
of feed gas mixtures. Thus, the used catalysts were not exposed to air until room temperature had been reached. No significant changes in catalyst surface area were observed following reaction. X-ray Diffraction. The X-ray powder diffraction patterns of the used catalysts having composition P/V = 0.9,l.O. 1.1,and 1.2 are shown in Figure 15. The used P/V = 0.9 (Figure 15a) catalyst was found to have broad XRD lines a t 7.2, 6.8, 3.58, 3.10, and 3.04 A. Diffraction lines at these positions correspond to mixtures of V5+ phases. Gopal and Calvo (1972) assign lines a t 3.57, 3.07, and 3.01 to a-VOPO,; diffraction lines a t 3.10, 3.00, and 1.96 A correspond to a'-VOPO,; broad diffraction lines near 7.2 A. are indicative of a VP0,.2H20 phase (Ladwig et al., 1965). The used P/V = 1.0 catalyst (Figure 15b) was found to have XRD lines at 7.2,6.8,3.90,3.15, 3.06, and 3.00 A. The
vanadium(1V) phase(s) described earlier is indicated by diffraction lines a t 7.2,6.8, and 3.06 A. The broad lines at 3.90,3.15, and 3.00 indicate the presence of a patented B-phase catalyst (Schneider, 1975) which seems to closely resemble the (VO)pPz07phase (Hodnett, 1985). Progressively broaded XRD lines at 3.90, 3.15, and 3.00 A indicate smaller B-phase crystallites were formed for used catalysts which incorporate "excess" phosphorus (Figure 15c,d). Laser Raman Spectroscopy. Characterization of used catalysts having P-to-V ratios of 0.9 and 1.2 was obtained by laser Raman spectroscopy (Figure 16). Unreacted Vz05 was observed in the ?/V = 0.9 catalyst (Figure Xa), as indicated by vibrational bands at 287,305,406,486,529, 702, and 994 cm-' (Sanchez et al., 1982). The presence of a second phase was indicated by broad Raman bands at 567, 872, 1014, and 1127 cm-'. These vibrational bands
Ind. Eng. Chem. Fundam.. Vol. 25. No. 4, 1986
I
536
532
S28 524 520 Bindinq .W" 1w .
617
516
Figure 8. X-ray photoelectron spectra of precipitated V-P-0 c a t alysts (a) P/V = 0.9, (b)P/V = 1.0, (e) P/V = 1.1, and (d) P/V = 1.2.
536
532
528 524 520 Binding mwqy I e V I
SI6
Figure 9. X-ray photoelectron spectra of air-activated V-P-0 catalysts (a) P/V = 0.9.(b) P/V = 1.0, (e) P/V = 1.1, and (d) P/V = 1.2.
Figure 7. Scanning electrun micrograph and X-ray energy dispersive spectra of the air-activated P/V = 0.9 catalyst: (a) elemental map for phosphorus; (h) scanning electron micrograph; (c) elemental map for vanadium.
are assigned to a nearly amorphous a-VOP04 phase prepared by precipitation. Raman bands were observed for the used P / V = 1.0 catalyst (Figure 16h) at 435,892,922,934 (shoulder),993, and 1018 cm-'. The strong 922- and 934-cm-l bands were assigned to P-0 stretching vibrations of the (VO)2P,0, phase (Moser and Schrader, 1985). The very weak bands a t 435, 892, and 1018 cm-' suggest P-0 vibrations; the 993-cm-I band can be assigned to the V=O vibration of an amorphous V-P-0 phase. The used P / V = 1.1and 1.2 catalysts (Figure 16, e and d) were found to exhibit very broad and weak Raman bands near 922 cm-' which are characteristic of the (VO)2Pz07phase. The progressive broadening of these P-0 stretching hands with increased P-to-V ratio indicates a
90
-
-8 --*- 8 0 -
-
.-> u
E
p/v=I. I
70-
VI
KO
50
-
PlV.O.9
350
375
400
Temperature
425
450
IT1
Figure 10. V-P-0 catalyst selectivity for maleic anhydride as a function of catalyst P-to-V ratio and reaction temperature.
distortion of the P20,p crystal environment by =excessn catalyst phosphorus.
Ind. Eng. Chern. Fundarn., Vol. 25, No. 4, 1986
618
22
t
PlV=I.I
I 0.5 P/V= 1.0
I
,
1
350
400
375
Temperature
425
I
P/V=I.l
0. I
PIV: 1.2
325
450
350
375
400
425
-
450
Temperature ("C1
("Ci
Figure 11. n-Butane conversion as a function of catalyst P-to-V ratio and reaction temperature.
h E
0.3
2.6 p?=0-9
Figure 14. Specific rates of carbon dioxide formation as a function of catalyst P-to-V ratio and reaction temperature.
I m
c
0 3
0
350
375
400
425
450
Temperature ("C)
8.838
Figure 12. Specific rates of n-butane disappearance as a function of catalyst P-to-V ratio and reaction temperature.
L
2'
E --z y z
1.3 I./
w u 2 0.9
-
I .o
-
a,
mp.
E
5 0.7
-c: E
1.1
-
0.5 -
01
' 0.3
-
1.2
c
0
=
0.1 -
350
375
400
425
450
Temperature ( " C )
Figure 13. Specific rates of maleic anhydride formation as a function of catalyst P-to-V ratio and reaction temperature.
Infrared Spectroscopy. Fourier-transform infrared spectroscopy was employed to characterize the used VP-0 catalysts (Figure 17). The used P / V = 0.9 catalyst (Figure 17a) was found to have infrared bands at 565,679, 749,956,990,1040,1088, and 1171 cm-'. Vibrational bands at these frequencies are in agreement with stretching vibrations of the PO:- group; a V=O stretching vibration is indicated by an infrared band at 990 cm-l. The infrared spectrum of the used P / V = 0.9 catalyst can be assigned to the a-VOPO, phase (Centi et al., 1984; Shimoda et al., 1985). The used catalysts having composition P / V = 1.0, 1.1, and 1.2 were found to have broad infrared bands a t 510, 576,635,743,967,1078,1137,1249, and 1268 cm-' (Figure
4.436
2.976
2.252
1.823
1.541
(A) Figure 15. X-ray diffraction patterns of used catalysts: (a) P / V = 0.9, (b) P / V = 1.0, (c) P/V = 1.1,and (d) P / V = 1.2. d Spacing
17b-d). The infrared spectrum of these catalysts is in agreement with that reported by Centi et al. (1984) for (VO)2P207and mixtures of (VO)2P207with the B phase. Scanning Electron Microscopy. Scanning electron micrographs of the used V-P-0 catalysts having P-to-V ratios of 0.9, 1.0, 1.1,and 1.2 are shown in Figure 18. A strong effect of nominal P-to-V ratio on catalyst morphology was observed. The used P / V = 0.9 catalyst (Figure 18a) was found to consist of irregular platelets while the used P / V = 1.2 (Figure 18d) catalyst displayed a block-style morphology. The used P / V = 1.0 and 1.1 catalysts (Figure 18, b and c) were found to have long-range structure which indicated a transition from platelet to block morphology. X-ray Photoelectron Spectroscopy. X-ray photoelectron spectra of the used V-P-0 catalysts are shown in Figure 19. The used catalysts were characterized by V 2pSi2binding energies of 517.6, 517.5, 517.4, and 517.3 eV for the P / V = 0.9, 1.0, 1.1, and 1.2 catalysts, respectively. A decrease in binding energy from 517.9 to 517.6 eV for the P/V = 0.9 catalyst (Figure 19a) and from 517.6 to 517.5 eV for the P/V = 1.0 catalyst (Figure 19b) suggests that a partial reduction of the surface occurred during hydrocarbon-air flow. No change in V 2p3 binding energy was observed for the used catalysts with "excess" phosphorus having composition P / V = 1.1and 1.2 (Figure 19, c and d). Discussion Kinetic measurements with vanadium-phosphorusoxygen catalysts of varying P-to-V ratio have shown that
Ind. Eng. Chem. Fundam., Vol. 25. No. 4, 1986
L
200
400 600 800 IO00 W0"en"nbn I L r n ~ l J
I
Figure 16. Raman spectra of used catalysts having composition (a) P/V = 0.9, (b) P/V = 1.0,(e) P/V = 1.1, and (d) P/V = 1.2.
Id 1
M
1350
1200
1050
900
750
600
450
300
Wovenumberr Icm-IJ
Figure 17. Infrared spectra of used catalyata having composition (a) P/V = 0.9, (b) P/V = 1.0, (e) P/V = 1.1, and (d) P/V = 1.2.
catalyst phosphorus loading is a key parameter in determining catalyst selectivity, catalyst activity, and specific rates of oxidation. Catalyst characterization studies provide structural and compositional explanations for the observed kinetic results. The precipitated P/V = 0.9 catalyat was found to consist of VO(HP0,).0.5H20 with vanadium(1V) present on the catalyst surface. Following activation in air, an oxidized B phase was formed with both surface and bulk vanadium(V). Used catalysts having composition P/V = 0.9 were found to contain unreacted V205and a-VOPO,; the surfaces of these catalysts possessed both vanadium(1V) and vanadium(V) states. The low selectivity for maleic anhydride and the moderate conversion of n-butane observed for the P/V = 0.9 catalyst are attributed to the formation of active, but relatively nonselective, a-VOPO,. High specific rates of n-butane consumption, maleic anhydride formation, and C02formation appear t o occur as a result
E19
620
Ind. Eng. Chem. Fundam., Vol. 25, No. 4, 1986
V-P-0 catalyst is made by incorporating just enough “excess” phosphorus to stabilize the active and selective (VO)zPz07phase. Too much excess phosphorus can dramatically reduce n-butane conversion by “blocking” the selective as well as the nonselective oxidation pathways. The very low specific rates of n-butane consumption, maleic anhydride formation, and C 0 2 formation observed for the P/V = 1.1and 1.2 catalysts indicate a dramatic reduction in catalyst activity with the incorporation of “excess” phosphorus.
I
536
I
532
1
I
1
528
524
516
520 Bindinq energy (eV)
Figure 19. X - r a y photoelectron spectra of used V-P-0 catalysts: (a) P / V = 0.9, (b) P / V = 1.0, (c) P / V = 1.1, a n d (d) P / V = 1.2.
of only mild suppression of activity by catalyst phosphorus. The precipitated and air-calcined P / V = 1.0 catalysts were found to contain VO(HPO4).0.5HzO;some B’ phase was detected in the air-calcined samples. XPS results indicated the surface of the precipitated catalyst was vanadium(IV), while both vanadium(1V) and vanadium(V) were observed for the air-calcined and used P / V = 1.0 catalysts. Intermediate values of maleic anhydride selectivity and n-butane conversion are attributed to the formation of (V0)2P207with some a-VOPO, in the used P/V = 1.0 catalyst. Intermediate specific rates of n-butane consumption, maleic anhydride formation, and COz formation result from a moderate suppression of activity by phosphorus. The precipitated and air-calcined P / V = 1.1 and 1.2 catalysts were found to contain VO(HP04).0.5H,0. XPS results indicated the surface of the precipitated catalysts was primarily vanadium(1V). A slight oxidation of the surface was indicated during air activation; no change in vanadium oxidation state was observed for the used VP-O catalysts containing “excess” phosphorus. High values for maleic anhydride selectivity were observed for both the P/V = 1.1and 1.2 catalysts. However, n-butane conversion was found to be highest for the P/V = 1.1catalyst and very low for the P/V = 1.2 catalyst. It appears that the optimal
Conclusions These studies have established that a slight “excess” of catalyst phosphorus (P/V = 1.1)is required to stabilize a (VO),PzO, phase which is both active and selective in the oxidation of n-butane to maleic anhydride. A larger excess of catalyst phosphorus was found to further enhance the selectivity of this phase, but this was accomplished only at the expense of a loss of catalyst activity. Used V-P-0 catalysts without “excess” phosphorus contained the active but nonselective a-VOPO, phase. Specific rates of COz formation and maleic anhydride formation were found to decrease with catalyst P-to-V ratio. Since COz formation rates were more strongly depressed by added phosphorus, maleic anhydride selectivity was favored by an increase in catalyst phosphorus. Registry No. V a n a d i u m oxide, 11099-11-9; phosphorus oxide, 12640-86-7;H3C(CH2)2CH3, 106-97-8; maleic anhydride, 108-31-6.
Literature Cited Bhargava, R. N.; Condrate, R. A. Appl. Spectrosc. 1977, 3 1 , 230. Bordes, E . ; Courtine, P. J . Catal. 1979, 57,236. Busca, G.; Cavani, F.; Centi, G.; Trifiro, F. J . Catal. 1986, 9 9 , 400. Cavani, F.; Centi, G.; Trifiro, F. Appl. Catal. 1984, 9 , 191. Centi, G.;Manenti, I.; Riva. A,; Trifiro, F. Appl. Catal. 1984, 9 , 177. Gopal, R.; Calvo, C. J . Solid State Chem. 1972, 5 , 432. Greenfield, S.;Clift, M. Analtyical Chemistry of the Condensed Phosphate; Pergamon: Oxford, 1975. Hodnett, B. K. Catal. Rev.-Sci. Eng. 1985, 27,373. Hutchings, G.; Higgins, R . US. Patent 4 209 423, 1980. Johnson, J.; Johnston, D.; Jacobson, A,; Brody, J. J . Am. Chem. SOC. 1984, 106, 8123. Katsumoto, K.; Marquis, M. U S . Patent 4 132670, 1979. Ladwig, V. Z . Anorg. Allg. Chem. 1985, 338, 266. Moser, T.; Schrader, G. J . Catal. 1985, 92. 216. Moser, T.; Schrader, G. J . Catal.. in press. Ozkan, U.; Schrader, G. J . Catal. 1985, 95, 120. Sanchez, C.; Livage, J.; Lucazeau, G. J . Raman Spectrosc. 1982, 72,68. Schneider, R. US. Patent 3 864 280, 1975. Shimoda, T.; Okuhara, T.; Misono, M. Bull. Chem. SOC. Jpn. 1985, 58. 2163. Stefani, G.; Fontana, P. U.S. Patent 4 100 106, 1978.
Received for reuiew J u n e 20, 1986 Accepted July 21, 1986
