EFFECT OF VANADIUM ON THE DEACTIVATION OF FCC CATALYSTS

EFFECT OF VANADIUM ON THE DEACTIVATION OF FCC CATALYSTS

R.E. Roncolatto and Y.L. Lam

Centro de Pesquisas da PETROBRAS- Divisão de Catalisadores, Cidade Universitária

Quadra 7 - Ilha do Fundão Rio de Janeiro- RJ- Brazil, CEP 21949-900

phone: [phone] - fax: [phone]

e-mail: [e-mail] - [e-mail]

(Received: November 5, 1997; Accepted: January 28, 1998)

Abstract - This work provides concrete evidence that vanadium causes the destruction of the zeolite in the FCC catalysts by a mechanism of acid attack or solid-solid transformation, as well as additional dealumination of the zeolite framework in the presence of steam and at high temperature. While these effects resulted in the reduction in crystallinity (zeolite Y content), specific area and unit cell size of the Y zeolite as the amount of vanadium in the catalysts increased, the reduction in activity was the most pronounced. The differences in these behaviors were interpreted and the model can be used for better catalyst formulation or screening.

Keywords: Deactivation of FCC catalysts, vanadium attack.

INTRODUCTION

Currently there is a worldwide tendency for refiners to crack heavier oil feedstocks. Heavy feeds differ from gas oil by their much higher boiling range and their higher content of polynuclear naphthenes and aromatics, resins, asphaltenes, contaminant metals (V, Ni, Fe, Cu), sulfur, nitrogen and Conradson carbon.

Vanadium, while not the only contributor to fluid-cracking catalyst deactivation, frequently dictates the amount of fresh catalyst added to the FCC unit to maintain activity.

Natural vanadium complexes are present in heavy gas oil in the form of porphyrins. In the FCC unit the complexes decompose and the metals are deposited on the catalyst. Vanadium reduces the activity and selectivity of the FCC catalyst due to its tendency to cause undesirable reactions of hydrogen and coke formation.

Various authors have proposed the mechanism of vanadium attack on catalysts. Using various catalysts and characterization techniques such as XRD, BET and electron microprobe Wormsbecher et al. (1986) proposed that V₂O₅ be transformed into volatile vanadic acid, H₃VO₄, in the regenerator by a reaction with water causing zeolite destruction due to the hydrolysis of the silica-alumina framework. While studying Y zeolite exchange with different cations and silicalite, Pine (1990) observed that vanadium attacks both structures in the presence of steam. XPS results indicated that vanadium does not penetrate the zeolite crystal, suggesting that the attack is mainly on the surface. According to these results, Pine proposed that the attack by vanadium is probably at the Si-OH bond and not the tetrahedral aluminum site of the lattice. Working with various catalysts, Huifang et al. (1996) proposed that the vanadium destruction mechanism of zeolite attack is based on accelerated framework dealumination. The V₂O₅-Na₂O system would dissolve alumina originating from zeolite framework aluminum and non-framework aluminum. Occelli (1991) studied various zeolite types and observed, with relation to rare earth (RE), that resistance to vanadium diminishes when RE ions are present since they have the tendency to react with vanadium, and this in turn causes destabilization of the framework. Pine (1990) observed that RE does not change the tolerance of vanadium, but does cause an indirect effect by increasing the base steam stability of the zeolite.

All mechanisms can exist depending on vanadium content, catalyst composition and RE content. Thus, this work has as its objective to verify the extent of the attack on PETROBRAS catalysts with increasing vanadium content, the influence of RE and the attack mechanism involved.

EXPERIMENTAL

FCC catalysts used in this study were produced at FCC S. A.. They contained Y zeolite incorporated into a matrix of SiO₂-Al₂O₃-kaolin, in the absence (code CAT) or presence (code CAT.R) of RE.

Both catalysts were impregnated with different vanadium contents (1000, 3000 and 5000 ppm) according to the Mitchell (1980) method, which uses vanadium naphthenate in benzene solution.

Catalysts were submitted to hydrothermal treatment in a fixed bed for about 1077 K during 5 hours, since the catalysts are exposed to this pretreatment before being evaluated. This procedure aims to simulate the properties of an equilibrium catalyst in the FCC units.

Chemical composition was determined by XRF. Crystallinity and unit cell size (Ao) were determined in the Phillips X-ray diffractometer using Ka Cu radiation. Specific surface area was determined by nitrogen adsorption at 77 K by the BET method in the range of p/po= 0.06-0.2.

Catalyst activities were measured by cumene cracking at 643 K, by passing stream of nitrogen containing a 0.8 mol % of cumene at 38 ml/min on about 100 mg of catalyst.

RESULTS

Table 1 shows the properties of the virgin catalysts and catalysts after hydrothermal treatment. It can be observed that with hydrothermal treatment in the absence of vanadium, there is a drastic loss of specific area and decrease in unit cell size, yet about 100% crystallinity is preserved. The significant Ao reduction indicates the drastic dealumination that the zeolite suffers due to calcination at a high temperature in the presence of steam. RE containing catalyst better preserves the Ao due to the stabilizing effect of RE in the framework. In spite of the strong dealumination, it was observed that the crystallinity was not altered. This is probably due to the fact, already reported by Gélin and Courières (1991), that silicon atoms in the matrix, in the form of volatile Si(OH)₄ species, are transported towards the zeolite component and subsequently reincorporated into the zeolite framework in the place of the expelled aluminum atoms, thus preventing the collapse of the zeolitic structure. In this way, the amount of crystalline material remains constant. Area reduction is justified by the blockage caused by NFA (nonframework alumina), normally located on the particle surface, to the N₂ molecules used in the BET method for area measurement, as well as to area loss of the matrix (Yang et al., 1994).

Table 1:
Properties of the catalysts

Catalyst	CAT1	CAT.R2
Na2O ( wt%)	0.76	0.99
RE2O3 ( wt%)	0.11	1.42
Al2O3 ( wt%)	29.3	29.5
SiO2 ( wt%)	68.9	67.1
SBET (m2/g)	303	289
A0 (A)	24.56	24.55
Hydrothermal Treatment with no V
%Yret3	100	100
SBET (m2/g)	184	185
A0 (A)	24.25	24.30
Hydrothermal Treatment with 1000 V ppm
ppmV	1073	1035
%Yret4	93	88
SBET (m2/g)	173	177
A0 (A)	24.24	24.27
Hydrothermal Treatment with 3000 V ppm
ppmV	2456	2751
%Yret4	63	71
SBET (m2/g)	136	144
A0 (A)	24.22	24.27
Hydrothermal Treatment with 5000 V ppm
ppmV	3208	4256
%Yret4	43	29
SBET (m2/g)	111	90
A0 (A)	24.20	24.23

CAT¹=Catalyst with no RE; CAT.R ²= Catalyst with RE

³ %Yret³=crystallinity retention relative to the virgin catalyst

⁴ %Yret⁴= crystallinity retention relative to the catalyst calcined with no vanadium

When heat treatment is preceded by vanadium impregnation in different contents, there is a progressive crystallinity, area and unit cell size loss. Figures 1 and 2 summarize these effects. In Figure 1, it can be observed that area retentions (%SBETret) are bigger than the crystallinity retentions (%Yret), which is in accordance with the results of Wormsbecher et al. (1986). The probable explanation is that the matrix is attacked less by vanadium (Yang et al., 1994) but contributes with part of the total area of the catalyst, while the crystallinities are solely due to the zeolites. It can also be noted that, with the increase in vanadium content there is a tendency to widen the differences between the relative retentions of area and crystallinity, presumably due to the more intense attack on the zeolites and not so much on the matrixes.

It can also be verified in Figure 1 that the catalyst with RE (CAT.R) presents a higher hydrothermal stability than the catalyst with no RE (CAT). In his study of the effect of vanadium in catalysts with and without RE with low sodium, Pine (1990) also obtained the same tendency, which can be explained in terms of the stabilizing effect of RE in the zeolite structure.

In Figure 2, it can be observed that vanadium also causes an extra dealumination of the zeolite contained in the catalyst, reflected in the Ao reduction. In fact, various authors (Yang et al., 1994 and Occelli, 1996) proposed that the vanadium in the gas oil be transformed in the regenerator into volatile vanadic acids (such as H₃VO₄ or H₄V₂O₇), which should be capable of leaching Al from the zeolite framework (probably as AlVO₄), this way justifying the reduction of Ao. This fact can contribute even more to the greater activity loss of the catalyst, since the number of acid sites is proportional to the concentration of framework aluminum.

In the cracking of cumene only benzene and propene were obtained as products. As expected, the catalyst without vanadium was the most active. Under test conditions, the activity of CAT.R for converting cumene was 1.54 m mol/gcat.seg. Figure 3 shows activity retention (%Aret) with vanadium content. Results of area and crystallinity retention were also included for comparison. It can be observed that the activity loss is more pronounced, except for the sample with the highest vanadium content. This behavior is in accordance with the speculation above, that besides the destruction of the active phase (the zeolite), the dealumination caused by vanadium reduces by even more the number of active sites per zeolite mass.

Figure 3 shows the influence of vanadium on the cracking ability of the catalyst containing RE. There is practically a linear relation between vanadium level and cumene cracking activity retention (%Aret). It can be observed that a reduction greater than 50% takes place for about 4000 ppm of vanadium.

DISCUSSION

Vanadium deactivation of zeolites and FCC catalysts is a classic problem in FCC, and it is expected that key catalyst properties such as area and activity decrease as vanadium contents increase. However, it is important to compare the variations in catalyst properties in order to identify the best parameter for use in catalyst selection and to create a physical model for better catalyst formulation.

Area retention is frequently used as an index for correlation with the catalyst performance. Figure 4 shows the variations in the three catalyst properties measured in this study. With an increase in vanadium content, activity and crystallinity losses are bigger than area loss. Thus, a catalytic test may be a better correlation with commercial performance than is textural characterization.

It is important to explore again the observation that loss of area and loss of crystallinity are both nonlinear as functions of vanadium. This implies that the matrix could be used to retain a fraction of vanadium for low levels. The more effective the matrix, the lower the decrease in area/crystallinity at low amounts of vanadium, causing more deviation from linearity. However, as the amount of vanadium increases, zeolites that are more stable for higher vanadium contents must be used. Other forms of catalyst protection against the deactivation of vanadium normally used by refiners are metal traps in the catalyst or passivators mixed in the gas oil, to prevent the contact of vanadium with the catalyst. If these materials are highly effective, an increase in nonlinearity of the decrease in area and crystallinity can be expected as a function of vanadium level.

For the catalysts studied, the correlation between dealumination and increasing vanadium content supports the mechanism of vanadium attack on the silica-alumina zeolite framework, which acts more directly on the aluminium. For this catalyst, RE increases hydrothermal stability in the presence of vanadium.

Figure 1:
Crystallinity (%Yret) and area retention (%SBETret) with relation to the catalysts that were hydrothermally treated without vanadium.

Figure 2:
Effect of vanadium on unit cell size with heat treatment.

Figure 3:
Cumene cracking activity retention (%Aret) as a function of vanadium content on CAT.R

Figure 4:
Correlation of cumene cracking activity retention (%Aret) and Crystallinity (%Yret) and area retention (%SBETret)

CONCLUSIONS

Two types of cracking catalysts, those with and those without RE, were shown to be heavily attacked by vanadium when submitted to hydrothermal treatment (high temperature and water vapor pressure). About 50% of catalyst area loss was observed for about 4000 ppm of vanadium. RE guarantees higher stability for the zeolitic structure. Vanadium attacked the silica-alumina framework of the zeolites, and more directly the aluminum, causing thereby a reduction in unit cell size and showing its function as an acid in the removal of aluminum from the framework. Consequently, the reduction of catalyst activity can be attributed to more than one cause, the loss of active phase as well as a reduction in the number of sites per mass of active phase.

• Gélin, P. and Des Couričres, T., Role of Amorphous Matrix in the Hydrothermal Aging of Fluid Catalytic Cracking Catalysts, Applied Catal., 72, 179 ( 1991).
• Huifang, P., Xiaofeng, W., Aijun, T., Zhihong, S. and Gaoshan, Z., The Design of Vanadium Trapping System for FCC Catalysts, Chinese J. Chem. Eng., 4, Nş 2, 120 (1996).
• Mitchell, B. R., Ind. Eng. Chem. Prod. Res. Dev., 209 (1980).
• Occelli, M. L., Metal-Resistant Fluid Cracking Catalysts, ACS Symp. Ser., 452, 343 (1991).
• Occelli, M. L., Vanadium Resistant Fluid Cracking Catalysts, Studies Surface Science and Catalysis, 100, 27 (1996).
• Pine, L. A., Vanadium-Catalyzed Destruction of USY Zeolites, J. Catal., 125, 514 (1990).
• Wormsbecher, R. F., Peters, A. W. and Maselli, J. M., Vanadium Poisoning of Cracking Catalysts: Mechanism of Poisoning and Design of Vanadium Tolerant Catalyst System, J. Catal., 100, 130 (1986).
• Yang, S.-J., Chen, Y.-W. and Li, C., Vanadium-nickel Interaction in REY Zeolite, Applied Catalysis, 117, 109 (1994).
• Yang, S. J., Chen, Y. C. and Li, C., Metal-resistant FCC Catalysts: Effect of Matrix, Applied Catalysis, 115, 59 (1994).

Publication Dates

• Publication in this collection
  09 Oct 1998
• Date of issue
  June 1998

History

• Accepted
  28 Jan 1998
• Received
  05 Nov 1997