FCC CATALYST INCREASES ISOBUTYLENE YIELD AT EUROPEAN REFINERY

Steve Benton
Engelhard Ltd.
Sutton, U.K.

The Clean Air Act of 1990 ushered in the reformulated gasoline era in the U.S. and turned attention to fluid catalytic cracking units (FCCUs) as a source of iso-olefins for producing methyl tertiary butyl ether (MTBE) and other oxygenates.

A new series of catalysts, called IsoPlus, has been developed to help refiners increase isobutylene yield in FCCUs.

Laboratory evaluations confirmed the selectivity benefits of IsoPlus catalysts for producing high iso-olefin yields for a wide variety of FCC feeds and conditions.1 Commercial trials at a number of refineries around the world have since verified the results attained in the laboratory.2

One recent commercial application at a European refinery showed, on average, a 30% increase in isobutylene yield and about a 10% increase in yield of other C4 olefins. The catalyst also increased RON and MON substantially and improved bottoms upgrading.

In the development work leading to this catalyst, Engelhard Corp. found that isobutylene yields of available FCCU catalysts differed by as much as a factor of two.2 The highest yields were derived from USY-zeolite catalyst with low rare earth content and relatively high matrix activities. High-rare-earth, low-matrix-activity catalysts showed the lowest isobutylene yields.

Compared to conventional catalysts, IsoPlus catalysts boost isobutylene yield 69-200% in microactivity tests (MAT). Two such catalysts-IsoPlus 1000 and IsoPlus 2000-have been produced and used commercially.

INCREASED ISOBUTYLENE

The striking increase in butenes production possible when cracking with IsoPlus catalysts can be explained by the carbenium ion mechanism (Fig. 1). This mechanism is considered the principal route by which catalytic cracking takes place.

The isomerization of linear butenes cannot occur directly and involves three intermediate stages.3 The intermediates involved in this process, however, are reactive and can be diverted easily by processes such as hydride transfer to produce butanes (Fig. 2). Reducing hydride transfer thus is important in maximizing isobutylene yield.

The reaction mechanisms involved in this process suggest two ways of increasing isobutylene: increase total C4 yield (e.g., through conversion), and improve selectivity to isobutylene at the expense of iso and normal butane. Isobutylene will not increase by the isomerization of n-butenes as the required intermediate is reactive and easily diverted by hydride transfer to produce butanes.

Fig. 1 also suggests that, if isomerization of the surface carbenium ions is rapid relative to the other reactions, the equilibrium value (expressed as isobutylene yield divided by total butenes yield) will be a limitation. As temperature increases, the equilibrium value decreases (Fig. 3).

At the temperatures typically found in FCCUs (950-1,000 F. or 510-538 C.), the equilibrium value should be about 0.45. But in most FCCUS, this ratio normally is 0.20-0.35, which suggests that it can be improved in many units.

Deeper cracking of hydrocarbon molecules is obtained when fewer hydride transfer reactions occur to interrupt carbenium ion cracking. This leads to production of more LPG, less gasoline, and a more olefinic product for C3, C4, and gasoline boiling-range species (i.e., more C5 iso-olefins, which can be used for synthesis of tertiary amyl methyl ether). Higher olefinicity also helps increase gasoline octane.

One further potential benefit of reduced hydride transfer is a reduction in the conversion of naphthenes to aromatics. The resulting lower aromatic level can help refiners better achieve the targets set for reformulated gasoline.

RESULTS

The 150,000 b/d European refinery that used IsoPlus catalyst to increase isobutylene instituted a series of steps to improve yields. These steps will provide an understanding of the role IsoPlus catalyst played in helping the refinery achieve its production goals.

The unit involved was a UOP high-efficiency FCCU processing a high-sulfur, low-metals vacuum gas oil before the IsoPlus trial. The refinery operated a 3,000 b/d C4 HF alkylation unit downstream of the FCCU.

A 1,000 b/d MTBE unit was being built when the IsoPlus catalyst was introduced. In addition, a new FCCU feed hydrotreater had been commissioned not long before the trial. This unit had greatly improved feed quality and altered product yields.

The FCCU was using a high-rare-earth gasoline catalyst (1.9 wt % ReO level) when the feed hydrotreater came on line. Table 1 shows yields relative to operation on nonhydrotreated feed.

The hydrotreated feed dramatically increased conversion, gasoline yield, and total LPG yield. Absolute coke increased slightly, but specific coke (coke conversion) decreased.

Total fuel gas yield declined, even though conversion increased almost 10%. Hydrogen sulfide yield decreased substantially because of the reduction in feed sulfur.

Although total LPG yield increased, its olefinicity declined substantially, particularly for C4S. In addition, gasoline RON decreased while MON increased.

The changeover to hydrotreated feed caused several problems:

The reduction in RON limited the refinery's gasoline output, a critical issue. This is why riser temperature was increased 12 C. relative to the nonhydrotreated feed case.
The 40% increase in total LPG overloaded the LPG recovery section and forced a 20% reduction in feed rate. Decreasing the riser temperature to reduce LPG yield was not an option, given the concerns about RON.
The decrease in isobutylene yield meant the MTBE unit, which was to start up shortly, would have to operate well below its design capacity.

In response to these problems, the refinery reduced the rare earth level in the fresh catalyst and added a ZSM-5 additive (Engelhard's Z-100) until it comprised 4% of effective catalyst inventory. The refinery also decreased riser temperature, used a slightly different feedstock, altered the fresh catalyst, and introduced other changes that make it hard to define the specific effects of the additive.

Table 2 illustrates the results of these changes relative to the hydrotreated feed case in Table 1.

LPG olefinicity, potential MTBE and alkylate production, and gasoline octanes improved. Although the refinery was satisfied with these improvements, it lost another 8% in feed rate as total LPG yield increased again. The refinery also still lacked sufficient isobutylene to supply the MTBE unit.

NEW CATALYST

The refinery decided to use IsoPlus 1000 catalyst to increase isobutylene yield. The early part of the trial reflects the influence of both IsoPlus catalyst and the ZSM-5 additive.

The data in Table 3 show two periods of operation using the new catalyst. ZSM-5 addition stopped midway through the first period. ZSM-5 activity declines fir faster than the rate at which it is removed from inventory, so its influence is probably less than is indicated by its percentage in the inventory.

The catalyst was effective at increasing isobutylene yield. The 0.84 wt % increase in the first period is almost a 60% increase over the hydrotreated-feed base case. This is higher than expected and suggests that ZSM-5 aided the new catalyst's performance.

Yields of C3 and other C4 olefins also increased in the first period. Although yields of saturated LPG products fell significantly, the overall increase in unsaturated compounds pushed total LPG yield above the level of the ZSM-5-only case and reduced feed rate by an additional 3%.

Although the feed rate was reduced, the refinery could satisfy the demands of both the MTBE and C4 alkylation plants. Large increases in RON and MON (respectively, 1.9 and 1.0) meant that gasoline octane was no longer a significant constraint, and reaction severity could be relaxed.

In the second operating period, several months into the trial and at the lower reaction severity, IsoPlus 1000 comprised more than 75% of catalyst inventory. Average riser temperature was 7 C. lower than in the first period and 22 C. lower than in the base case. Conversion was 5% lower than in the base case and total LPG yield was about 3.6 wt % less.

Although, in the second period, isobutylene yield increase was less than in the first period, the improvement still is well above the base case. Absolute isobutylene yield averaged 2 wt % during the second period, compared to 1.46 wt % before the trial.

Yields of other butenes were slightly lower than in the base case, the decline in each being about the same (0.25-0.32 wt %). More significant reductions are apparent in the isobutane and propylene yields.

The net reduction in LPG freed the key restraint and enabled the feed rate to increase 21% above the first period and 10% above the base case. This increase in feed rate and isobutylene yield gave an MTBE yield that was 51% greater than in the base case. This is much higher than in the maximum feed rate, nonhydrotreated case.

Although C4 alkylate yield was less than in the previous cases, it met the needs of the alkylation unit and was better than in the nonhydrotreated feed case.

NET EFFECTS

The dynamic and complex nature of the FCC process makes it hard to separate the yield effects of IsoPlus 1000 catalyst from those of ZSM-5 additive in the data shown in Tables 1, 2 and 3.

To separate the effects, Engelhard analyzed the data using a multivariable linear regression technique, which screens out other operating effects, such as riser temperature and feed quality. The results of this analysis are shown in Table 4.

The yield differences shown are pure catalytic effects relative to the unit operating with a maximum-octane-barrel gasoline catalyst. The results confirm much of what was expected, with the exception of a few surprises that warrant further explanation.

The increased butenes yield with the new catalyst, particularly the isobutylene fraction, and the reduced saturated LPG yields were at the upper end of what was anticipated from laboratory work. This increase would be the expected result from a reduction in the hydride transfer mechanism.

The apparent reduction in propylene yield was unexpected because test work suggested that propylene would increase. The decline may result from the reduced conversion levels the unit experienced with the new catalyst.

Table 4 shows that actual conversion decreased by about 1.5 wt % with the new catalyst. MAT activity was 6 numbers lower for the IsoPlus 1000 catalyst compared to the base gasoline catalyst (Table 2).

In some circumstances, a drop in activity would be a disadvantage. Here it is an advantage because it appears to have allowed LPG selectivities to improve beyond what was expected.

It may seem strange that the difference between the 1.5 wt % decrease in observed conversion and the 6 number drop in MAT activity is so large. MAT activity is, after all, a measure of the conversion on a MAT unit operated at standard conditions (at Engelhard, 910 F. and a catalyst-to-oil ratio of 5) and with standard feed. Therefore, MAT test results are relevant to one set of conditions and can only indicate activity trends.

Previous studies show that apparent differences in activity are less significant as MAT catalyst-to-oil ratio is increased. Because the FCCU operated at catalyst-to-oil ratios in the 7-9 range, the large difference is not so Surprising. Indeed, the information indicates that, for this unit at these operating conditions, only 25% of catalyst MAT activity actually increases unit conversion.

The reduction in gasoline yield from the IsoPlus catalyst was less than half that anticipated from the initial laboratory testing. This suggests that this unit may be in the overcracking region, where a reduction in conversion results in an increase in gasoline yield.

Light cycle oil (LCO) and bottoms yields in Table 4 show that the new catalyst appears to have a high bottoms upgrading potential. Bottoms yield is almost 0.5 wt % less with the new catalyst, even though conversion is almost 1.5 wt % lower overall. Better bottoms upgrading and lower conversion mean a significant increase in LCO yield (about 1.9 wt %).

One of the interesting features of IsoPlus 1000 in this trial was that there was no real change in coke and gas (C2-) selectivities. Subsequent MAT studies on various equilibrium catalyst samples from the unit indicate that the catalyst actually improved coke selectivity (Table 5).

Another positive aspect of the catalyst switch is the substantial increase in gasoline octanes achieved-almost 2 RON and 0.7 MON (clear). Some refiners may find this attractive in its own right.

ZSM-5 EFFECTS

Yield changes attributable to ZSM-5 are similar to those seen in other units (Table 4). Yields of all LPG components except propane increased significantly. The biggest effect was seen for propylene.

The next largest yield shift was seen in normal butene, followed by trans-2-butene, then cis-2-butene. ZSM-5 had the least impact on isobutylene production, only about half that achieved by IsoPlus 1000.

Most of the increase in LPG yield was balanced by a reduction in gasoline yield. There was also, however, an appreciable reduction in LCO yield. This is not unexpected, for the front end of the LCO cut contains hydrocarbon molecules small enough to react with ZSM-5.

This also explains the apparent increase in unit conversion with ZSM-5. Unlike IsoPlus 1000, ZSM-5 cannot convert the large molecules in LCO or the fractionator bottoms. Coke and fuel gas yields are effectively unchanged by ZSM-5. ZSM-5 also boosted RON and MON about the same amount. The additive increased the absolute yield of LPG olefins, but it was much less advantageous than IsoPlus 1000 because of the net increase in LPG. The constraint on total LPG processing capacity meant that the maximum feed rate with ZSM-5 was much lower than with the IsoPlus catalyst.

The new catalyst significantly increased olefin concentration in LPG without increasing total LPG yield, whereas ZSM-5 tended to increase yields of all LPG components. Indeed, the olefinicity in the total C4 cut actually declined slightly with ZSM-5, although it obviously increased in the case of the C3 fraction.

E-CAT SELECTIVITIES

Table 5 shows MAT selectivity details from equilibrium catalyst samples from the commercial trial. The test followed Engelhard's standard MAT procedure except a typical hydrotreated feed sample supplied by the refinery was used.

The trends were similar to those found in the commercial operation: increased C4 olefins, improved bottoms upgrading, and reduced LPG yields overall (C3, C3=,iC4, and nC4).

The refiner continues to use the IsoPlus 1000 catalyst in its FCCU. At the beginning of 1995, the catalyst had been in use for about 18 months.