Catalysis 2022 issue

Data quality obtained in refinery catalyst testing

With high data quality (repeatability, reproducibility, and scalability) refiners can confidently determine themost efficient catalyst for their process units

TIAGO VILELA Avantium Catalysis

A ccurate evaluation of cat- alysts performance is cru- cial in optimising catalytic refinery processes with respect to product yields, run length, energy efficiency, and overall product qual - ity. Utilising high-throughput mul - tiple parallel reactors with excellent reactor-to-reactor repeatability is key to achieving the desired high data quality. 1 High data quality means that the test results are reproducible and thus reliable to evaluate the best performing catalysts. To guarantee high data quality in a parallel sys - tem, we need to obtain a good reac - tor-to-reactor repeatability where duplicate reactors within a run (loaded with the same catalyst sys - tem) yield the same results (within the experimental error margins). We also need to obtain good run-to- run reproducibility where the same catalyst system tested in different runs yields comparable results. As the catalyst packing in the reactors is straightforward (single-pellet string-reactor, or SPSR, loading) and does not require any special pro - cedures, we avoid typical packing issues like wall effect or channel - ling. This results in excellent reactor loading repeatability, which trans - lates into the high quality of the test results. Avantium developed small-scale parallel fixed bed reactor systems designed for catalyst intake of up to 1 ml, trade name Flowrence, to enhance catalyst development and selection for refinery applications. Flowrence high-throughput tech - nology with 16 reactors in parallel is extensively used to simulate refinery

catalytic processes, such as hydro - treating, hydrocracking, reforming, isomerisation, and dewaxing, over a wide range of process conditions and applications. These small-scale reactors are ideal in terms of heat transfer and hydrodynamics compared to larger reactors and therefore provide data that is reliable and intrinsically eas - ier to translate to an industrial scale. 2 It is intuitive to expect that larger reactors are less susceptible to size-related limitations. However, recent research by Moonen et al . shows that SPSRs are no more susceptible to wall effects, chan - nelling, and back mixing than properly utilised bench-scale reac - tors. Through an experimental pro - gramme, Moonen et a l. showed an excellent correspondence for gas oil hydrodesulphurisation (HDS) between an Avantium SPSR unit and a bench-scale unit with a cata - lyst volume of 225 ml – more than 300 times the volume of an SPSR. With rigorous modelling of the cor - responding hydrodynamics, they explained why results from the smaller unit are so similar to the larger one. 3 In addition, a recent paper for a lubricant base-oil hydrotreater showed that comparable results are obtained when using the Flowrence unit with 16 parallel SPSRs and a conventional pilot plant with a sin - gle-reactor pilot plant with a catalyst volume between 0.5L and 1L. For both catalyst systems, the relative average deviations were less than 1 wt% for HDS and HDN. 4 In the Flowrence unit, the schemes were evaluated in parallel – at two

different space velocities and in quadruplicate for increased accu - racy – while in the conventional unit the catalyst schemes were tested one at a time without replication. Due to excellent hydrodynamics of the SPSR and sophisticated process con - trol, Avantium’s unit achieves high repeatability, resulting in average deviations of less than 0.2 wt ppm for HDS and HDN across the quad - ruple reactors with the same loading scheme. 5 The engineering concepts of the Flowrence parallel small-scale reac - tor systems are discussed in the book chapter by van der Waal et al . This includes the influence of cat - alyst particle size, flow patterns, pressure drop, and temperature pro- files on the quality of catalytic per - formance results and is exemplified by multiple case studies on Fischer- Tropsch, oxidative coupling of methane, hydrocracking, and hydro - treating applications. Parallel testing allows for replica - tion – determining the statistical sig - nificance of results obtained – and simply evaluating more catalyst options simultaneously (side-by- side. In addition, smaller volumes will reduce the amount of feed required, avoiding the typical issues associated with obtaining large quantities such as handling, ship - ping, and storage (also for longer term availability of reference feed material). The catalyst testing data presented in this article were obtained in col - laboration with the major catalyst suppliers Albemarle, ART, Axens/ IFPEN, Haldor Topsoe, Shell, and UOP. All catalyst suppliers have

54 Catalysis 2022