Metallic cobalt derived from cobalt nitrides for Fischer-Tropsch Synthesis

Padayachee, Veroushia

Metallic cobalt derived from cobalt nitrides for Fischer-Tropsch Synthesis

Master Thesis

2022

Abstract

The Fischer-Tropsch synthesis (FTS) is a catalytic surface polymerisation reaction that converts synthesis gas, a mixture of carbon monoxide (CO) and hydrogen, into long chain hydrocarbons, namely α-olefins and n-paraffins. The product selectivity of this reaction is dependent on the nano effects of the catalyst used in addition to the reaction conditions, making FTS a highly structure-sensitive reaction. The transition elements namely ruthenium, iron, cobalt and nickel are highly active catalysts in FTS, however cobalt will be the focus of this research. Cobalt (Co) has two common crystalline phases: hexagonal close-packed (hcp) and face-centred cubic (fcc). Both these phases are active under FTS conditions, however, theoretical and reactor testing studies have found the hcp Co phase to be more active as a result of the hcp crystal structure having more favourable facets and active sites available for CO dissociation than the fcc Co phase. The conventional route for synthesising hcp Co nanoparticles is through a reduction carburization-reduction (RCR) route or via the reduction of hexagonal cobalt oxide. The former route produces a catalyst that is predominantly hcp Co, however not all of the carbon is able to be removed, especially at lower second reduction end temperatures of the RCR process. The latter route produces a reduced catalyst with an intergrowth (which occurs when there is more than one way in which a close packed layer of atoms can be put together) and could be an indication of a mixture of hcp and fcc Co. As a result, the objective of this study was to produce pristine hcp Co without an intergrowth or unfavourable surface species and determine the selectivity of this pure phase under Low temperature FTS (LTFTS) conditions. To complete this, a novel technique to producing metallic cobalt catalysts in FTS was proposed and implemented: cobalt nitride decomposition. This novel route involved producing a cobalt nitride with the desired crystal phase (e.g. hexagonal cobalt nitride, Co3N) using ammonia, and decomposing this nitride in a hydrogen atmosphere to the metallic phase that retained that crystal structure (e.g. hcp Co). Using this method, a pure hcp Co phase of an applicable FT crystal size of 17 nm was synthesised. Interestingly, cobalt hydroxide (α-Co(OH)2), the precursor to the Co3N, also reduced to a pure hcp Co phase of 32 nm. Both hcp Co catalysts had no indication of an intergrowth, nor were there any nitrogen species detected on the surface of the catalyst (according to the PXRD and SEM-EDX data), indicating a successful decomposition or reduction. The fcc Co catalyst was derived from exposing cobalt oxide (Co3O4) to ammonia, which reduced the catalyst to pure fcc Co instead of nitriding the catalyst to the fcc cobalt nitride, Co4N. When Co3O4 was reduced in hydrogen, the final catalyst was a mixture of hcp and fcc Co, with a predominant fcc Co phase (90 %). Thus, the four catalysts studied in this research are the two fcc Co and two hcp Co catalysts derived from the four different reduction techniques: CAT 1: fcc Co – H, through a pure hydrogen reduction of Co3O4, CAT 2: fcc Co – NH, through an ammonia treatment followed by a hydrogen reduction of Co3O4, CAT 3: hcp Co – H, the direct reduction of α-Co(OH)2 and finally CAT 4: hcp Co – NH, through the thermal decomposition of Co3N. These catalysts were tested under low temperature Fischer-Tropsch conditions (T=220 °C, P=20 bar, H2: CO =2 and GHSV = 0.5 NL/ gcat hr-1 ). The FT results revealed that CAT 2: fcc Co – NH achieved the highest CO conversion of 31 %, followed by CAT 1: fcc Co – H at 23 %, CAT 3: hcp Co – H at 5 %, and CAT 4: hcp Co – NH at 3 %. CAT 2: fcc Co – NH achieved the highest α value of 0.69, with the largest selectivity to long chain hydrocarbons and the smallest selectivity to methane, possibly as a result of the high water formation at the relatively high CO conversion (Claeys & van Steen, 2004). CAT 1: fcc Co – H had an α value of 0.62 and was more selective to methane. This indicated CAT 1: fcc Co – H had more favourable active sites available for methanation instead of chain growth. The hcp Co catalysts appeared to deactivate quickly and thus have little to no activity for chain growth, and as a result the selectivity towards methane was high. The α value for CAT 3: hcp Co – H and CAT: hcp Co – NH was 0.64 and 0.54 respectively. CAT 3: hcp Co – H did have a comparable alpha value to CAT 1: fcc Co – H, however as a result of the large difference in their CO conversions, the yield for the product classes for CAT 1: fcc Co – H was much higher than that of the hcp Co catalyst. CAT 2: fcc Co – NH subsequently achieved the highest turnover frequency (TOF) at the end of time on stream (TOS). The PXRD of the spent catalyst showed that wax was formed on this catalyst, which supported the selectivity to the longer hydrocarbon chains. The major difference between the preparation methods between CAT 1: fcc Co – H and CAT 2: fcc Co – NH, is that the former catalyst used solely a hydrogen atmosphere, whilst the latter used an ammonia and then a hydrogen atmosphere for the reduction. The difference in the reduction atmospheres influenced the purity of the resulting crystal phase achieved (CAT 1: fcc Co – H: 90% fcc Co, CAT 2: fcc Co – NH: 100 % fcc Co), and as a result, the activity of the catalyst. When the spent catalysts were characterised to determine the source of the deactivation, an increase in the crystal size of the catalysts was noticed. The samples did appear to have sintered, but to a certain extent. The XRD patterns show only the active metallic phase present on the catalyst after 48 hours, with the absence of carbon species and oxidized cobalt. Surface nitrogen was not detected according to the SEM-EDX results of both the catalysts that required an ammonia step during their preparation (CAT 2: fcc Co – NH and CAT 4: hcp Co – NH). It was proposed that the deactivation of the catalysts were perhaps a result of the choice in the metal salt used during synthesis preparation. According to a report conducted by Rosynek & Polansky (1991), residual chlorine ions from the cobalt chloride salt (similar to the precursor salt used in this research) blocked a considerable amount of the reduced metallic catalyst during FT studies. The SEM-EDX conducted on the spent catalysts confirmed the presence of chlorine on all four catalysts. The results showed that the fcc Co catalysts derived from Co3O4 (CAT 1: fcc Co – H and CAT 2: fcc Co – NH) had a lower chlorine concentration than the hcp Co catalysts derived from α-Co(OH)2 (CAT 3: hcp Co – H and CAT 4: hcp – NH). The calcination step in synthesising Co3O4 helped remove some of the chlorine ions on the surface of the catalyst, similarly suggested by Panpranot et al. (2003) and Rosynek & Polansky (1991). The α-Co(OH)2 precursor was deliberately not calcined in air in order to maintain the stacking sequence and purity of the hcp phase, which was the main aim of this research. Cobalt nitrides as FT catalyst themselves have not been thoroughly investigated. However, the nitrogen content on the catalyst has been reported to present additional active sites that aids in the desorption of reactants and thus improving the activity of catalysts in CO methanation (Razzaq et al., 2015). Two cobalt nitrides, Co3N (hexagonal) and Co2N (orthorhombic), were successfully synthesised to be tested under LTFT conditions. However, when their stabilities were investigated using the in situ XRD under LTFT conditions, it was found that the nitrides were unstable. Neither catalyst was able to remain a nitride for more than 2 hours at 190 °C before hcp Co was detected. Decreasing the temperature would not have offered useful results as the activity would be extremely low at those conditions. As a result, the cobalt nitrides were not tested under LTFT.

Keywords

Chemical Engineering

Reference:

Collections

Masters

Full item page