Journal of the Chilean Chemical Society
versión On-line ISSN 0717-9707
J. Chil. Chem. Soc. v.55 n.1 Concepción 2010
J. Chil. Chem. Soc, 55, N° 1 (2010), pág: 35-38
HYDROCARBONS SYNTHESIS FROM A SIMULATED BIOSYNGAS FEED OVER FE/SIO2, CATALYSTS
P. UBILLA1, R. GARCÍA1, J.L.G. FIERRO2, N. ESCALONA1*.
1 Universidad de Concepción, Facultad de Ciencias Químicas, Casilla 160C, Concepción, Chile
2 Instituto de Catálisis y Petroleoquímica, CSIC, Cantoblanco, 28049, Madrid, Spain
The Fischer Tropsch reaction, using a simulated gas mixture similar to that obtained from biomass transformation (biosyngas), was studied over Fe/SiO2catalysts. The reaction was carried out in a stainless steel fixed bed reactor at 300 °C and 1 MPa. Excess solvent impregnation was used to prepare the catalysts. Nitrogen adsorption at -196 °C, thermal programmed reduction (TPR), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM) were used as characterization methods. The highest CO conversion was obtained for the catalyst with Fe-loading of 15wt% whereas it decreased at higher Fe loadings. The decrease ofthe activity is explained in terms ofthe drop ofthe active sites due to the formation of large Femetallic aggregates. The Fe/SiO2 catalysts presented a wide distribution of liquid hydrocarbons (C6 to C 18) and high methane formation. The liquid product distribution was related to the average size of iron particles. The increase of iron particle size favours the formation of hydrocarbons centred in C 11-14 chain length.
Keywords: Fischer Tropsch synthesis, biosyngas; Fe/SiO2 catalysts, biomass
The Fischer Tropsch synthesis (FTS) is the most important catalytic process in the synthesis of gasoline and/or diesel from syngas (CO/H2) mixture . While the first commercial FTS plants operated with syngas mixtures produced from coal gasification, modern FTS units use CO/H2 mixtures mainly obtained from the methane steam reforming, where the molar H2/CO ratio is around of 2 . The metals used as catalysts inthe FTS are mainly Fe and Co promoted by K, Re, Cu or Zn [2, 3] . On the other hand, the environmental regulations and the depletion of oil reserves have motivated the search of new technologiesforthe synthesis ofenvironmentally friendlyfuels. In this context, the production of synthetic hydrocarbons using the biomass gasification streams as source of CO and H2 appears highly attractive. The biomass can be gasified in a gas mixture consisting mainly in CO2, CO, CH4, H2 and N . This mixture is called "biosyngas" and the H2/CO ratio is close to 1. The hydrocarbons produced according to the FT synthesis using biosyngas mixtures are free of S and N, and in principie can be considered neutral in the CO2 greenhouse effect. The body of work related to hydrocarbon synthesis using syngas mixtures produced according to the steam methane reforming technology is enormous [1,2,3,4,5,6], however the literature describing the production of hydrocarbons from biosyngas is much more limited [7,8,9,10,11]. In the line Jun et al.  reported the synthesis of liquid hydrocarbons over Fe/Cu/Al/K (100/6/16/4) catalysts using biosyngas mixtures. These authors found that this catalyst produces high fuel yield and selectivity to olefins. Recently, Escalona et al.  have studied the synthesis of hydrocarbons from biosyngas over Co/ SiO2 catalysts promoted by Cu, Re, Ru and Cu . They found that the CO conversion over these catalysts is mainly determined by the dispersion of Co species. In addition, they observed that the higher metal particles result in a decrease of the production of long hydrocarbon chain (C14+) and promote the synthesis of smaller hydrocarbon chain lengths (C8-C9). Also, Re and Ru increases significantly both the activity and the selectivity to long chain hydrocarbons, whereas in those Cu- and Zn-promoted the effect is strongly affected by the amount of promoter. On the other hand, the Fe supported catalysts have not been reported in the FT synthesis using the biosyngas as feed. Therefore, the focus of this work was to study the effect of Re content in the hydrocarbon synthesis over Fe/SiO2 catalysts using the biosyngas as feed.
2.1 Catalyst preparation
Iron-based catalysts containing metal loading of 10-25 wt.% Fe were prepared by wet impregnation of commercial silica support (BASF Dll-11. SBET = 136 m2 g-1) with appropriate amounts of aqueous solution of Fe(NO3)2 (Aldrich, p.a.), in a rotary evaporator. After impregnation, the samples were dried at 110 °C for 12 h and then calcined at 450 °C for 5 h. Prior to characterization and testing, the catalysts were activated in flowing hydrogen at 500 °C for 12 h. The catalysts and the nomenclature used are listed in Table 1.
2.2 Catalyst characterization
The BET specific surface area (SBET) was evaluated from nitrogen adsorption isotherms recorded at -196 °C in an automatic Micromeritics system Model ASAP 2010. Temperature programmed reduction (TPR) studies were carried out in a conventional system having a thermal conductivity detector and a quartz cell. In each experiment 500 mg of the sample was used, with a heating rate of 10 °C min-1 and the temperature range from 25 to 1000 °C. The mixture used in the reduction experiments was a 5%H2/Ar with a flow of 50 cm3/min. TEM micrographs (methanol dispersion method) were obtained in a transmission electron microscopy (TEM) taken with a Jeol Model JEM-1200 EXII System. In order to obtain the Fe particle size, a histogram study, using software Image Tool 3.0 from TEM, using over 400 particles of each catalyst was carried out.
X-ray photoelectron spectra (XPS) was acquired with a Escalab 200R spectrometer equipped with a hemispherical electron analyzer and an A1Ka (hv = 1486.6 eV, 1 eV = 1.602 x 10-19 J) 120 W X-ray source. Prior to the analysis, the samples were reduced in situ in H2 flow at 500 °C for 12 h. AU binding energies (BE) were referenced to the Si 2p line at 103.4 eV. This reference gave BE values within an accuracy of ± 0.2 eV. Atomic ratios were calculated from the intensity ratios normalized by atomic sensitivity factors.
2.3 Catalytic reaction
Activity tests were carried out in a fixed bed stainless steel reactor. The reaction conditions were: space velocity (GHSV) of 1800 mi g-1 h-1, pressure of 1 MPa and reaction temperature of 300 °C. The feed was a representative mixture obtained from biomass gasification having H2/CO/CO2/CH4/N2 in the molar proportion 32/32/12/18/6, respectively . Prior to the reaction, the catalysts were reduced in situ at 500 °C during 12 h under hydrogen flow. Activity data were taken at steady-state conditions, approximately after 96 h on-stream. The analysis of products were performed by gas chromatography using a Perkin Elmer 3920B system provided with a thermal conductivity detector, using a n-octane Porasil-C column for liquids and a Carbosieve II column for the gaseous products.
RESULTS AND DISCUSSION
3.1. Catalyst characterization
The SBET and total pore volume (obtained at a relative pressure P/P° = 0.95) of calcined Fe/SiO2 catalysts are shown in Table 1. The SBET values decrease slightly with the Fe loading. This result suggests that Fe species were highly dispersed into the pores of the silica substrate and that pore blockage was negligible.
The Figure 1 shows the TEM micrograph recorded from the Fe(x)/SiO2catalyst. The EDX analysis showed the presence of Fe particles (not shown here), The Fe metal displays a black colour, such as is shown in the Figure 1. The histogram studies shown a broad particle size distribution and the máximum of each histogram is summarized in the Table 1. In general, the Fe/SiO2 catalysts present a broad particle size distribution upon increasing Fe-loading and the average particle was found to increase gradually with Fe loading from 29 to 40 nm (Table 1). This trend is similar to that found with Co/SiO2 catalysts . However, the average particle size obtained for Co/SiO2 is higher than for Fe/ SiO2 catalysts.
TPR profiles of the oxide precursors are shown in Figure 2. It can be observed that the reduction process of the catalysts occurs according the three characteristic steps of the reduction of Fe2O3 species. The first peak centred at 420 °C is related to the transformation of Fe2O3 to Fe3O4, the second stage centred to 615 °C represents the transformation of Fe3O4 to FeO and the third stage centred around 720 °C corresponds to the transformation of FeO to metallic Fe [13,14]. The intensity of the three peaks increases gradually with the Fe loading. Figure 2 also shows that the position of the máximum reduction of three peaks does not change significantly with increasing iron loading. This observation suggests that iron species are homogeneously dispersed on the surface of the silica substrate.
Photoelectron spectroscopy has been used to determine the chemical state and relative concentration of the reduced Fe/SiO2 catalysts. The binding energies of core-level electrons and the Fe/Si surface atomic ratios are compiled in Table 2. The Fe 2p core-level spectra (not shown) of all catalysts show the characteristic doublet (Fe 2p3/2-Fe 2p1/2) of iron species, whose splitting is about 13.6 eV. On looking at the most intense Fe2p3/2 component, a minor peak placed at 707.3 eV belonging to metallic iron is observed for all samples , whereas a major peak whose máximum intensity is located at 710.5 eV is usually assigned to iron oxide . As no satellite line is observed somewhere around 719.0 eV indicative of the presence of Fe3+ ions, it is inferred that the iron oxides responsible for the peak around 710.5 eV in the reduced catalysts comes from partially reduced iron oxides such as Fe3O4 (magnetite) species. The relative intensity of the two Fe 2p components (peaks at 707.3 and 710.5 eV) is also included in Table 2 (in parentheses). It can be seen, that the fraction of metallic iron determined on the surface region of these catalysts is much lower than the fraction of Fe oxides. In addition, the fraction of reduced iron to metallic state (peak at 707.3 eV) increases upon increasing the iron loading in the catalysts. This behaviour suggests that at low Fe content, the ionic iron species interact strongly with the SiO2 surface and therefore are difficult to reduce to the metal state under the experimental conditions employed in this work On the contrary, in the catalysts with higher Fe loadings a higher proportion of tridimensional iron oxide structures are developed and therefore they can be easily reduced to metal state.
In order to get an estimate of the extent of dispersion of the active phase over the silica surface, the Fe/Si atomic ratios were calculated (see Table 2). The variation of the Fe/Si atomic surface ratio as a function of Fe-loading of the Fe(x)/SiO2 catalysts is shown in Figure 3. Clearly, the Fe phase appears as ratherlargecrystallites(around 32 nm)upto 15 %of Fe. Thistrendis similar to that found with Co/SiO2 catalysts previously , where the activity increased almost linearly with increasing Co-loading, reaching the máximum at about 20 wt% Co and then levelled off The observed deviation from linearity above 15 % of Fe suggests the formation of large segregated crystalline particles (over 37 nm) mainly on the external surface of the silica particles. The higher Fe/Si ratios observed at higher Fe loading, especially for Fe(20)/SiO2 and Fe(25)/SiO2 catalysts, results likely from the presence of a high density of iron oxide particles and therefore keep less exposed silica surface. This is in good agreement with TEM results.
3.2 Catalytic activity
The conversion of CO over Fe(x)/SiO2 catalysts as a function of Fe-loading at 300 °C is shown in Figure 4. The activity, expressed as percent of CO conversion, increases linearly with increasing Fe-loading up to about 15 wt%, and then slightly decreases. Similar behaviour has been found for Co/ SiO2 catalysts, where the máximum activity is reached at 20 wt% Co-loading . The initial increase of CO conversion is associated with the increase inthe amount of well-dispersed Fe species. Such increase in performance for catalysts with Fe-loadings up to 15 wt% is consistent with the observation that Fe species are deposited relatively uniformly in both the external and internal surface of silica pores in the catalysts as it was revealed by the Fe/Si ratios derived from XPS analysis (cf Fig. 3). However, a different picture emerges for higher Fe-loadings. Above 15wt% Fe-loading, some segregation of iron species toward the external surface of SiO2 particles occurs; this phenomenon being more marked for the catalysts with the highest (25wt%) Fe-loading. Thus, for the 20 and 25 wt% Fe catalysts the lower CO conversion achieved can be related to segregation of an important fraction of the iron species from the pores of the substrate toward the external surface (cf. Fig. 3) where they develop larger iron oxide crystallites (Table 1). Indeed, the lowest CO conversion achieved on the catalyst containing 25 wt% Fe is achieved whose average crystallite size is the highest (over 37 nm) and for which a strong increase in the Fe/Si atomic ratio is observed.
The morphology of the iron phase in these catalysts plays a major role on their performance. This is illustrated by comparing the highest CO conversion of the Fe(15)/SiO2 catalyst, whose metal fraction in the reduced state is the lowest (14% metal by XPS), with the lowest CO conversion achieved on the Fe(15)/SiO2 catalyst for which the fraction of metal area increased up to 24%. Therefore, the absence of direct correlation between the fraction of Fe metal and CO conversion indícate that factors other than metal dispersion must be considered to explain catalytic performance. To explain this trend, it is suggested that the highest CO conversion achieved on the 15 wt% Fe catalyst is the result of two consecutive reactions:
Thus it is likely that the highly dispersed iron species within the pores of the substrate catalyse both the synthesis of hydrocarbons and WGS reactions with the subsequent increase in CO conversion. Notwithstanding CO2 selectivity data summarized in Table 3 indicate that CO2 proportion decreases markedly upon increasing Fe-loading. Thus, it can be inferred that reactivity is basically associated to the dispersed iron species within the pores of the substrate.
Table 3 shows the selectivity to CH4, CO2 and C5+ hydrocarbons for the Fe/SiO2 catalysts as a function of Fe-loading. These results indícate that Fe/ SiO2 catalysts present a high selectivity to CH4 which also increases with Fe-loading. However, these values are lower than those recorded for Co/SiO2catalysts (95 %) . On the other hand, the Fe(10)/SiO2 catalyst presents a high formation rate of CO2, which decreases with the increase of Fe content. In addition, the Fe/SiO2 catalysts show higher formation rates of CO2 than the Co/ SiO2 catalysts, reported previously . This behaviour is expected because it is well known that iron catalysts perform the WGS reaction. The participation of other CO2 forming reactions to CO2 selectivity cannot be precluded. For instance, a fraction of the CO2 formed might well arise from recombination of the oxygen fragment coming from CO dissociation with other CO molecules as proposed by Krishnamoorthy et al. .
The decrease of the CO2 formation as the increase in the CH4 formation with the Fe content in the Fe/SiO2 catalysts is not clear yet. Tentatively, this change in selectivity is associated to changes in morphology of iron species when increasing Fe-loading. i.e. increase in metal particle size and segregation of iron phase toward the particle surface.
Table 3 also shows the highest formation rate of CJ+ hydrocarbons on Fe(15)/SiO2 catalyst. In general, Fe/SiO2 catalysts record values of selectivity of C5+ much larger than on Co/SiO2 catalysts (around of 5 %) . This result suggests that Fe/SiO2 catalysts are more selective than Co/SiO2 ones for the production of liquid hydrocarbon from biosyngas. However, in the classic FT where the H2/CO ratio equals 2, Co/SiO2 catalysts are more active than Fe/SiO2counterparts . This different behaviour may result from the feed mixture employed. The feed mixture (biosyngas) employed here is deficient in H2 (H2/ CO ratio equal 1) and Fe/SiO2 catalyst may increase the formations of H2 "in situ" via WGS reaction and therefore may improve the selectivity to C5+.
Figure 5 shows the distribution of condensable products over Fe(x)/SiO2catalysts as function of Fe-loading. An increase in Fe-loading produces a shift in the distribution of condensable products. Thus, the C9-C10 production of hydrocarbon chain increases with a decrease of Fe-loading. Similarly, the production of C hydrocarbons increases with Fe-loading. This behaviour cannot be due to changes in the acidity of the catalysts upon increasing the reduction of Fe2O3 species with the raise of Fe content. Wan et al.  reported that surface basicity suppresses the chemisorption of CO and facilitates the chemisorption of H2, so it favours the production of low molecular weight products and enhances the hydrogenation capability. Therefore, the variation in the yield to hydrocarbon chain also may be related to metal particle size. The increase of iron particle size could favour the formation of hydrocarbons centred in C chain length. In general, this variation in the hydrocarbon selectivity on larger Fe particles can be attributed to the rate of secondary reactions such as olefin readsorption, as has been proposed earlier .
The results reported in this contribution show that Fe/SiO2 catalysts are active in the Fischer-Tropsch synthesis from a simulated biosyngas feed. The máximum conversion of CO is reached for metal loading of 15 wt%, and then decreases at higher metal loadings. The activity drop is associated to the loss of active sites by formation of large metallic Fe aggregates, mainly on the external surface of silica particles. The selectivity to methane increases with Fe-loading. The distribution of condensable products shifts lo longer-chain hydrocarbons with the increase of iron content. These behaviours may be due to changes in the distribution of average particle size of Fe. Fe/SiO2 catalysts with a large average iron particle size, favours the formation of longer chain hydrocarbons.
The authors thank CONICYT for the financial support (FONDECYT 1070548 grant).
 M. Dry, in Catalysis Science and Technology, J.R. Anderson and M. Boudart, Eds., Springer, Berlin, Vol I, Chapter 4, 1981. [ Links ]
 M.E. Dry, Catal. Today 71, 227 (2002). [ Links ]
 M. Ojeda, T. Herranz, F.J. Pérez-Alonso, J.M. Gonzalez-Carballo, S. Rojas, J.L.G. Fierro, in Syngas Production Methods, Post Treatments and Conversion" (A. Kurucz and I. Bencik, Eds.), Nova Science Publishers, Inc, Chapter 4, 2009, in press. [ Links ]
 S. Krishnamoorthy, A. Li, E. Iglesia, Catal. Letters, 80, 77, (2002). [ Links ]
 B. H. Davis, Catal. Today, 141, 25, (2009). [ Links ]
 M.E. Dry, Appl. Catal. A:Gen, 138, 319, (1996). [ Links ]
 K.-W. Jun, H.-S. Roh, K.-S. Kim, J.-S. Ryu, K.-W. Lee, Appl. Catal. A: Gen, 259, 221(2004) . [ Links ]
 N. Escalona, C. Medina, R. Garcia, P. Reyes, Catal. Today, 143, 76 (2009). [ Links ]
 C. Medina, R. García, P. Reyes, J. L. G. Fierro, N. Escalona, Appl. Catal. A: Gen 373, 71(2010). [ Links ]
 M.J.A. Tijmensen, A.P.C. Faaij, C.N. Hamelinck, M.R.M. van Hardeveld, Biomass Bioenergy, 23, 129 ( 2002). [ Links ]
 C.N. Hamelinck, A.P.C. Faaij, H. den Uil, H. Boerrigter, Energy 29, 1743 (2004). [ Links ]
 S.T. Chaudhari, S.K. Bej, N.N. Bakhshi, A.K. Dalai, Energy Fuels 15, 736(2001). [ Links ]
 H. Hayashi, L. Zhe Chen, T. Tago, M. Kishida, K. Wakabayashi, Appl. Catal. A: Gen, 231, 81 (2002) [ Links ]
 D.B. Bukur, C. Sivaraj, Appl. Catal A: Gen. 231, 201(2002). [ Links ]
 C.S. Kuivila, J.B. Butt, P.C. Stair, Appl. Surf. Sci. 32, 99 (1988). [ Links ]
 F. Tihay, G. Pourroy, M. Richard-Plouet, A.C. Roger, A. Kiennemann, Appl. Catal. A: Gen. 206, 29 (2001). [ Links ]
 S. Krishnamoorthy, A.W. Li, E. Iglesia, Catal. Lett. 80, 77 (2002). [ Links ]
 H. Wan, B. Wu, C. Zhang, B. Teng, Z. Tao, Y. Yang, Y. Zhu, H. Xiang, Y. Li. Fuel, 85, 1371 (2006). [ Links ]
 O. Borg, S. Eri, E.A. Blekkan, S. Storsarter, H. Wigum, E. Rytter, A. Holmen, J. Catal. 248, 89 (2007). [ Links ]
(Received: December 1, 2009 - Accepted: January 20, 2010)