Method Predicts Activity of Vanadium-Contaminated FCC Catalysts
A model has been developed that can predict the activity of vanadium-contaminated fluid cataIytic cracking (FCC) catalysts from laboratory data. It was derived from hydrothermal deactivation studies on metals-contaminated catalysts with silica/alumina matrices.
Commercial data are presented that confirm the model's predictions and demonstrate the susceptability of low-surface-area catalysts to deactivation by vanadium contamination.
Published laboratory studies(Ref. 1) have shown that commercially available fluid cracking catalysts with a higher surface area and high alumina matrix are more resistant to deactivation when contaminated with metals than catalysts with lower surface area and high silica matrices.
Recently, low matrix surface area process catalysts, made with an alumina-rich gel, have become widely used. The vanadium tolerance of these high alumina matrix catalysts was tested to see if their alumina content helped them maintain activity.
The experimental technique that was used to deactivate the catalysts is a modification of one originally published by Mitchell(Ref. 2). Fresh catalysts are impregnated with metals and then steam treated to simulate aging of the metals and deactivation of the catalyst that occurs in a commercial unit.
Speronello and Reagan found that steaming conditions had a strong influence on catalyst deactivation by metals. The laboratory studies of this article used the 1,450° F., 4 hr steaming conditions and the vanadium-to-nickel ratio of 2 that these authors found to be most representative of commercial units. The test methods used were identical to those described by Speronello and Reagan(Ref. 1).
Catalyst particles in a commercial unit are expected to have different levels of vanadium contamination depending upon their residence time in the unit. The experimental technique is accurate to measure the activity of particles at one vanadium level.
In order to find the average activity of the catalyst inventory, it is necessary to average the contributions of particles with a distribution of vanadium levels. A steady state model was derived to calculate the average activity for a unit operating at constant inventory, constant catalyst loss rate, and constant catalyst addition rate.
The physical and chemical properties of the five commercial catalysts chosen for this report are presented in Table 1. Catalysts EA and EB have alumina contents over 50 wt %. These alumina contents indicate the matrix contains a high level of alumina. Catalysts FH and BC have alumina contents nearly as high as the EA and EB catalysts.
The matrix surface areas of the five catalysts are indicated by their total surface areas at equal activity. In Table 1, 75% conversion on the micro activity test unit was chosen as the activity level for the comparison.
The catalysts are arranged in Table 1 with the highest surface area sample on the left.
Catalysts EA and EB have the highest matrix surface area. Catalyst FH has a moderate surface area, while CD and BC have low surface areas. All five of the samples contained rare earth exchange Y zeolite to catalyze the gas oil cracking reaction.
Fig. 1 presents the results of the laboratory hydrothermal deactivation studies. The intermediate and high matrix surface area products maintained enough zeolitic cracking activity to convert at least 50 wt % of the test gas oil at metals loadings up to 4,000 ppm vanadium.
The low surface area products could convert only 40 wt % after an identical deactivation. The spread between the low surface area catalysts and the high surface area catalysts is larger at high ppm vanadium contamination levels.
Fig. 1 shows that the reaction in the second order cracking activity (conversion divided by 100% minus conversion) can be adequately expressed by the relationship:
where r is the ratio of activity at vanadium loading V to the zero vanadium level activity.
This representation has the advantage of expressing the metals deactivation rate in terms of a single constant, V37, that can be read off the ln r vs. vanadium plot in Fig. 1.
V37 represents the amount of vanadium contaminant required to reduce the activity to 37% of the zero vanadium activity. The values of V37 for each of the four catalysts as determined from Fig. 1 are where their plots cross the r = -1 line. These values are:
The values of V37 increase for these five catalysts as surface area increases. Catalyst BC has the lowest value of V37 despite its high alumina content.
The catalyst deactivation equation described previously is not an adequate representation of a commercial unit where fresh catalyst is continuously added and metals-contaminated catalyst is constantly withdrawn.
The problem of calculating the average steady state activity for such a unit without metals has a solution that is well known in the industry(Ref. 3,4). Table 2 summarizes the derivation of an extension of this solution to the problem of metals-contaminated catalyst.
If vanadium contamination is not present, catalyst activity decays exponentially with time. To calculate the activity of a catalyst particle after t days, multiply the fresh activity by the decay fraction .
Vanadium contamination also accelerates deactivation exponentially. The same particle with vanadium poisoning has its activity further reduced by the fraction:
In order to find the average activity of a catalyst inventory, the contribution of each particle must be integrated over the range of particle ages. When the catalyst inventory, losses, and makeup are at steady state, the fraction of catalyst that has been in the unit more than time t is:
For the noncontaminated catalyst with this age distribution, the activity integral is:
For the metals contaminated case, the vanadium level V depends on the length of time the catalyst was in the unit. At any age t, the vanadium on the catalyst is:
where V is the average vanadium content of the inventory.
The integral for the vanadium contaminated average activity is then:
The results of both integrations are given in Table 2. The equation of deactivation with vanadium reduces to the well known zero vanadium deactivation equation when V is zero.
Table 3 presents commercial data on average catalyst activity vs. metals level and catalyst type for six commercial units. When laboratory data are used for the nondeactivated catalyst activity and V37, a hydrothermal deactivation constant, can be calculated for each case in Table 3. An average value was then found for each unit. The average was used in the equation in Table 2 to predict the activity. The calculations for all six refineries shows that the model predicts activities with the +/- 2 wt % accuracy of the data.
Note that refineries 1 through 4 used both a high matrix surface area catalyst and a low matrix surface area catalyst. For refineries 1 and 3, significant activity losses occurred on the low matrix surface area catalyst.
Refinery 2 was able to hold activity on the low matrix surface area catalyst only by doubling makeup to reduce metals levels which increased the fraction of uncontaminated catalyst in the inventory.
In refinery 4, a lower fresh activity high matrix surface area catalyst was able to hold the same activity as a higher fresh activity, low matrix surface area catalyst. In all four refineries, the high surface area catalyst was better able to retain its activity when contaminated with vanadium.
1. Speronello, B.K., and Reagan, W.J., OGJ, Jan. 30, 1984, p. 139.
2. Mitchell, B.R., Industrial and Engineering Chemistry Product Research and Development, Vol. 19, No. 2, 1980, pp. 209-13.
3. Anderson, S.L. and Matthias, R.H. Industrial and Engineering Chemistry, Vol. 46, 1954, p. 1296.
4. John, G.S., and Mikovsky R.J. Chemical Engineering Science, Vol.15 1961, pp. 172-5.
[Reprinted with permission from Oil & Gas Journal, July 15, 1985]