Improved Co/Ti Catalysts By Direct Reduction of Cobalt Nitrate and the Incorporation of Promoters

Fischer-Tropsch synthesis (FTS) is a collection of a heterogenously catalyzed reactions which converts synthesis gas (SynGas) to transportation fuels such as diesel and jet fuel. This reaction is also used to make premium lubricants and waxes. The most common metals which are used for FTS are iron (Fe), cobalt (Co) and ruthenium (Ru). The surfaces of these metal (or metal carbide, in the case of Fe) nanoparticles are active for FTS, but Fe has higher intrinsic water gas shift activity compared to Co and Ru. Thus, Co and Ru are often preferred for converting natural gas derived syngas, which has a higher H₂/CO ratio. Commercially, Co is more desirable, because it is considerably less expensive than Ru.

These metals are often supported on a metal oxide carrier to increase their reactive surface area. The most common supports are titania (TiO₂), alumina (Al₂O₃), and silica (SiO₂). For typical air calcined catalysts, Co has been found to have a stronger interaction with Al₂O₃ in comparison to TiO₂ or SiO₂. This interaction allows H₂-activated Co metal nanoparticles to be smaller in size compared to those supported on TiO₂ or SiO₂. Titania has a number of attractive characteristics including moderate surface area, mesoporosity, and stability. It is therefore not surprising that Shell® company incorporates TiO₂ into their FTS catalyst formulation.

Air calcination is a typical standard pretreatment method to convert cobalt nitrate to cobalt oxide and remove NO_X gases, as well as H₂O. However, Co oxide species in calcined Co/SiO₂ catalysts were found to have a weak interaction with SiO₂ support, and they were found to agglomerate into large clusters (e.g., 50+ nm) upon reduction. Previous studies revealed that direct reduction of Co(NO₃)₂•xH₂O/SiO₂, either unpromoted or accompanied by a reduction promoter such as platinum (Pt), ruthenium (Ru), rhenium (Re), or and silver (Ag), resulted in smaller cobalt metal nanoparticles than those produced from the reduction of the respective air calcined catalysts. Reduction of the cobalt oxides generated from cobalt nitrate decomposition was facilitated by the addition of promoter, leading to higher active site densities. C₅+ selectivity and productivity in FTS reactor testing were improved as well.

Since using TiO₂ as a catalyst support is currently more practical than SiO₂ from an industrial standpoint, so this study revisits direct reduction of cobalt nitrate (Co(NO₃)₂), but in this case for unpromoted Co/TiO₂ as well as promoted catalysts incorporating reduction promoters such as Pt, Ru, Re, and Ag.

Catalyst preparation through the typical calcination method involves wet impregnation methods - e.g., incipient wetness impregnation (IWI) or slurry impregnation (SI) - of cobalt nitrate, followed by drying and air calcination in flowing air. Interestingly because of the vastly different degrees of interaction between different metal oxide supports and cobalt oxide species, activation of air calcined catalysts results in widely different average cobalt nanoparticle sizes.

In past studies, researchers have examined the effect of some promoters on Co/TiO₂. For example, the effects of boron (B), ruthenium (Ru) and rhenium were investigated in 2002. No effect of boron was observed on catalyst activity at lower space time but syngas conversion increased slightly at higher space time. While addition of ruthenium into Co/TiO₂ and CoB/TiO₂ enhanced both catalyst activity and selectivity, the study revealed that rhenium promoter exhibited the greatest effect on syngas conversion for Co/TiO₂ catalyst. Furthermore, the effect of water (in syngas) on the catalyst was investigated. While no significant change was observed for CO conversion at high space velocity (SV= 4~8 NL per g catalyst per h), the result showed that CO conversion was reduced at lower space velocity.

In 2007, the influence of support and promoter was investigated using temperature programmed reduction during EXAFS/XANES (TPR-EXAFS/XANES). The experiments illustrated that the nature of reduction of cobalt oxide and also demonstrated that the size of cobalt crystallites following activation depended on the strength of the interaction between the catalyst support and the cobalt oxide species prior to reduction, with silica offering a weak interaction and alumina providing a strong one. These experiments revealed a two step reduction process from Co₃O₄ to CoO and from CoO to Co⁰. Adding platinum (Pt) played an important role in enhancing the reducibility of cobalt oxide species, in agreement with prior studies.

The most important properties of the catalyst considered in the current investigation are stability, selectivity and carbon monoxide conversion per gram of catalyst (conversion rate). Cobalt particle size may plays an important role in tuning the CO conversion rate; while studies shows that catalysts having particle sizes between 10 and 210 nm have almost identical turnover frequency (TOF), reported results for particle sizes smaller than 10 nm are not consistent. For instance, applying a deposition-precipitation method resulted in a Co particle size range of 3-5 nm, with results consistent with a linear trend between activity and cobalt surface area per gram of catalyst. However, other researchers reported that TOF and selectivity diminished on cobalt/carbon nanofiber (CNF) catalysts when the cobalt particle size was decreased from 16 nm to 2.6 nm. Thus, regarding prior results from the literature, the question remains as to whether changing from a conventional air calcination prior to activation to a direct cobalt nitrate reduction technique with the application of reduction promoters will result in smaller cobalt nanoparticles, and whether this change in cobalt metal cluster size on TiO₂ has an effect on activity and C₅+ selectivity.

Regarding catalyst stability, previous studies have illustrated that heavily loaded Co catalysts display greater stability against the catalyst oxidation (i.e., and cobalt-support compound formation) in Co/Al₂O₃. EXAFS/XANES experiments on catalysts exposed to FTS conditions containing water showed that higher loaded cobalt catalysts (e.g., 25% Co) with Pt promoter exhibited greater resistance toward oxidation and cobalt-support compound formation compared to lower loaded cobalt catalysts (e.g., 15%Co, and especially those containing Pt promoter that facilitated reduction of smaller CoO species). In 2016, Hughes et al. performed EXAFS/XANES experiments on Co/TiO₂ catalysts and reported that while catalysts with small average nanoparticle size (2-3 nm) prepared by CVD of cobalt underwent oxidation under conditions mimicking 50% conveersion, catalysts having larger cobalt clusters (nanoparticles ≥ 10nm) underwent net reduction at the same conditions. Thermodynamic studies suggest that small cobalt nanoparticles (cobalt particles ≤ 4 nm) may re-oxidize under FT reactions conditions, and that this may be exacerbated by the nature of the support.

In this study we assessed the effect of direct reduction of cobalt nitrate versus conventional calcination/reduction and incorporate reduction promoters to produce different extents of reduction and different cobalt crystallite sizes (i.e., Co-Co coordination in EXAFS) for different catalysts. Measurements include TPR-EXAFS/XANES, hydrogen chemisorption with pulse reoxidation, BET surface area and porosity measurements, and results of catalyst testing using a CSTR. CO conversion for the catalysts after H₂ activation and testing for FTS at steady state (process conditions: T = 220^oC, P= 300 psi, H₂/CO= 2 mol/mol, SV = 3.8 slph per g_cat) revealed the following trend: 0.5%Pt-12%Co/TiO₂ uncalcined (X_CO = 56.4%) > 12%Co/TiO₂ uncalcined (X_CO = 49%) > 0.5%Pt-12%Co/TiO₂ calcined (46.6%) > 12%Co/TiO₂ calcined (X_CO = 29.8%). Methane selectivity was highest for the 12%Co/TiO₂ calcined catalyst (6.5%); the other catalysts were ~5%. TPR EXAFS/XANES results showed that while a typical two-step reduction occurs for calcined Co/TiO₂ catalysts, direct reduction of uncalcined catalysts proceeds through Co(NO₃)₂ decomposition to CoO_X, oxidation to a spinel resembling Co₃O₄, and then reduction of the spinel to more highly interacting CoO species (i.e., relative to the calcined case) prior to reduction to Co⁰. Moreover, adding the metal promoter facilitates reduction, especially for the CoO to Co⁰ step. EXAFS showed that the Co⁰ species produced from direct reduction of cobalt nitrate had significantly lower Co-Co metal coordination numbers in the first shell (6.4 - 7.4 relative to 10.2 for the calcined Pt-containing catalyst). The results suggest that direct reduction of nitrate leads to smaller, more strongly interacting Co oxides, and that their reduction is facilitated by incorporation of a promoter. These catalysts outperformed their calcined counterparts in CSTR reactor tests.