Vanadia-titania systems: morphological and structural properties

MaterialsResearchBulletin,Vol.31,No.5, pp.513-520,1996 CopyrightO 1996ElsevierSeienoeLtd PrintedintheUSA.Allrightsreserved 0025-5408/96$15.00+ .00

Pergamon

PII S0025-5408(96)00029-3

VANADIA-TITANIA SYSTEMS: M O R P H O L O G I C A L AND STRUCTURAL PROPERTIES

L. J. Alemany I, M. A. Bafiares 2, M. A. Larrubia 1, M. C. Jim6nez I, F. Delgado I and J. M. Blasco ! (1) Departamento de Ingenieda Quimica, Universidad de Mfilaga, Campus de Teatinos, 29071 M~aga, Spain. (2) Instituto de Cat~disis y Petroleoquimiea, CSIC, Campus UAM, Cantoblaneo, 28049 Madrid, Spain. (Received June 5, 1995; Accepted December 28, 1995) (Refereed)

ABSTRACT The changes of structure of TiO2 in line with vanadium loading and thermal treatment have been studied. A correlation is proposed between vanadium surface coverage, crystal size and the Raman intensity. Modification of anatase cell parameters by vanadium substitution is reported. KEYWORDS: A. oxides, C. Raman spectroscopy, C. X-ray diffraction, D. crystal structure

INTRODUCTION Many investigations have been published on the characteristics of vanadia-titania catalysts in the current literature (1-7). The structure and reactivity have been extensively studied, but the influence of vanadia on the structural and morphological evolution has been less studied (8, 9). In this paper we present an investigation on the influence of temperature and the vanadium loading on the morphological and structural properties of the vanadia-titania catalytic systems, A correlation is attempted between the microscopic properties of material and the macroscopic morphological and structural ones. EXPERIMENTAL Pure TiO 2 and V205-TiO 2 samples have been prepared and characterized. Pure TiO2 has been obtained by neutralization at a pH constant (pH = 6) of an acid solution of high purity grade

513

514

L.J. ALEMANY et aL

Vol. 31, No. 5

of TiC14. The precipitate after filtering and washing with water to remove chloride ions, has been dried overnight at 383 K and calcined at different temperatures up to 1073 K. V2Os-TiO2 samples with different loading have been prepared by dry impregnation technique. A hot aqueous solution of ammonium metavanadate and oxalic acid was used as a complexing agent on the titania dried at 383 K. Afterwards, the samples were dried and calcined at different temperatures. Surface area, and micro- and mesopore size distribution have been determined by nitrogen adsorption with BET method using a Carlo Erba Sorptomatic 1800 series instrument. Powder X-ray diffraction analysis have been performed with a Siemens D-501 instrument. Powdered silicon served as an internal standard. The lattice parameters were computed using the least squares technique to increase the accuracy of the minimized quantity being (20exp-2 0calc)2. The microstrcture of the samples has been studied in terms of crystallite dimensions. The influences of particle size and of the lattice disorder on the peak shape have been separated using a single line method, as proposed in the literature (10). The anatase and rutile relative contents have been calculated according to the equation reported in literature (11): %A= I(101)A/(I(101)A +1"4I(110)R) Raman Spectra were recorded on a Brucker FT-Raman instrument with an exciting line of 1064 nm. Power on the sample was 40 mW.

RESULTS AND DISCUSSION Table 1 shows mean crystallite dimensions (Dc), specific surface area (Sa), pore volume (Vp), and mean pore radius (rp) obtained from experimental pore size distribution and calculated according to the equation rp = 2Vp/Sa ofTiO 2 and V205-TiO 2 (1.5% and 7.5% w/w) samples, calcined at different temperatures. Pure TiO 2 samples show only anatase phase at low temperatures. TiO 2 rutile is observed once the calcination temperature is higher than that for the initial sintering of TiO 2 anatase (973 K). Line broadening analysis reveals that the crystal size of titania anatase phase increases with increasing vanadia content and calcination temperature. Reflections corresponding to the anatase phase are observed below monolayer surface coverage. Crystalline vanadia segregates and titania rutile phase becomes evident for vanadium oxide loadings above monolayer coverage. The crystallite dimensions increase with the temperature for all samples. Meanwhile V2Os-TiO2 crystallite dimensions are always higher than those corresponding to TiO 2. The specific surface area decreases, as the calcination temperature and vanadium oxide loading increases. Pore volume decreases and the mean pore radius increases and the distribution is unimodal. The measured value of rp compares well with the calculated ones, indicating a consistency among pore size distribution, pore volume and surface area measurements. The decrease of specific surface area with temperature can be interpreted as being due to the growth of crystallites (12). Figure 1 shows average particle size determined by XRD plotted vs. average particle size determined out of the BET surface area. These were calculated assuming that the particles are spherical. There is a good correlation between both techniques.

Vol. 31, No. 5

VANADIUM TITANIUM OXIDE

515

TABLE 1 Characteristics o f V 205/TIO 2 Materials V205 (%) 0 0 0 0 0 1.5 1.5 1.5 1.5 1.5 7.5

Temp. (K) 383 623 873 973 1073 383 623 873 973 1073 383 623 873

7.5

7.5

Phase Am A A A + R tr 8A + 92R Am A A 96A + 4R 6A + 94R Am A 28A+72R V 2 0 5 Ir

De (A) 105 155 350 950 110 182 550 970

Sa (m2/~) 284 165 85 28 10 295 118 52 15 5

-

288

230 695

V la (cm~/g) 0.33 0.27 0.39 0.35 -

Rpexp

50 68

73 120

-

48 8

0.42 0.41

190

Am: Amorphous; A: Anatase; R: Rutile; tr: traces; nnA: nn% Anatase.

1000

•

~

~'~

• • •

400

-0 • 0

*

0

°

i

.

.

.

.

|

.

.

.

.

|

,

•

°

,

"

1000 2000 3000 Crystal size obtained from Specific Surface Area

4000 (A)

FIG. 1 Average particle size values determined by XRD measurements vs. those determined from BET surface area values.

Table 2 shows the morphological characteristics of the vanadia/titania samples calcined at 873 K. The theoretical surface coverage has been calculated assuming a monolayer capacity corresponding at 7.9 ~xnol V 20 5/m 2 in line with the literature report (13). Significant differences are observed in the crystallites dimensions o f anatase as vanadium content increases. A

516

L.J. ALEMANYet

Voi. 31, No. 5

al.

remarkable growth of crystal size is measured due to the rutile presence on V2Os/TiO2 sample with the highest vanadium loading. The average crystal size of the anatase phase increases with growing vanadium loading. The morphological characteristics of the samples clearly depend on the V20 s content. The specific surface area decreases from 85 m2/g (sample without vanadia) to 30 mZ/g (3.5% w/w vanadia/titania sample). A pronounced reduction of specific surface area is observed for the sample with the highest vanadia content where TiO2 rutile is dominant. The pore volume and the mean pore radius increase on increasing the vanadium content in line with the increasing of surface coverage. The presence of vanadium on the surface of TiO 2 decreases the specific surface area proportionally to its loading on the catalysts. The variation of Sa is in agreement with the corresponding change in the average crystal size, pore volume, and pore radius. TABLE 2 Morphological characterization of V 205/TIO 2 samples calcined at 873 K V205 Sa (%) O (m2/g) 0 0 85 0.3 0.03 71 1 0.11 62 1.5 0.20 52 2.5 0.36 48 3.5 0.69 30 7.5 >>1 8 O: Theoretical surface coverage;

Vp (cm3/g) 0.27 0.28 0.30 0.35 0.37 0.40 0.41

Dc (A) 155 158 165 182 190 235 695

IRam (638 cm "1) 0.614 0.614 0.512 0.502 0.283 0.064 0.006

IRam:(638 cm'l: Raman relative intensity measured at 638 cm'l). The anatase cell parameters a and c derived for the samples with different vanadium loading and calcined at 873 K are shown in Table 3. The axial ratio (c/a) is also included. Whereas the c parameter in the catalysts decreases with the vanadium loading, the a paramenter values are constant within experimental error. The observation of changes of the unit cell volume of titania (anatase) with vanadium loading indicates that vanadium is combining with the titania lattice, TABLE 3 Lattice Parameters of Anatase in V2Os/TiO2 calcined at 873 K V205 (%) 0 0.3 1.0 1.5 2.5 3.5

a (nm) 0.3783 0.3783 0.3784 0.3782 0.3783 0.3783

c (nm) 0.9516 0.9515 0.9513 0.9513 0.9509 0.9506

c/a 2.5155 2.5152 2.5140 2.5153 2.5136 2.5128

Volume (nm3) 0.13618(8) 0.13617(4) 0.13621(1) 0.13607(6) 0.13608(2) 0.13604(5)

Ag0 0.005 0.004 0.004 0.004 0.004 0.003

z~20 = (l/N) E (20exp-20ealc),where 20exp is the experimental diffraction angle, 20talc is the one calculated from lattice parameters, and N is the number of investigated diffraction lines.

Vol. 31, No. 5

VANADIUM TITANIUM OXIDE

517

producing a slight reduction of cell volume and leaving the diffractogram of the anatase phase unchannged. Since the ionic radii of V (IV) and Ti are very close, it is reasonable to assume that vanadia ions substitutionally replace the Ti ions. Probably, there is a limit to the m o u n t of vanadia that can dissolve into anatase structure, when this limit is reached no decrease in the unit cell value n~,,ght be observed. FT-Raman spectra of titania-supperted vanadia catalysts are presented in Figure 2. The major Raman bands observed at 143, 192, 396, 515, and 638 cm" 1 are characteristic for anatase phase (7,15). Figure 3 shows Raman spectrum of reference bulk vanadia. It shows Raman bands at 102, 144, 196, 284, 405, 480, 527, 701, and 993 cm "l, which agrees well with previous reports (13). The Raman bands corresponding to the crystalline V205 are visible for the samples of higher loading levels, which also show Raman modes for the rutile phase at 138, 231,443, 606, and 882 cm -1 (Fig. 3), as expected from the structural modifications. Segregation of crystalline vanadia and the onset of TiO2 rutile has been observed by several authors (8,9,17) Raman modes for crystalline vanadium pentoxide are not observed below the monolayer surface coverage. Relative intensities associated to the anatase-TiO 2 phase, decrease by increasing the vanadium content. This effect was observed by Bond at al. (18) and has been attributed to the Raman scattering process becoming less efficient as the color deepness with increasing amount of vanadium (19). However, the modification from anatase to ruffle is promoted by vanadium oxide loading (8,2022) the amount of rutile increases with vanadia loading, for the same calcination temperature. There is a good linear correlation (R =-0.992) between the Raman intensity and the theoretical surface coverage while the vanadium content is below a monolayer.

,I I i

a

1000 950

900 850 800 758 700

650

600

550 500

Wavenumber

450

,180 350

300

250

200 150

90

c m -~

FIG. 2 FT-Raman Spectra of representative vanadia/titania catalysts: (a) TiO2, (b) 0.3% (C) 3.5% V2Os/TiO2, and (d) 7.5% V2OsJTiO 2.

V2Os/ZiO2,

518

L.J. ALEMANY etal.

I

i

l

I

I

Vol. 31, No. 5

I

I

I

J I

B tzot

tioo

1(30~

900

80o

700

600

50o

40o

300

2OO

I

I

I

I00

Wavenumber cm ~

FIG. 3 FT-Raman spectra of fresh V205.

I

1200

I

I

i

!

1100

1000

500

I

I

I

I

|

i

!

i

!

i

i

800

700

600

500

400

300

200

Wavenumber cm ~

FIG. 4 Magnification of the 7.5% VzOs/TiO2 system.

90

Vol. 3 I, No. 5

VANADIUMTITANIUMOXIDE

519

The increase of the crystallite size must result in a decrease of the crystal density and, therefore, of the density of active Raman vibrational modes. Figure 5 evidences this trend. Even when this appears to be the main parameter accounting for the decrease of Raman intensity, structural disorder should not be ruled out as an additional explanation. Modification in the elementary cell of TiO 2 anatase upon addition of vanadium oxide have been observed (14). The variation in the elementary cell evidence the interaction between both oxides. High Resolution Electron Microscopy fttrther provides evidence of the existence of irregular lattice due to a misfit between TiO2 anatase and V205 phase (9), prior to the onset of TiO2 ruffle phase. There is a worse definition of the structures, and vibration modes are not so well defined. The modification of crystal parameters and progressive transformation into rutile must account for the decrease of Raman signal.

1.2

i

|

i

i

1.l~

0.8 0.6 0,4 f 0.2 0.0" 0.0

0.2 0.4 0.6 0.8 Theoretical Surface C o v e r a g e

1°0

FIG. 5 Dependence of theoretical surface coverage on the Raman relative intensities measured at 638 cm -1 and crystal density.

CONCLUSIONS Vanadium oxide affects the initial sintering of TiO 2 and the anatase-rutile transformation. These processes are faster at high calcination temperature. The crystal size of the anatase TiO 2 increases in the range below theoretical surface coverage with increasing V205 content The thermal treatment of the sample can give rise to the diffusion of vanadium into the TiO 2 crystal The lower transformation temperature observed on the vanadium-containing samples is justified by the high mobility of vanadium into the structure by substitutional replacements of Ti sites. Owing to the fact that the vanadium, as V(IV), is smaller than titania atom (0.64 vs. 0.68 A), this substitution results in an decrease of the elementary cell volume. A correlation is attempted between surface coverage, crystal size, and the Raman intensity.

520

L.J. ALEMANYet

al.

Vol. 31, No. 5

ACKNOWLEDGMENT

The financial support of CICYT Project PB93-1009 is gratefully acknowledged.

REFERENCES

1. J. Haber, A. Kozlowska, R. Kozlowski, J. Catal. 102, 52 (1986). 2. A. Vejeoux and P. Courtine, J. Solid State Chem. 63, 179 (1986). 3. G. Busca, P. Tittarelli, P. Forzatti and E. Tronconi, J. Solid State Chem. 67, 91 (1987). 4. G. Hausinger, H. Schmelz and H. Knozinger, Appl. Catal. 39 (1988) 267. 5. G.C. Bond, S. Flamers and R. Schukri, Faraday Discussion Chem Soc. 87, 65 (1989). 6. H. Eckert and I.E. Wachs, J. Phys. Chem. 93, 6796 (1989). 7. G.T. Went,.S.T. Oyama and A.T. Bell, J. Phys. Chem. 94, 4240 (1990). 8. R.Y. Saleh, I.E. Wachs. S.S. Chart, and C.C. Chersich, J. Catal. 98, 102 (1986). 9. Z.C. Kang and Q.X. Bao, Appl. Catal. 26, 251 (1986). 10 J. Mignot and D. Rondot, Acta Crystallogr. A 33,327 (1977). 11. R.A. Spurr and H. Myers, Anal. Chem. 31, 2481 (1978). 12. L.J. Alemany, F. Delgado and J.M. Blasco, React. Kinet. Catal. Lett. 53, 177 (1994). 13. G.C. Bond and S.F. Tahir, Appl. Catal. 71, 1 (1991). 14. P. Lostak, Z. Cernosek, E. Cernoskova, L. Benes and J. Krotuil, J. Mat. Sci. 28, 1189 (1993). 15. J.M. Lopez Nieto, G. Kremenic and J.L. Fierro, Appl. Catal. 61, 235 (1990). 16. I.E. Wachs, R.Y. Saleh, S.S. Chan and C.C. Chersich, Appl. Catal. 15, 339 (1985). 17. J. Haber, T. Machej, E.M. Serwicka and I.E. Wachs: Catal. Lett. 32, 101 (1995). 18. G.C. Bond, J. P6rez Zurita, S. Flamerz, P.J. Gellings, H. Bosch, J.G. Van Ommen and B.J. Kip, Appl. Catal. 22, 361(1986). 19. D.N. Waters, Spectrochimica Acta 50A, 1833 (1994). 20. M.A. Bafiares, L.J. Alemany, M. Carmen Jim6nez, M. Angeles Larrubia, F. Delgado, J.M. Blasco and J.L.G. Fierr0, J. Solid State Chem. (submitted). 21. Z.X. Lui, Q.X. Bao and N.J.J. Wu,. Catal. 113, 45 (1988). 22. L. Rey, R. Martino, L. Gambaro and H. Thomas, Latin Ing. Quim. Apl. 16, 231(1986).