J. Phys. Chem. 1988, 92, 6052-6058
6052
Surface Active Sites of V,O,-WO,
Catalysts
Atsushi Satsuma,* Atsushi Hattori, Koichi Mizutani, Akio Furuta,t Akira Miyamoto,* Tadashi Hattori, and Yuichi Murakami Department of Synthetic Chemistry, Faculty of Engineering, Nagoya University, Furo-cho, Chikusa- ku, Nagoya 464, Japan, and Kinu-ura Research Department, JGC Corporation, Sunosaki-cho, Handa, Aichi 475, Japan (Received: December 16, 1987)
The surface active sites of V2O5-WO3catalysts were investigated by means of secondary ion mass spectrometry (SIMS), IR spectroscopy of adsorbed NH,, temperature-programmeddesorption (TPD) of NH,, and the NARP (NO-NH, rectangular pulse) technique, and the bulk and surface of the catalysts were characterized by means of X-ray diffraction (XRD), IR, electron spin resonance (ESR), and X-ray photoelectron spectroscopy (XPS). Although V20sand WO, are not well-mixed in the bulk of the catalysts, V and W ions are well-mixed in the surface of the catalysts, causing a remarkable promoting effect of WO, on the surface active sites. The surface concentration of the redox sites measured by the NARP technique increased with increasing W 0 3 content, and, at high content of WO,, the surface concentration of the redox sites was much larger than the surface concentration of V ions. By means of SIMS, it was verified that the oxygen species bonded to the surface W ions also act as the redox sites in the reduction of NO with NH3 at such low temperature that the reaction cannot proceed on WO, itself. It was concluded that the redox sites on V2OS-WO3catalysts are not only the surface V 4 species but also the surface W 4 species interacting with vanadium ions. As for the acid sites examined by TPD of NH,, both the surface concentration and the strength are not changed by the addition of WO,.
Introduction Since the catalytic property is directly related to the surface active sites of catalysts, the investigation of the active sites will lead to a better understanding of the catalysis and to the science-based improvement of catalysts. As for V2OScatalysts, it is well-known that the surface V V species are active as the redox sites for the selective oxidations of hydrocarbons’ and the reduction of N O with NH3.233 Various oxides are used as promoters for the V2O5 catalyst to improve the catalytic property. However, little is known about the effect of the promoters on the active sites, especially on the redox sites. We have shown in a previous paper4 that the addition of P2OS to V20s increases the concentration of the surface V=O species, and, at high P205content, all of the surface vanadium and phosphorus ions act as the active sites: all the surface P ions form P-OH species which act as Brcansted acid, and all the surface V ions form the surface V=O species, Le., the redox sites. Thus, although the addition of P2OSon V20s catalyst increases the redox sites, phosphorus ions do not form the redox sites. This seems to be reasonable, since P20sitself does not have the redox property. WO, is also used as a promoter of V2OScatalysts for the oxidation of benzene’ and o-xylene,6 the ammoxidation of mxylene,’ and the reduction of N O with NH3 at high temperature.8 In contrast to P2O5, W 0 3 itself has a redox activity and is a good catalyst for the selective oxidation of hydrocarbon^.^ Therefore, it can be expected that W 0 3 will give a different promoting effect on V2OScatalysts from that of PZOS, especially on the redox property. The purpose of the present study is to reveal the effect of W 0 3 on V,0s-W03 catalysts, especially the surface active sites. The redox sites are characterized by the NARP technique and SIMS, and the acid sites are characterized by TPD of N H 3 and IR of adsorbed NH3. Experimental Section The V2OS-WO3catalysts containing various amounts of WO, were prepared by mixing an oxalic acid solution of NH4V03with (NH4),oW12041.5H20 followed by drying and subsequent calcination in flowing O2 at 773 K for 3 h. The WO, contents of W/(V + W) were 0.1,0.3,0.5,0.7,and 0.9. V2Os and WO, were prepared by the thermal decomposition of NH4VO3 and (NH4)loW,204,-5H20in flowing O2at 773 K for 3 h, respectively. JGC CorDoration. Present address: Department of Hydrocarbon Chemistry, Faculty of Engineering, Kyoto University, Sakyo-ku, Kyoto 606, Japan. f
0022-3654188 12092-6052$01.SO10
The catalysts were then pressed and sieved in the range 28-48 mesh. X-ray diffraction diagrams of the catalysts were obtained with a Rigaku G F 2035 X-ray diffractometer with Cu target. IR spectra of catalysts were recorded with a Jasco FT/IR-3 Fourier transfer infrared spectrometer with KBr as a diluent. ESR absorption measurements were made at X band with a JEOL M E 1X spectrometer at room temperature. X-ray photoelectron spectra (XPS) of the catalyst were measured with a Shimadzu ESCA-750 with Mg KP radiation (1253.6 eV). The binding energy was corrected with reference to the peak at 285.0 eV for C( Is), and the surface composition was determined from the observed peak intensities by the known value of the total cross section of individual atoms.10 SIMS spectra of catalysts were recorded with a VG ESCA LAB-5 under the following conditions: primary ion, Ar’; ion current, 100 nA, resolution, 6.5or 8.6; source energy, 4.5 kV. To avoid charging of the samples, charge neutralizer was used under optimum conditions for each sample. (1) (a) Cole, D. J.; Cullis, C. F.; Hucknall, D. J. J. Chem. SOC.,Faraday Trans. 1 1976, 72, 2185. (b) Akimoto, M.; Usami, M.; Echigoya, E. Bull. Chem. SOC.Jpn. 1978,51,2195. (c) Bond, G. C.; Sirkiny, A. J.; Parfitt, G. D. J. Catal. 1979,57,476. Bond, G . C.; Konig, P. Ibid. 1982, 77, 309. (d) Anpo, M.; Tanahashi, I.; Kubokawa, Y. J . Phys. Chem. 1980,84,3440. (e) Anderson, A,; Lundin, S. T. J . Catal. 1979, 58, 383; 1980, 65, 9. Jonson, B.; Rebenstorf, B.; Larsson, R.; Anderson, S. L. T.; Lundin, S. T. J. Chem. SOC.,Faraday Trans. 1 1986,82, 767. (f) Mori, K.; Inomata, M.; Miyamoto, A,; Murakami, Y. J. Phys. Chem. 1983,87, 4560; J . Chem. SOC.,Faraday Trans I 1984, 80, 2655. Mori, K.; Miyamoto, A.; Murakami, Y. J . Phys. Chem. 1984,88,2735,2741,5232;1985,89,4265. (g) Tanaka, T.; Om,M.; Funabiki, T.; Yoshida, S. J . Chem. SOC.,Faraday Trans. I 1986, 82, 35. Tanaka, T.; Tsuchitani, R.; Om, M.; Funabiki, T.; Yoshida, S. J . Phys. Chem. 1986,90,4905. (h) Hengstun, A. J.; Pranger, J.; Hengstum-Nijhuis, S. M.; Ommen, J. G.; Gellings, P. J. J . Catal. 1986, 101, 323. (2) (a) Miyamoto, A.; Kobayashi, K.; Inomata, M.; Murakami, Y. J . Phys. Chem. 1982,86, 2945. Miyamoto, A,: Yamazaki, Y.; Hattori, T.: Inomata, M.; Murakami, Y. J. Catal. 1982, 7 4 , 144. (b) Bosch, H.; Janssen, F. J. J. G . ; Kerkhof, F. M. G.; Oldenziel, J.; Ommen, J. G.; Ross, J. R. H. Appl. Catal. 1986, 25, 239. (3) Inomata, M.; Miyamoto, A.; Murakami, Y. J. Catal. 1980, 62, 140. Miyamoto, A,; Inomata, M.; Hattori, A.; Ui, T.; Murakami, Y. J . Mol. Catal. 1982, 16, 315. (4) Satsuma, A,; Hattori, A,; Miyamoto, A.; Hattori, T.; Murakami, Y.
(8) Jpn. Kokai 46 463, 1976; 124 88 48230,48231,97547, 1981. (9) Rossi, S.; Iguchi, E.; Schiavello, M.; Tilley, R. J. D. Z . Phys. Chem. (Munich) 1976, 103, 193. Ai, M. J . Calal. 1977, 49, 313; Jpn. Kokai 15051, Patent 672 199, 1980. Haber, J.; Janas, J.; Schiavello, M.; 1978; U.S.S.R. Tilley, R. J. D. J. Catal. 1983, 82, 395. (IO) Scofield, J. H. J. Electron Spectrosc. Relat. Phenom. 1976, 8, 129.
0 1988 American Chemical Societv
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988 6053
Surface Active Sites of V205-W03 Catalysts
TABLE I: Relative Amount of V4+, BET Surface Area (SBm), Number of Redox Sites" ( L ) , Surface Concentration of Redox Sites (L/SBm), Number of Acid Sites ( A ), Surface Concentration of Acid Sites ( A /SBm),and Sum of the Surface Concentration of Redox Sites and Acid Sites ( ( L A )/SBm) of Vz05-W03 Catalysts
+
amount of V4+ (arbitrary) 0.0 1.o 0.92
catalyst VZOS W/(V W/(V W/(V W/(V W/(V
wo3
+ W ) = 0.1 + W ) = 0.3 + W ) = 0.5 + W ) = 0.7 + W ) = 0.9
m2/g 5.8 5.1 2.8 3.1 2.6 3.5 15.6
SBET,
0.57 0.29
0.01 0.0
L, rmol/g 22 32 34 43 42 48 0
(L + ~ / S B E T , L/SBET,rmol/m2 3.8 6.3 12.1 13.9 16.2 13.7 0
A, rmol/g 43 35 20 26 20 30
AISBET,rmol/m2 7.4
6.9 7.1 8.4
7.7 8.6 6.4
100
rmol/m2 11.2 13.2 19.2 22.3 23.9 22.3 6.4
"See text.
I
28
1200
(degree1
I
20
30
40
50
60
I
I
I
1
I
1000 I
I
800 I
I
600 I
I
(''-')
I
0
-0.7
Figure 1. XRD diagrams of V 2 0 s - W 0 3catalysts with various compositions: (0)v20s;(A) wo3.
In situ infrared spectra of adsorbed N H 3 were measured with Jasco FT/IR-3 spectrometers equipped with an in situ cell connected to a conventional high-vacuum system. TPD spectra of adsorbed NH3 were measured with a conventional TPD apparatus similar to that described by Amenomiya et al." The number of redox sites, Le., the surface V=O species on the catalyst, was determined by using the N A R P (NO-NH3 rectangular pulse) technique.I2 All the experiments were carried out under the same conditions as described previ~usly.~
Results X-ray Diffraction. Figure 1 shows XRD diagrams of V2O5WO, catalysts with various compositions. At W/(V W) = 0.1, the catalyst was composed of V205 phase and a small amount of WO,. The intensity of the diffraction lines of WO, increased with W 0 3 content, and at W/(V + W) = 0.5, diffraction lines of W 0 3 became more intense than those of V205. At W/(W + V) = 0.7, diffraction lines of V205 disappeared, and only those of W 0 3 could be observed. For all of the catalysts, no other diffraction lines except VzOs and W 0 3 phases were observed in the diagrams, indicating that little or no intermediate compounds were formed in the V,05-W03 catalysts. No broad line indicative of amorphous phase could be observed at a noticeable intensity. Since the width of the diffraction lines was rather broad, the formation of solid solution
+
I
1200
I
I
1000
I
I
I
800
I
I
630
Wavenumber ( c m - l )
Figure 2. IR spectra of V20s-W03 catalysts with various compositions.
could not be confirmed by XRD diagrams. IR Spectra. Figure 2 shows IR spectra of V205-W03 catalysts. Vz05 exhibited a band at 1022 cm-' which is assignable to the V=O stretching ~ i b r a t i o n . ' ~ When ~ ' ~ a small amount of W 0 3 was added to V20s, the spectrum did not undergo essential changes. The band position did not change, although the intensity of the V = O absorption gradually decreased with increasing W 0 3 content. Below 1000 cm-', V205 exhibited the band assignable to the V-0-V stretching at 900-750 ~ m - ' . ' ~ ,The ' ~ addition of WO, to V 2 0 5broadened the band because of the combination with the W-0-W stretching band at 900-600 cm-'.15 ESR Spectra. As shown in Figure 3, ESR signals were observed at g = 1.96 on all catalysts except V205 and W03. These ESR signals were assigned to the V4+ ion, indicating that V ions are partially reduced in the v205-Wo3 catalysts. At W/(V W) = 0.9, hyperfine structure was observed, suggesting that the vanadium ions are well-isolated in the catalyst. ESR parameters of the perpendicular principal components were as follows: g, = 1.991 G and A , = 43 G. The value of g, agrees
+
(1 1) CvetanoviE, R. J.; Amenomiya, Y. Adv. Carol. 1967, 17, 103; Card. Rev. 1972, 6, 21. (12) Miyamoto, A,; Yamazaki, Y.; Inomata, M.; Murakami, Y. J . Phys. Chem. 1981, 85, 2366. Inomata, M.; Miyamoto, A.; Murakami, Y. Ibid.
(13) (a) Frederickson, L. D.; Hansen, D. M. Anal. Chem. 1963, 35,818. (b) Miyamoto, A,; Yamazaki, Y.; Murakami, Y. Nippon Kagaku Kaishi 1977, 619. Inomata, M.; Miyamoto, A,; Murakami, Y. J . Catal. 1980,62, 140; Chem. Lett. 1978, 799. (14) Tarama, K.; Teranishi, S.; Yoshida, S.; Tarnura, N. Proc. Inf. Congr. Caral., 3rd 1965,282, Tarama, K.; Yoshida, S.; Ishida, S.; Kakioka, H. Bull. Chem. SOC.Jpn. 1968,41, 2840. (1 5) Rocchiccioli-Deltcheff,C.; Thouvenot, R.; Franck, R. Spectrochim. Acta, Part A 1976, 32, 587. Highfield, J. G.; Moffat, J. B. J. Catal. 1984,
1981, 85, 2372.
88, 171.
6054
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988
Satsuma et al.
0.2
0
0.4
0.6
W/V+W
In bulk
0.8
1.0
Figure 5. Relationship between surface composition and bulk composition.
vi
Range =
Range =
3x10'
Range
300
=
wo+
w/v+w
U
H Figure 3. ESR spectra of V205-W03 catalysts with various compositions.
100
7
I-+
wo;
wv0+
=0.3
=0,9
Mechan I ca I
0
0.2
0.4
0.6
0.8
1.0
In bulk Figure 4. Binding energy of V ( ~ P ~ , W(4d3/,), ~), and W(4d5/,) based upon C(1s) = 285.0 eV. W/W+V
40
60
'/
200
220
240
260
280
m / e (a.rn,u.)
fairly with those of V4' observed in V205-P20514316 and supported V2O5,I7although the value of A , lies between those catalysts and crystalline V2O5.l' Since the S / N ratio of the ESR signal was low, the parallel principal components of gll and All could not be determined. Table I shows the relative amount of V4' ions determined by the double integrations. The amount of V4+ions was the largest at W/(V W) = 0.1, and it decreased gradually with increasing W 0 3 content. XPS Spectra. Figure 4 shows the binding energies of XPS peaks of V(2p3/2), W(4d3/2), and W(4d5/2) corrected by the C(1s) peak as 285.0 eV. As shown in Figure 4, the binding energy of V(2p3/,) was relatively low between W/(W V) = 0.1 and 0.5. On the other hand, the chemical shifts of W(4d3 2) and W(4ds/z) peaks could be neglected in all the catalysts. The trend of these
+
+
(16) (a) Nakamura, M.; Kawai, K.; Fujiwara, Y . J . Coral. 1974, 34, 345. (b) Martini, G.; Trifirb, F.; Vaccarl, A. J . Phys. Chem. 1982,86, 1573. ( c ) Ballutaud, D.; Bordes, E.; Courtine, P. Mater. Res. Bull. 1982, 27, 519. (17) (a) Inomata, M.; Mori, K.; Miyamoto, A.; Ui, T.; Murakami, Y . J . Phys. Chem. 1983, 87, 754. Inomata, M.; Mori, K.; Miyamoto, A,; Murakami, Y . Ibid. 1983, 87, 761. (b) Sharma, V. K.; Wokaun, A,; Baiker, A. Ibid. 1986, 90, 2715. (c) Cordischi, D.; Vielhaber, B.; Knozinger, H. Appl. Caral. 1987, 30, 265. (18) Gillis, E.; Boesman, E. Phys. Status Solidi 1966, 14, 337.
Figure 6. SIMS spectra of V205-W03 catalysts with various compositions (resolution = 6.5).
results was similar to that of ESR. Figure 5 shows the surface composition determined from the relative intensity of V(2p3/2) and W(4d5/2)with various compositions by using reported photoionization cross section factors.I0 Surface compositions and bulk compositions were fairly equal to each other. SZMS Spectra. Figure 6 shows SIMS spectra of V205-W03 catalysts. All of the spectra indicate secondary ions of V', VO', W', WO', and W02+. The peak ascribed to Fe' is due to the contaminants which were admixed when the sample was charged in the SIMS cell. At W/(V W) = 0.3,0.5,and 0.7, composite ions including both vanadium and tungsten, Le., WVO' and WV02+, were also observed. These composite ions suggest that V-0-W bonds formed on the surface of the catalysts. It should be noted that such composite ions were not observed in the case of the mechanical mixture of V205 and W 0 3 , though VO' and WO' ions were clearly observed, as shown in Figure 6. Thus, these results indicate the mixing of vanadium ions and tungsten ions on an atomic scale on the catalyst surface. Figure 7 shows SIMS spectra of (a) W/(V + W) = 0.5, (b) W/(V + W) = 0.5 after the redox treatment described below,
+
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988 6055
Surface Active Sites of V2Os-WO, Catalysts V@+
1800
I
Vt
I
wo+
,.I,
v205
I
1600 I
I
(cm-l)
11100 1200 I I I I I
---+-I
I1222
I
u -
40
50
60
70
180
190
Wavenumber ( c r n - l )
200
210
220
Figure 8. Infrared spectra of NH3 adsorbed on V20,-W03 catalysts at
230
room temperature.
m / e (a.m,u.)
+
+
Figure 7. SIMS spectra of (a) W/(V W) = 0.5, (b) W/(V W) = 0.5 after the redox treatment, and (c) the mechanical mixture of V,05
and W03 after the redox treatment (resolution = 8.6). and (c) the mechanical mixture of V 2 0 5and W 0 3 containing 50 atom % tungsten after the redox treatment. The redox treatment was done as follows: The catalyst, which had been preoxidized in flowing I6O2at 773 K for 20 min, was reduced by injecting several pulses of the NO-NH, mixture at 588 K. Then the catalyst was reoxidized with 10.7 kPa of 1802 at 573 K for 18 h. As for VO+ ion, the fresh catalyst (a) gave the VO+ ion peak at m / e 67 and the VOH' ion peak at m / e 67, while the redoxtreated catalysts (b) and (c) gave another VO+ ion peak at m / e 69. The ion peak at m / e 69 can be assignable to the VI80+ ion. This result indicates that the V=O species was replaced with gaseous l8Ozin the redox treatment. According to the reaction mechanism of the reduction of N O with N H 3 in the presence of O2 proposed by us,, the surface V=O species are reduced to V-OH as follows: V=O
+ N O + NH3
-+
N2
+ H2O + V-OH
(1)
The V-OH species thus formed is reoxidized by gaseous 02: 2V-OH
+ !I202
-+
I I I I I I I I 1800 1600 1400 1200
+ H2O
2V=O
(2)
Thus, the surface oxygen species having a redox function must be substituted for gaseous oxygens after the reaction. Therefore, the SIMS spectra in Figure 7 show that the surface V=O species act as the redox sites. As for the WO+ ion peak, the mechanical mixture after the redox treatment gave essentially the same spectrum as that of the untreated V205-W03 catalyst, as shown in Figure 7a,c. Thus, the oxygen species bonded to surface W ions were not substituted with the gaseous oxygen during the redox treatment, which agrees with the fact that W 0 3 is inactive by itself for the reduction of N O with N H 3 at the temperature of the redox treatment.Ig On the other hand, the V2O5-WO3catalyst after the redox treatment ~
gave a shoulder peak at m / e 204, as shown in Figure 7b. This peak was assigned to 186W'80+,indicating that the surface. oxygens bonded to surface W ions in the catalyst were substituted with gaseous oxygen during the redox treatment, probably in the same way as the V=O species. In other words, the surface oxygens bonded to W ions in the V2O5-WO3catalyst act as the redox site under the reaction condition. The fact that the mechanical mixture did not give the la6WI80+ion peak after the same treatment indicates that the oxygen species bonded to the surface W ions are activated under the effect of V ions mixed in atomic order. IR Spectra of Adsorbed NH3. Figure 8 shows infrared spectra of NH, adsorbed on VZOS, V205-W03catalysts of W/(V W) = 0.7 and 0.9, and W 0 3 . Since the IR disks were very dark, IR spectra could not be measured for W/(V + W) = 0.1, 0.3, and 0.5. An adsorption band assignable to the bending vibration of NH4+ ion was observed at around 1420 cm-1,20indicating the presence of the B r ~ n s t e dacid sites on the surface of all the V20s-W0, catalysts. Rands assignable to coordinately held N H 3 ion were also observed around 1610 and 1260 cm-' at W/(V W) = 0.7 and 0.9 and WO,, suggesting the presence of Lewis acid sites on the surface of the catalysts. The relative concentration of Lewis acid sites and Bransted acid sites was calculated from the intensity of the bands at around 1610 and 1420 cm-' with the ratio of absorption coefficient of ~ ( 1 6 1 0cm-')/t(1420 cm-l) = 1/7.21 V205contained only 14% Lewis acid sites, and Lewis acid sites increased with W 0 3 content: the relative concentrations of Lewis acid sites were 26, 45, and 54% at W/(V W) = 0.7,0.9, and WO,, respectively. Temperature-Programmed Desorption of NH,. Figure 9 shows TPD spectra of NH,. For the unpromoted V205,the desorption maximum was observed at 350 K, and the desorption was completed below 520 K. The TPD profile on W 0 3 was considerably different from that on V2O5. Thus, the desorption peak was very broad and flat, and the desorption of N H 3 continued to 670 K. TPD profiles of desorbed NH3 from the V205-W03 catalysts of various compositions were essentially the same as that from V2Os.
+
+
+
~~~~~
(19) Murakami, Y.; Hayashi, K.; Yasuda, K.; Ito, T.; Minami, T.; Miyamoto, A. Nippon Kagaku Kaishi 1977, 173.
(20) Krozinger, H. Adu. Catal. 1976, 25, 184. (21) Basila, M. R.; Kantner, T. R. J . Phys. Chern. 1967, 7 1 , 467.
6056
Satsuma et al.
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988
A
5b3
5EO
58@
TemperotLre
6GC 1
622
K
Figure 11. Amount of the initial sharp N2measured at various temperatures: ( 0 )V,O,; (0)W/(V W) = 0.1; (V) W/(V W) = 0.3; (0) w / ( v w) = 0.5; (A) w / ( v w ) = 0.7; (0)w / ( v w) = 0.9; (V) WO3.
+
3@@
400
500
600
700
Temoeroture/ K
Figure 9. TPD spectra of desorbed NH3 from V205-W03 catalysts with various compositions.
+
+
+
+
catalysts with a large P/V ratio: the tailing N2 was so large under the conventional condition that the reliable results could be obtained only at lower t e m p e r a t ~ r e . ~ Since very similar profiles of N, were obtained in the same temperature range, the initial N2 could be separated from the tailing N2 in the same way as that in the case of the unpromoted v205.4~12 The initial sharp N2 can be separated from the tailing N2 by the dotted line indicated in Figure 10, and the number of the redox sites is given by the amount of the initial sharp N2. Figure 11 shows the reliability of the amount of the initial N2thus obtained. The figure shows that the amount of the initial N2 is constant and is independent of the reaction temperature, indicating that the amount of determined initial sharp N2 under the present condition did not depend on the rate of the NO-NH3 reaction. Therefore, the NARP technique can be applied for the V205-W03 catalysts to determine the number of the surface redox sites; that is, the amount of initial sharp N2 represents the number of the surface species. The surface concentration of the redox sites, Le., the surface V=O species (L/SBET),was determined from the number of the redox sites ( L ) and BET surface area (SBET),and they are summarized in Table I. The surface concentration of the redox sites increased with increasing WO, content below W/(V + W) = 0.5 and was almost constant from W/(V W) = 0.5 to 0.9.
+
Time / sec
Figure 10. Concentration profiles of N2produced by the reaction of the rectangular pulse of NO and NH3with W/(V + W) = 0.5 at various
temperatures. Table I shows the number of acid sites ( A ) and the surface concentration of the acid sites (AISBET), determined from the number of the desorbed N H 3 in TPD experiments. The surface concentration of the acid sites did not essentially depend on the WO, content. NARP (NO-NH, Rectangular Pulse) Technique. When a rectangular pulse of a NO-NH, mixture is introduced onto the V20scatalyst, N, with a concentration profile consisting of initial N2 and tailing N2 is obtained in the temperature range 550-610 K. And the number of the redox sites, Le., the surface V=O species, can be determined from the amount of the initial N2.4*12 In contrast to this, WO, did not give any Nz peak, which agrees well with the fact that WO, is active only at higher temperature by ca. 150 K for the reduction of NO with ",.I9 Figure 10 shows a typical example of concentration profiles of N2 produced in the NARP technique, Le., N2 produced by the reaction of the rectangular pulse of the N O and N H 3 mixture with the catalyst of W/(V + W) = 0.5. The profiles of N 2 shown in Figure 10, which are very similar to those on the V20s catalysts, were obtained in a temperature range completely identical with that for V205. This is in marked contrast to the results on the v,O5-P205
Discussion I . Structure of V205-W03Catalysts. The results of the bulk characterizations indicate that the v205-Wo3 catalysts are mainly composed of the crystallites of V2O5 and WO,, and there are little or no intermediate compounds in the bulk. This is because in the XRD patterns (Figure 1) only the lines assignable to V2OS and WO, were detected, and because in the IR spectra (Figure 2), the addition of WO, decreases the absorption band of V=O stretching and increases that of W-0-W stretching. Therefore, V and W ions are not well-mixed in the bulk; that is, the V205-WO, catalysts are similar to a mechanical mixture of V205 and W 0 3 at first glance. Only ESR signals assignable to V4+ (Figure 3) show the effect of WO, in the bulk. The formation of V4+ions suggests a possibility of the formation of solid solutions, because the dissolution of 6-valent W ions in the matrix of 5-valent vanadium oxide will cause the formation of 4-valent V ions due to the valence control. However, the results of the surface characterizations indicate that V and W ions are well-mixed in the surface. The strong acid sites are not observed in the V205-W03 catalysts even at the high W 0 3 content, although the WO, itself contains the strong acid sites. This result indicates that the crystal plane of WO, is not exposed on the surface of the V205-W03 catalysts. One might suspect that the surface of the VZO5-WO3catalysts are covered with a thin layer of V205, since the strength and the surface concentration of the acid sites of the V205-W03 catalysts are
The Journal of Physical Chemistry, Vol. 92, No. 21, 1988 6057
Surface Active Sites of V205-W03 Catalysts -.i
-E" E
Surfoce H content
(W/V+WI
Figure 12. Relation of the surface concentration of the active sites with surface W content: (0)redox sites (see text); (A) acid sites; (0)sum of the surface concentrationsof the redox sites and the acid sites; (---) calculated surface concentration of V ions on all planes; calculated surface concentration of V and W ions. (-e)
similar to those of V2O5. However, XPS spectra (Figure 5) show that the surface W contents are fairly equal to that in the bulk, indicating that the surface layers of the V205-W03 catalysts are composed of V and W ions. Thus, all of these results certainly indicate that V and W ions are well-mixed in the surface. Furthermore, the secondary ions of V', VO+, W', and WO' observed in SIMS spectra (Figure 6 ) indicate that the top layer of the v ~ o ~ - wcatalysts o ~ must contain both V and W ions. It was also confirmed that the V-0-W bonds are formed in the surface, since the secondary ions of WVO+ and WV02+ were observed in SIMS spectra. In conclusion, the surface structure of the V205-W03 catalysts is entirely different from that in the bulk; that is, V and W ions are well-mixed in the surface of the catalysts, though VZO5 and W 0 3 are not well-mixed in the bulk. 2. Surface Redox Sites on V205-W 0 3Catalysts. Figure 12 shows the surface concentration of active sites as a function of surface W content determined by XPS. The surface concentration of the redox sites, Le., the surface V=O species, remarkably increased by the addition of WO,, indicating a strong synergistic effect of W 0 3 on the surface redox sites. 2.1. Surface V=O Species. The concentration of the surface V=O species on V205-W03 catalysts was higher than that on the unpromoted V2O5 catalyst, as shown in Figure 12. A similar result was obtained in the case of the V205-P205catalysts, and the increase in the surface concentration of the V = O species was attributed to the formation of lower valent vanadium ions in the bulk.4 AnderssonZZshowed that, in the case of V205,the surface V=O species is formed only on the (010) plane, but the surface V = O species are formed on various crystalline planes on slightly oxidized V6013and V204, resulting in the increase of the surface V = O species. And, actually, it has been experimentally confirmed that the oxidized V,Ol3 and V 2 0 3 have a higher concentration of surface V=O species.23 ESR spectra (Figure 3) indicated that the addition of W 0 3 to V205 forms V4+ ions, and XPS spectra (Figure 4) showed that low-valent vanadium ions are formed in surface layers. The formation of V4+ ions may be due to the dissolved W ions forming solid solution. Lower valent vanadium ions in the bulk may result in the formation of the surface V = O species on various crystalline planes in the same way as that in the lower valent vanadium oxides, leading to the increase in the concentration of the surface V=O species. 2.2. Surface W=O Species. Broken lines in Figure 12 represent the calculated surface concentration of V ions. The upper line was calculated from the closest packing of V 0 6 octahedra, and the lower line was calculated from the surface concentration (22) Andersson, A. J . Solid State Chem. 1982, 42, 263. (23) Miyamoto, A,; Hattori, A,; Murakami, Y . J . Solid State Chem. 1983, 47, 373.
of V ions on the (010) plane of V205. Since the concentration of the redox sites shown by the open circles was lower than that indicated by the broken lines at surface W content less than 0.5, the increase in the surface redox sites can be explained by the increase of V4+ as mentioned above. However, at surface W content above 0.5, the surface concentration of the redox sites exceeded the broken lines representing the concentration of surface V ions. One of the possible explanations of this discrepancy is the difference between the surface concentration of V ions in the top layer and that in the surface layers measured by XPS. But the surface concentration of the redox sites appears too large as an experimental error of the XPS analysis at surface W content of 0.9. Another possibility is that surface W ions also have active oxygens for the formation of N z in the NARP experiment. W 0 3 itself was inactive in the N A R P experiment; that is, any N 2 was not formed when the pulse of NO-NH3 mixture was introduced onto the oxidized WO, catalyst at the temperature region shown in Figure 11. However, it is quite probable that the surface W ions become active under the influence of neighboring V ions. This possibility was examined by using SIMS spectra (Figure 7). The preoxidized catalyst was slightly reduced by injecting several pulses of the NO-NH3 mixture and then oxidized with I8O2. The SIMS spectrum of V205-W03 catalyst treated as above gave W i 8 0 + ion as well as V i 8 0 +ion, while a mechanical mixture of V2O5 and W 0 3 did not give any Wl80+ ion. These results suggest that the latter is the case; that is, oxygen bonded to the surface W ion of the v205-Wo3 catalysts is active for the reduction of N O with "3.
Since the V205-W03 catalysts with high surface W content gave essentially the same N, profiles as that on Vz05 at the same temperature region, the surface oxygen species on W ions should be as active as unpromoted V 2 0 5for NO-NH, reaction. This may lead to the assumption that the oxygen species on the surface W ions of V205-W03 catalysts must be similar to the surface V=O species of V205, that is, the surface W=O species. The presence of W = O species has been confirmed for supported W 0 3 catalyst^,^^^^^ compounds such as Al,(WO,), or NazW04,25 and heteropolyanions of tungsten.26 Thus, the surface W=O species may be formed under the effect of neighboring surface V ions. Composite ions of WVO+ and WVO,+ observed in SIMS spectra indicate that V and W ions are well-mixed on an atomic scale on the catalyst surface. Although the composite ions of WVO' and WV02+were not observed in W/(V + W) = 0.9, the hyperfine structure in the ESR spectrum of W/(V W) = 0.9 indicates that vanadium ions are isolated in the W 0 3 lattice; that is, V and W ions are well-mixed. The mixing on an atomic scale enables strong interaction of W ions with neighboring V ions on the surface of the V205-W03 catalysts, which may lead to the formation of the active surface W=O species. Thus, the redox sites include not only the surface V = O species but also the surface W=O species on the v205-Wo3 catalysts, especially at high W 0 3 content. 3. Surface Acid Sites on V205-W03Catalysts. The acidic property of the catalysts were characterized by the TPD of NH, and the IR spectra of adsorbed NH,. The IR spectra (Figure 8) indicate that Br~nstedand Lewis acid sites exist on the surface of all the V ~ O S - W Ocatalysts. ~ The spectra also show gradual change of acid type; that is, the addition of W 0 3 increases the relative concentration of Lewis acid sites on the surface. The ratio of Lewis acid sites to B r ~ n s t e dacid sites was smaller on V205 than that of W 0 3 , and it increased gradually with the content of W 0 3 .
+
(24) Chan, S. S.; Wachs, I. E.; Murrell, L. L.; Wang, L.; Keith Hall, W. J . Phys. Chem. 1984, 88, 5831. (25) (a) Salvati, L., Jr.; Makovsky, L. E.; Stencel, J. M.; Brown, F. R.; Hercules, D. M. J . Phys. Chem. 1981, 85, 3700. (b) Chan, S . S.; Wachs, I. E.; Murrell, L. L. J . Catal. 1984, 90, 150. (26) (a) Keggin, J. F. Proc. R. Soc. London, A 1934,144,75. (b) Bradley, A. J.; Illingworth, J. M. Ibid. 1936, 157, 113. (c) Brown, G. M.; Noe-Spirlet, M. R.; Busing, W. R.; Davy, H. A. Acta Crystallogr., Sect. B 1977, 33, 1038. (d) Smith, D. P.; So, H.; Bender, J.; Pope, M. T. Inorg. Chem. 1973.12.685. Altenau, J. J.; Pope, M. T.; Prados, R. A,; So, H. Ibid. 1975, 14, 417.
J . Phys. Chem. 1988, 92, 6058-6065 In spite of the gradual change of the ratio of Lewis acid sites to Br~nstedacid sites, the surface concentration of the acid sites on the V205-W03 catalysts was independent of the W 0 3 content, as shown in Table I. It is quite probable, because the surface concentration of the acid sites on V205was not so much different from that on WO,. This is in marked contrast to the trend in the strength of the acid sites shown in Figure 9. The addition of WO, does not give any effect on the strength of the acid sites of the V205-W03 catalysts, though unpromoted VzO5 and WO, were significantly different in the strength of the acid sites from each other. Especially, it is astonishing that even at high WO, content, the profile of desorbed NH3 was similar to that of Vz05 but not to that of WO,. This may be rationalized if the strength of the acid sites on the surface W ions is weakened by the interaction with neighboring V ions, when the surface W ions are mixed with V ions on an atomic scale. 4. Sum of the Surface Concentration of Redox Sites and Acid Sites. The sum of the surface concentration of the redox sites and the acid sites increased with surface W content, and it remained constant in the range of surface W content from 0.5 to 0.9, as shown in Figure 12. The dotted line in the figure represents the surface concentration of V and W ions calculated on the assumption that the surface was covered by V 0 6 and W 0 6 octahedra with the surface concentration of the closest packing of V 0 6 octahedral. Since the sum of the surface concentrations of both sites is close to the dotted line at surface W content from 0.5 to 0.9, all the surface V and W ions act as the active sites: either of the redox sites (the surface V = O species and the surface W=O species) or of the acid sites. A similar result was observed
in the case of the Vz05-Pz05catalysts shown in a previous paper.4
Conclusion The vanadium and tungsten ions are well-mixed on an atomic scale on the surface of V205-WO, catalysts, although V2O5 and WO, are not well-mixed in the bulk. The atomic scale mixing leads to the changes in the concentration and the type of active sites of V205-W03 catalysts, which are summarized as follows: 1. The surface concentration of the redox sites gradually increased with surface W content. The most important reason for the increase in the redox sites is the activation of the surface W 4 species as the redox sites by the synergistic effect of neighboring V ions. Another reason is the formation of the V=O species on the other planes under the effect of V4+,although the surface redox sites are only the V=O species on the (010) plane in the case of v2°5.
2. The strengths of the acid sites on V2O5 and W 0 3 catalysts were different from one another, but the surface concentration and strength of the acid sites of V2O5 were not significantly modified by the addition of WO,. 3. Above the surface W content of ca. 0.5, all the surface V and W ions act as the active sites: either of the surface redox sites or of the acid sites.
Acknowledgment. This work was partially supported by a Grant-in-Aid from the Ministry of Education, Science, and Culture, Japan (No. 57470055). Registry No. V 2 0 5 , 1314-62-1; WO,, 1314-35-8; NH,, 7664-41-7; NO, 10102-43-9.
Dark Electrochemistry and Photoelectrochemistry of Molecularly Doped Ion-Exchange Polymer Blends Andrew M. Crouch, Ishmael Ordoiiez, Cooper H. Langford,* and Marcus F. Lawrence* Chemistry Department, Concordia University, 1455 West, de Maisonneuve Boulevard, Montreal, Quebec, Canada H3G l M 8 (Received: January 13, 1988; In Final Form: March 12, 1988)
Ion-exchange polymer blends have been shown to produce modified electrode surfaces with high affinities for ionic reactants. The main feature of these blends is their spontaneous tendency to segregate into hydrophilic and hydrophobic domains. It is now believed that, when appropriate dye molecules are incorporated into such films and then illuminated, these systems operate under both ionic and “dry” electronic conduction mechanisms. The dark electrochemical measurements performed on an ion-exchangepolymer containing ZnTPPS4-, ZnTPP, ZnPc(OPh),, or CuPcTS”, in contact with a Fe(CN)63-/‘+redox solution, show that the dye molecules within the film are immobile and that the high ion-exchange capability is maintained. The photoelectrochemical results obtained with the dye-loaded films indicate that electrons may be transferred from the photoexcited dyes to the polymer matrix and transported to the Sn02 substrate electrode. The oxidized dye molecules are reduced by accepting electrons from the Fe(CN):- species. The electronic conduction following charge separation is assumed to be intimately related to the ion-exchange polymer’s tendency to segregate into hydrophilic and hydrophobic domains and also to the excited-state energetics of the dye. A model which invokes the existence of large distributions of molecular ion states is proposed to explain the conduction of electrons through the hydrophobic domains of the polymer film and a detailed energy level diagram is presented to summarize the overall situation.
Introduction
The design of synthetic organic systems capable of efficiently performing the spatial separation of photogenerated charges (electron-hole Dairs) has lone been and still remains an activelv pursued subjeit. A simple iersion of this process is shown h Scheme I, which illustrates that by irradiating a dye molecule in the presence of appropriate donors and acceptors, one may expect vectorial charge migration. A wide variety of metalloporphyrins and metal phthalocyanines have been used to produce dye-sensitized systems for the conversion and storage of solar energy.’” (1) Minami, N.; Watanabe, T.; Fijishima, A,; Honda, K. Ber. Bunsen-Ges. Phys. Chem. 1979,83,476. ( 2 ) Jaeger, C. D.; Fan, F. F.; Bard, A. J. J . Am. Chem. SOC.1980, 102, 2592.
0022-3654/88/2092-6058!$01.50/0
Within this context, polymers have also attracted much attention because of their potential use as a solid support for the chemical mdification of electrode (3) Darwent, J. R.; Douglas, P.; Harriman, A,; Porter, G . ;Richoux, M.-C. Coord. Chem. Rev. 1982, 44, 83. (4) Kirk, A. D.; Langford, C. H.; Saint-Joly, C.; Lesage, R.; Sharma, D. K. J . Chem. SOC.,Chem. Commun. 1984, 961. (5) Langford, C. H.; Saint-Joly, C.; Pelletier, E.; Arbour, C. Inorg. Chim.
,-.
Acta 1984. 87. 13 - 1.
-(6)Bettelheim, A,; White, B. A,; Raybuck, S. A,; Murray, R. W. Inorg. Chem. 1987, 26, 1009. (7) Ikeda, T.; Schmehl, R.; Denisevich, P.; Willman, K.; Murray, R. W. J . Am. Chem. SOC.1982, 104, 2683. (8) Montgomery, D. D.; Anson, F. C. J . Am. Chem. SOC.1 9 5 , 107, 3431. ( 9 ) Sumi, K.; Anson, F. C . J . Phys. Chem. 1986, 90, 3845.
0 1988 American Chemical Society
