Introduction
Present efforts in Fischer Tropsch (FT) technology are devoted to increasing the selectivity of certain product classes and slates over others. While the traditional goal is to produce long-chain hydrocarbons for transport fuel applications, more recent research is focused on favoring either olefins or oxygenates production. As to the latter, long-chain terminal alcohols are important feedstocks for plasticizers, detergents and lubricants. The large-scale production of these compounds is based on homogeneous hydroformylation of Cn (n≥3) petroleum-derived 1-alkenes. One of the issues of this process is noble metal recovery; another is the limited regioselectivity dictated by the Markovnikov rule and predicting considerable product fractions to be branched rather than straight-chained. With this background, it is highly desirable to design a heterogeneous “one-pot” process based on CO hydrogenation according to Fischer-Tropsch (FT), in the absence of a target 1-alkene. Due to the polymerization-type hydrocarbon growth mechanism in FT, predominantly terminal alcohols with little chain branching are obtained. We made use of this strategy in the recent past1-4 and provide here results demonstrating that the selectivity of heterogeneous CO hydrogenation can be tuned to long-chain terminal alcohols using alkali-promoted promoted CoMn catalysts4. In addition, we show that these catalysts are also active in long-chain aldehyde production. The systematic variation of the H2/CO partial pressure ratios actually allows optimizing the various product classes without changing the overall Anderson-Schulz-Flory (ASF) chain lengthening probability. Interestingly, a reaction-induced change in the catalyst chemical composition is observed upon such partial pressure changes. High yields of chain-lengthened aldehydes under hydrogen-deficient conditions seems to be associated with the transformation of metallic Co into Co2C.
Materials and Methods
Alkali-promoted CoMn catalysts were prepared using oxalate co-precipitation from metal nitrates in acetonic solution1, 4. Mixed-metal Co-Mn oxalate precursors are polymers with Metal Organic Framework (MOF) structure. Alkali precipitation was enforced by a solubility/entrainment effect since KNO3 is soluble in water, but not in acetone. Oxalate presursors were subjected to temperature-programmed decomposition (TPDec) in the presence of hydrogen. Metal (Co) vs. metal oxide (Mn5O8, MnO, K2O) formation was evaluated from the relative amounts of gaseous CO and CO2 released during the decomposition. Active catalysts contained nanosized metal particles and metal oxide phases acting as dispersant and promoter. Specific surface areas of catalysts were determined using the BET (Brunauer-Emmett-Teller) method. Values were found to be in the range between 35 and 80 m2g-1 depending on the composition of the catalysts. Metallic surface areas were determined from H2/D2 dynamic exchange measurements and were usually in the range of 2-5 m2/g.
Catalysts were characterized by in-situ X-Ray Diffraction (XRD) during the ongoing reaction. High resolution Transmission Electron Microscopy (HRTEM) and Electron Energy Loss Spectroscopy (EELS) were employed before and after reaction to receive information about catalysts’ morphology and chemical phase composition.
High-pressure catalytic tests were performed in a fixed-bed reactor consisting of a quartz tubule (Φinner = 7 mm) inserted into a stainless steel housing. Oxalate precursors were diluted with SiC, followed by in situ thermal decomposition in H2 at 370 °C for 1 h. Then the reactor was cooled to <100 °C before introducing syngas at a H2/CO ratio of 1.5. The temperatures for the catalytic tests were approached using low heating rates of 1 °C min-1. The CO conversion and product selectivities were measured by online by GCMS.
Results and Discussion
We show that the selectivity of the “one-pot” heterogeneous CO hydrogenation over alkali metal promoted CoMn catalysts can be tuned to receive various product classes, including paraffins, 1-alkenes, 1-alcohols and aldehydes. The relative atomic amounts of metal in these catalysts was systematically varied. Remarkably, low H2/CO partial pressures (≤1.5) enabled oxygenate selectivities between 50 and 60 wt% (ex CO2) could be obtained, thereof 70~97% (depending on the reaction temperature) as straight aldehydes. The best performance results in terms of oxygenates production were obtained by using small amounts of potassium as a promoter. Replacing K for Li shifted the product spectrum from mainly oxygenates to olefins.
Quite generally, while oxygenates production is favored by low temperatures (and high pressures) for thermodynamic reasons, high CO conversions can only be achieved at high temperatures. Thus, there is a selectivity-activity tradeoff that can only be mitigated by operating the synthesis at medium reaction temperatures. With this background, most of our studies were performed in the range between 220 and 240 °C where CO conversions up to 20% could be reached.
Strictly linear ASF distributions were obtained for C4+ products over ‘CoMnK’. Major deviations from the ASF behavior were obtained for C1-C3 products: very low selectivities for C1 (methanol+formaldehyde) and C3 oxygenates were contrasted by high acetaldehyde selectivities.
Detailed kinetic studies were performed with Co4Mn1K0.1 catalysts (indices indicating atomic ratios). Varying the H2/CO partial pressure ratio had a significant effect on the catalytic performance. According to Figure 1, while aldehydes are clearly dominating at low H2/CO ratios, they become less abundant when the H2/CO ratio increases. The alcohol fraction is strongly increasing at the expense of aldehydes and reaches a maximum at H2/CO=5. Paraffins increase from initially 20% and dominate for H2/CO>5 while olefins starting from nearly 20% disappear at such high partial pressure ratios. Turning to the ASF chain lengthening characteristics, it is seen that a unique α value of 0.5 for the total Cn production (resulting from paraffins, terminal olefins and alcohols as well as aldehydes) is obtained independent of the H2/CO pressure ratio. Because of the strong deviations of short-chain products from the general (linear) ASF behavior, only C4+ products are considered here.
An intriguing kinetic hysteresis effect was observed on cycling the partial pressure ratios from initially high (H2/CO= 9) to low (H2/CO=0.5) and back to high. Accordingly, the catalyst activity followed a clockwise hysteresis, i.e. overall low CO conversion when decreasing the H2/CO ratio from high to low and increasingly higher CO conversion on the way back. Closed hysteresis loops were obtained when subjecting the catalyst to very high H2/CO pressure ratios or pure H2. Fascinatingly, chain-lengthened paraffins and alcohols showed counterclockwise and, respectively, clockwise hysteresis. Selectivity hysteresis was also observed, though less pronounced, for aldehydes, olefins and carbon dioxide. Note that a 60% aldehyde production at low H2/CO pressure ratios was never before reported in the FT literature and makes the Co4Mn1K0.1 formulation a unique catalyst.
To correlate the observed hysteresis with structural and chemical changes of the catalyst during the ongoing reaction, we performed in-operando XRD studies and found a Co fcc phase at high H2/CO ratios (beginning of the loop) and a Co2C phase at low such ratios. This Co2C phase turned out to be rather reluctant when reducing it to metallic (hcp) cobalt in excess hydrogen. On the other hand, our results provide strong evidence that high aldehyde selectivities are favored by a Co2C phase while high paraffin selectivities prefer a metallic (either fcc or hcp) Co phase.
HRTEM and EELS studies with a Co4Mn1K0.1 catalyst provided further support for the formation of Co2C after reaction with low H2/CO ratios. Furthermore, a mixed-valence Mn2(II)Mn3(IV)O8 phase was identified. We therefore advocate a synergistic interaction between Mn5O8 oxide and bulk Co2C phases. Recent HRTEM studies focussed on the Mn-oxide phase and showed that part of the Mn5O8 underwent a reaction-induced decomposition forming highly dispersed MnO aggregates.
Conclusions
Potassium-promoted CoMn catalysts, in particular Co4Mn1K0.1, show unprecedented
catalytic performance in terms of straight-chain aldehyde production at low H2/CO partial pressures. Catalysts undergo a reaction-induced restructuring and the bulk composition for optimal aldehyde formation is Co2C/Mn5O8. Activity-selectivity hystereses are observed upon H2/CO variation. High selectivities in oxygenates formation at low H2/CO ratios may help design a ‘heterogeneous hydroformylation’ process in the absence of terminal olefins. To make this a viable option, further efforts are necessary to increase the activity of the catalysts.
Figure 1. Effect
of H2 to CO ratio. (A) Products
selectivity without CO2, (B) CO conversion and CO2
selectivity, (C) ASF behavior for
total Cn of the same catalyst showing a chain lengthening
probability independent of the H2/CO pressure ratio. Reaction
conditions: Co4Mn1K0.1
catalyst 0.5 g, 220°C
and 40 bar. References 1. Xiang, Y.; Chitry, V.; Liddicoat,
P.; Felfer, P.; Cairney, J.; Ringer, S.; Kruse, N., Long-Chain Terminal
Alcohols through Catalytic CO Hydrogenation. Journal of the American Chemical Society 2013, 135 (19),
7114-7117. 2. Xiang, Y. Z.; Barbosa, R.; Kruse,
N., Higher Alcohols through CO Hydrogenation over CoCu Catalysts: Influence of
Precursor Activation. ACS Catalysis 2014, 4 (8), 2792-2800. 3. Xiang, Y. Z.; Barbosa, R.; Li, X.
N.; Kruse, N., Ternary Cobalt-Copper-Niobium Catalysts for the Selective CO
Hydrogenation to Higher Alcohols. ACS
Catalysis 2015, 5 (5), 2929-2934. 4. Xiang, Y.; Kruse, N., Tuning the
catalytic CO hydrogenation to straight- and long-chain aldehydes/alcohols and
olefins/paraffins. Nature Communications 2016, 7, 13058.
See more of this Group/Topical: General Submissions