Apr-2015

Hydrogen transfer and ZSM-5 in maximising propylene

Used at high concentrations to maximise propylene from the fluid catalytic cracking unit, ZSM-5 strongly influences a complex matrix of reactions.

BART DE GRAAF, MEHDI ALLAHVERDI, CHARLES RADCLIFFE, MARTIN EVANS and PAUL DIDDAMS
Johnson Matthey Process Technologies

Article Summary

The fluid catalytic cracking (FCC) unit is a versatile tool that can readily accommodate the need for increasing propylene demand. In the Middle East and Asia, new petrochemical FCC units are being built to add capacity, but in the US the advent of cheap and abundant natural gas liquids (from shale gas) displacing heavier feeds in steam crackers has reduced the production of propylene. Maximising propylene production in the FCC can be achieved by changes in operation and especially the use of ZSM-5 containing additives.

For maximum (or just increased) propylene operations, the feedstocks can be as diverse as a light hydrotreated VGO to blends containing atmospheric and vacuum residues. The optimal cracking catalyst for each feed is very different. Typically, it is assumed that ZSM-5 effects are additive to those of the base catalyst. The effects of hydrogen transfer in the FCC, ZSM-5 and combinations of both are reasonably accurately described at low concentrations (up to about 10%) of ZSM-5 additive in the circulating inventory. However, for operations where high concentrations of ZSM-5 additives are used to maximise propylene the simple models of these effects no longer hold true. Both catalysts (ZSM-5 and base catalyst) in the inventory guide an elaborate but elegant interplaying reaction matrix between all components from gasoline to LPG range. This has a surprising influence on ZSM-5 effects at high concentrations and helps to explain apparent contradictions in the literature. In a previous article1 we outlined mechanisms for ZSM-5 in FCC; here we expand this picture to include the effects of base catalysts with various levels of hydrogen transfer on ZSM-5.

Hydrogen transfer in the FCC
The role of hydrogen transfer in FCC is a well-known phenomenon and its effect on LPG olefins is generally assumed to be well understood. Hydrogen transfer reactions in FCC are bimolecular reactions. Naphthenes donate hydrogen to olefins and become aromatics while the olefins are hydrogenated into paraffins. Paraffins are more stable than olefins which reduces the rate of cracking reactions. Therefore hydrogen transfer increases gasoline yield and lowers LPG yield. Simultaneously, olefinicity is reduced in both fractions. This simple model is sufficient to describe the effects of oil cracking over base catalyst, or over base catalyst with low concentrations of ZSM-5 blended in. However, this model is insufficient to describe the catalytic interactions in high propylene operations, where high amounts of ZSM-5 containing additives are present with the base catalyst in the circulating inventory.

Hydrogen transfer is affected by several factors:
• Base catalyst: the higher the rare earth stabilisation of the base catalyst, the higher the acid site concentration in Y-sieve and the higher the rate of the bimolecular hydrogen transfer reactions
• Feed: feeds that contain a lot of hydrogen donors (naphthenes, typically hydrotreated feeds) have higher rates of hydrogen transfer
• Unit design/configuration: higher contact times and more back mixing increase hydrogen transfer in the FCC unit
• Unit operation: decreasing the hydrocarbon partial pressure (for instance, using steam in the riser) reduces the rate of hydrogen transfer and is a common feature in maximum propylene FCC unit designs.

These effects combine to determine which of the various pathways, such as cracking and hydrogen transfer, are balanced against each other.

The role of ZSM-5 on LPG yield and gasoline composition
Hydrogen transfer reduces the gasoline range olefins. As cracking rates of olefins over ZSM-5 are much higher than those of paraffins, an increase in hydrogen transfer reduces the effectiveness of ZSM-5 additives. Additionally, not all gasoline range olefins are equally crackable. Straight-chain normal olefins (C7 and higher carbon number) are very crackable. Their branched isomers suffer from diffusion limitation and are therefore much less crackable. Therefore, once all the straight-chain olefins have reacted, no further increases in propylene can be expected.²

The acid sites of ZSM-5 can readily crack olefins or isomerise in monomolecular reactions, but they can also oligomerise small olefins. The Mobil Olefin to Gasoline and Distillate (MOGD) process³ utilises coupling of light olefins. Lower temperatures and higher pressures favour oligomerisation of propylene and butylene. Propylene can readily form large molecules under the right conditions as the existence of polypropylene shows. At higher temperatures and lower pressures, thermodynamics heavily favour small olefins, but with a distribution tail for higher olefins. For example, cracking of pentene over ZSM-5 is feasible and occurs readily, but only via a dimerisation step.⁴ At elevated temperatures, adsorption of large olefins is reduced and the likelihood for a bimolecular reaction involving two large olefins becomes small and cracking becomes the dominant mechanism to shape the olefins distribution curve. Note that at both low and high temperatures both mechanisms occur, but relative reaction rates change with temperature.

The key question is whether LPG range olefins and their distribution in the FCC is in thermodynamic equilibrium or is kinetically controlled? Kinetic control is demonstrated by the fact that adding ZSM-5 additive to an FCC unit not only shifts LPG yields, but also substantially changes the ratio between propylene and butylenes. If the olefin distribution were controlled by thermodynamic equilibrium, this shift in composition would not have been possible.

Pilot plant study
An ACE pilot plant study was conducted to establish the effects of ZSM-5 additives on light olefin yields and gasoline composition. For this purpose, FCC catalysts with rare earth levels of 1, 3 and 5 wt% were selected. These catalysts were then deactivated over 10-20 hours in 100% steam to match catalyst activity, and then blended with various levels of equilibrated Super Z Excel (a leading, commercially available ZSM-5 additive). ZSM-5 is more stable under laboratory deactivation conditions than the REUSY containing FCC base catalyst. Because ZSM-5 zeolite has a higher silica alumina ratio (SAR) and smaller micropores than REUSY, it requires more severe deactivation. Deactivation conditions recommended for ZSM-5 additives are 815°C for 20 hours in 100% steam, a protocol that would be overly destructive for the zeolite-Y present in regular FCC base catalysts.

The study was carried out at 1060°F (570°C) to mimic petrochemical FCC operating conditions (equivalent to a riser outlet temperature of about 550°C) using a VGO feed.

Effects of ZSM-5 and base catalyst on gasoline composition
Previous studies have shown that adding small amounts of ZSM-5 additive to the base catalyst (around 5 wt%) results in a slight increase in gasoline aromatics as calculated on feed basis;^5,6 and when adding high amounts of ZSM-5 additive to an FCC base catalyst, gasoline aromatics as calculated on feed basis decrease slightly.^1,7,8 However, this study shows the effect is strongly dependent on the hydrogen transfer activity of the catalyst system. The higher the hydrogen transfer activity of the base catalyst, the more pronounced is the increase in gasoline aromatic cores on feed basis when a small amount of ZSM-5 is added (see Figure 1). However, when higher concentrations of ZSM-5 are blended with base catalyst, gasoline aromatic cores on feed basis remain constant compared to the base case without ZSM-5 for intermediate and high levels of rare earth on the base catalyst, and decrease for the low rare earth on base catalyst.