Product yield in methanol conversion over ZSM-5 is predominantly independent of coke content

Abstract

The conversion of methanol to hydrocarbons (MTH) over ZSM-5 with time on stream and resulting product distribution were monitored at 350 °C in a series of experiments in which catalyst mass was varied while other conditions were held constant. This method enabled product analysis with respect to degree of deactivation: product distributions at low conversions due to small catalyst loadings could be compared to product distributions at the same conversion due to deactivation. It was found that product selectivity during methanol conversion over ZSM-5 is predominantly independent of coke content. A few products showed a slight sensitivity to deposited coke, and their individual changes suggested that 1,2,4-trimethylbenzene and 1,2,4,5-tetramethylbenzene are intermediates for ethene formation. Yield versus conversion plots further suggested that ethene is converted to higher alkenes by oligomerisation.

Introduction

The conversion of methanol to hydrocarbons (MTH) over zeolite based catalysts offers a potential route for the production of synthetic gasoline. The process was first reported by Chang and Silvestri [1], researchers at Mobil Central Research, in the 1970’s, and the discovery soon gained huge interest from both industry and academia. During the 1980’s the methanol-to-gasoline process, based on ZSM-5 catalysts, was implemented in full industrial scale in New Zealand, as gasoline production from natural gas became favourable due to high oil prices. The plant was, however, soon shut down as oil prices decreased [2]. Now, with the reoccurrence of high oil prices and the realisation that crude oil reserves are being depleted, upgrading of natural gas to products formerly derived from oil is receiving renewed commercial interest, including building of new plants for methanol conversion to higher hydrocarbons [3].

The hydrocarbons formed in the reaction consist of a mixture of mainly light olefins and aromatic compounds and some paraffins [1], [4], and result from a complex reaction network [3]. Briefly described, the MTH reaction is autocatalytic, with alkenes (mainly C₃₊) and arenes (mainly polymethylated benzenes) acting as main autocatalytic species. Autocatalytic species which reside in the catalyst are commonly referred to as the “hydrocarbon pool”. The zeolite catalyst provides the strong acid sites required for the reaction to occur, and product distribution is to a large degree determined by the topology of the zeolite structure, as large molecules can not be formed or escape from small-pore zeolites due to geometric constraints.

Besides the desired hydrocarbon products, coke deposits are generally formed on the catalyst’s inner and outer surface [4], and can be regarded as an undesired by-product of reaction. Internal coke deposits consist of molecules that are too large to diffuse out of the catalyst pores, while graphite-like coke may be deposited on the external surface of the catalyst. It is known that these coke deposits lead to deactivation of the zeolite catalyst; deactivation due to coke formation is reversible and activity can be restored by burning off the coke. For many years, the common belief has been that for ZSM-5, the coke deposits on the outer surface of the catalyst particles simply block the access to the active acid sites inside the zeolite channels. However, recently it has been pointed out that a small amount of hydrocarbons inside the zeolite channels may effectively reduce the catalytic activity [4], [5]. This observation points to a more intricate relation between coke formation and catalytic activity, which is not yet fully understood.

A way to quantify catalyst deactivation is to describe an effective rate of loss of active catalyst mass. Such a description seems reasonable, since the changes in product selectivity with time on stream during deactivation of a ZSM-5 catalyst showed a similar tendency to those observed with a continuously increasing space velocity [1], [6]. A model based on the assumption of a first order reaction for loss of active sites with time on stream was developed, and applied to relate the deactivation of a large number of ZSM-5 based catalysts with the presence of internal silanol groups inside the zeolite channels [7].

The presence of coke on or in the zeolite catalyst may also affect the product distribution in the effluent stream. Possible reasons for such coke-dependent variations are changes in transition-state shape selectivity or product shape selectivity. For instance partial pore-blocking may force products to remain inside a catalyst for a longer period of time, thus increasing the probability of secondary reactions. The presence of coke adjacent to an active site may also reduce the available space for a transition-state. Previous studies on SAPO-34, the main commercial system together with ZSM-5, have shown clear indications of coke induced selectivity changes [8], [9], [10]. It has been debated whether this is due to transition-state shape selectivity [8], [9], [10], [11] or product shape selectivity [12], [13]. Recently, conclusive evidence for product shape selectivity being dominant in SAPO-34 has been reported [14].

Information on the effect of coke deposits on the product selectivity can be obtained by a method developed by Wojciechowski [15], [16], [17], [18], which allows for the identification of stable and unstable products and discrimination between primary and secondary products at different degrees of deactivation. In this study the deactivation pattern of a ZSM-5 catalyst along with product distributions were studied at various contact times, following the method mentioned above. At high catalyst loadings, the initial conversion was 100%, followed by a decay of the conversion. With lower catalyst loadings, initial conversions less than 100% were obtained. It was thus possible to compare product yields at, for instance, 80% conversion with different degrees of coke/deactivation, which revealed the effect of the presence of coke on the product distribution. Together with a validation of the deactivation model presented by Janssens [6], this approach provides a deeper insight in the role of coke deposits on the catalytic properties of ZSM-5 based catalysts for the conversion of methanol to hydrocarbons.