Sep-2017

FCC gasoline desulphurisation for Tier 3 sulphur compliance

FCC gasoline desulphurisation is used widely in refineries around the world. It is growing in importance because of new regulations that require sulphur in gasoline to be reduced to 10 ppm.

George Hoekstra
Hoekstra Trading

Article Summary

FCC gasoline is desulphurised in fixed bed hydrotreaters. Side reactions occur that cause octane loss. The key to the process is to remove the sulphur with minimum octane loss.

When designing and optimising these units, it is critical to know how much octane will be lost and how octane loss is affected by changes in gasoline feed quality, catalysts, process design, and process severity.

Hoekstra Trading has done three years of research on this process culminating in development of a new process model that makes modern octane analysis available to refining engineers for the first time, using modern chemical engineering tools and the most detailed data ever used for this purpose.

The model is being used today by refiners, technology suppliers, and catalyst suppliers to optimise feed, sulphur specifications, unit operations, and catalysts for this process; and to evaluate revamps, new unit investments, and Tier 3 gasoline sulphur strategies.

This paper describes how two refiners have used our data, methods, and software to help make decisions relating to their Tier 3 gasoline strategy.

Case 1 - Tier 3 gasoline sulfur credit strategy
This US refinery has an FCC gasoline desulphuriser with extra desulphurisation capacity. They qualify as a small refiner, and so are not required to make 10 ppm sulphur gasoline until 2020. But they are considering making sub-10 ppm sulphur gasoline now and banking Tier 3 sulphur credits for future sale to other refiners.

This strategy would involve increasing reactor severity and taking a penalty in octane loss today in exchange for revenue from future sale of Tier 3 credits. To evaluate this option, they needed reliable estimates of how much octane loss would occur in more severe operation on different feeds.

We analysed samples of the current feed and product using a detailed chemical analysis that identifies every compound in the gasoline. This detailed analysis is a cornerstone of our model. The data was run through Hoekstra Trading’s software which analyses the individual reactions occurring in the unit, the octane impact of those reactions, the effects of feed composition and process severity on process performance and octane loss.

Detailed feed composition - case 1
The feed to this gasoline desulphuriser contains 528 compounds, and the product contains 455 compounds. The data was fed into the input table of the Hoekstra Trading software. Table 1 shows a segment of the input table. This input table continues for 759 rows which is the total number of different compounds we have found in 70 gasolines we have analysed.

Reaction analysis - case 1
We are especially interested in olefins because olefin saturation is a major cause of octane loss.

On Figure 1, the black curve shows the olefin-by-carbon number distribution in the feed. The feed contains C4-C10 olefins, mostly C5s.

The tan curve is the olefin distribution in the product. Comparison of the two curves shows that no C5- olefins have been converted in the unit. This is because C5- olefins have been split out in an interstage splitter, bypassing the desulphurisation reactor. C6, C7, and C8 olefins have been partly converted. The feed contains a total of 20 wt% olefins, and the product contains 15.6% olefins, so the total olefin saturation is 4.4 wt% on feed.
 
Octane impact - case 1
How much octane is lost by saturation of 4.4 wt% olefins? The answer depends strongly on which particular olefins are being saturated.

The saturation of linear olefins - produces normal paraffins, which are very low in octane. The extent of linear olefin conversion is seen by comparing the normal paraffin distribution curves of feed and product (Figure 2): Comparison of the tan and black curves shows that normal paraffins are produced to the extent of 1.1 wt% on feed. It is mostly C6 (n-hexane) that is being formed by saturation of linear C6 olefins.

The saturation of iso-olefins - produces iso-paraffins, which are higher in octane than normal paraffins. The extent of these reactions is seen by comparing the feed and product iso-paraffin distributions (Figure 3): We see iso-paraffins being produced across a wide range, C6-C11. The feed contains a total of 28.6% isoparaffins, and an additional 3.2 wt% has been produced, mostly by saturation of iso-olefins.

Similarly, we compared the distribution curves for reaction of other compound groups which include cyclic olefins, naphthenes, aromatics, unknowns, oxygenates, and C14+ hydrocarbons. All these charts are produced immediately in the Hoekstra software upon entering the feed analytical data.

Octane/sulphur performance curve - case 1
The software contains an octane model that calculates octane and octane loss from the detailed gasoline composition at the molecular level. Figure 4 summarises the factors that contributed to octane loss in this unit on the day the samples were taken: Starting at the left, consumption of olefins caused 2.7 octane loss. Production of normal paraffins reduced octane another 0.6 RON. Those octane losses were partly offset by production of relatively high octane isoparaffins and naphthenes. Another 0.7 octane loss came from reactions involving higher molecular weight aromatics and unknowns (contrary to common perception, these reactions often contribute appreciably to octane changes in this process).

The total octane loss was 1.3 RON units, mostly due to saturation of linear C6-olefins and dienes to form normal hexane.

 The Hoekstra model then predicts the octane/sulphur performance curve for different scenarios based on feed composition, gasoline cut points, unit design, catalyst and process selectivity. The prediction is based on desulphurisation and olefin saturation rates and rates of other reactions that we have measured in pilot plant and commercial unit tests. Figure 5 shows the predicted octane/sulphur performance curve for Case 1, given the feed processed on the day of our test. The left-most point on this curve shows that at 10 ppm sulphur, the predicted octane loss is 3 RON.

We compared this octane/sulphur performance curve to those for a wide range of cases from our pilot plant and commercial tests. By comparison with other units, this feed is easy to desulphurise. The olefin type distribution is favourable from the standpoint  of minimising octane loss. The unit has excess kinetic capacity, and deep desulphurisation can be achieved at moderate temperature.

We concluded the octane/sulphur performance capability of this unit is high compared to others in the United States, and that this refinery is well-positioned to make sub-10 ppm gasoline at relatively low incremental cost. This is a competitive advantage that can be exploited by generating and banking standard Tier 3 sulphur credits