- Email: [email protected]
Studies on cobalt-based Fischer–Tropsch catalyst and characterization using SEM and XPS techniques
Applied Catalysis A: General 211 (2001) 203–211
Studies on cobalt-based Fischer–Tropsch catalyst and characterization using SEM and XPS techniques Bijay K. Sharma a,∗ , Mahendra P. Sharma a , Suresh Kumar a , Shyam K. Roy a , Sisir K. Roy a , Sundararajan Badrinarayanan b , Sudhakar R. Sainkar b , Anadarao B. Mandale b , Sadgopal K. Date b a
b
Central Fuel Research Institute, Dhanbad 828108, Bihar, India Special Instruments Lab, National Chemical Laboratory, Pune 411008, India
Received 29 May 2000; received in revised form 27 October 2000; accepted 30 October 2000
Abstract This paper reports the results of an experimental study involving Fischer–Tropsch (FT) synthesis on modified Co/SiO2 -Al2 O3 catalyst. The objectives are to increase the selectivity of C5 + liquid hydrocarbons (HC) and to decrease the formation of gaseous hydrocarbons during the reaction. The catalysts were characterized by X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and powder X-ray diffraction (XRD) techniques. The XPS results indicated the formation of a new surface species of cobalt complex on the surface of the catalyst. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Fischer–Tropsch synthesis; Catalyst modification; XPS and SEM studies
1. Introduction The catalytic conversion of coal into liquid and gaseous fuel by Fischer–Tropsch (FT) synthesis has been well known for the last 50 years. Iron and cobalt-based catalysts with promoters are most widely used for this reaction [1–9]. The major roles played by these metals or metal ions, in particular, cobalt, are (i) to hydrogenate the dissociating carbon species, (ii) to promote the chain growth and (iii) to restrict the de-activation by carbonaceous deposits. However, the efficiency of hydrogenation of carbon monoxide by cobalt ions is very much dependent upon their interactions with the support material, as well as upon the active species formed on the surface during the preparation and catalytic reaction. The formation of ∗ Corresponding author. Fax: +91-326-364350. E-mail address: [email protected] (B.K. Sharma).
active species on the surface of a cobalt-based catalyst depends on various processing parameters, such as method of impregnation, calcination temperature and time, pH of the precipitating medium, pre-treatment and the role of promoters [10–16]. Several research workers have also reported that the hydrocarbon formation in FT is not only dependent on support or on the cobalt dispersion, but also very much dependent on the amount of active site present in the catalyst system [2,3,11]. Different groups have reported the performance of Co/Al 2 O3 [17–19] and Co/SiO2 [20,21] catalysts in FT reaction. The catalytic activity was correlated with the physico-chemical properties that were obtained through SEM and X-ray photoelectron spectroscopy (XPS). In the Co/Al2 O3 system, XPS results suggested the interaction of cobalt with alumina (CoO/Al2 O3 ) on the surface. Also in case of cobalt on silica support, various complexes of CoO/SiO2
0926-860X/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 8 6 0 X ( 0 0 ) 0 0 8 6 0 - 7
204
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
species are formed on the surface of the catalysts as a reactive component. The design and development of catalysts, performance with respect to activity and selectivity has been reviewed by Adesina [22]. They suggested the application of multi-metallic catalyst as a promising substitute in FT synthesis. In a recent review, the mechanistic aspects of hydrogenation of carbon monoxide have been discussed [23] in detail for carbon bonding with respect to metal surfaces. The most accepted view implies that the tilted position of carbon monoxide configuration is a key one in CO dissociation responsible for hydrocarbon formation. However, in this model, the formation of a special surface species, which would affect the selectivity on the C5 + liquid hydrocarbon formation during the FT synthesis, is not well understood. In our recent publication [24], we have studied the performance of Co/SiO2 -Al2 O3 and determined the ideal conditions that led to better performance of the catalyst with respect to C5 + selectivity in FT reaction. We also reported a theoretical model, wherein the optimum selectivity with respect to C5 + liquid hydrocarbon has been discussed by a steady state modeling using artificial neural networks. However, this catalyst was found to produce some solid hydrocarbons and a higher percentage of methane. In the present communication, we report our study on the modification of the catalyst with a view to increase the yield of C5 + fraction, particularly C10 –C20 hydrocarbons (which is a major constituent of diesel fuel). The details of the experimental work on preparation of catalysts, catalytic activity as well as selectivity as regards to liquid hydrocarbon formation and also their physicochemical characterization have been discussed. It has been possible to identify the active surface species formed during the catalytic reaction on the SiO2 -Al2 O3 supported modified cobalt catalyst. 2. Experimental 2.1. Catalyst, apparatus and procedures The cobalt supported on SiO2 -Al2 O3 catalyst was prepared by co-precipitation and impregnation method, wherein Na, Mg, and thorium ions were doped as promoters, details were given in our earlier publication [24]. In the present study, this catalyst is referred to as Catalyst A (Cat A). The modification of
Table 1 Chemical composition of catalysta Samples
Cat A Cat B Heat-treated Cat A a
Chemical compositions (wt.%) SiO2
Al2 O3
Co
64.58 66.52 68.60
13.75 13.32 13.00
12.45 12.75 12.85
Rest → promoters.
the catalyst was carried out by heating Cat A above 1000◦ C for 4 h in a muffle furnace. It was then allowed to cool at ambient temperature. This heated mass was mixed with Cat A (Cat A/Cat A heated (3/2)), which is classified as modified catalyst (Cat B). The chemical compositions of the catalyst samples for the main constituent of support and the catalytic components, i.e. cobalt metal determined by conventional methods are shown in Table 1. The catalyst was calcined at 400◦ C for 4 h and packed in the Catatest reactor. The reduction of the catalyst was carried out at 450◦ C for 8 h before every run for in situ regeneration of the catalysts. The experimental runs were conducted in the Catatest unit made by Geomechanique, France. It consists of two units: basic unit and microprocessor-controlled unit. It has got a fixed bed reactor (ca. 140 ml) which can be operated at pressures up to 150 bar, temperatures up to 550◦ C and gas flows up to 300 l/h. The gas flow rate, pressure and temperature were maintained at desired levels by an electronic control system. The duration of each run was 50 h; the syngas feed with H2 :CO (v/v) ratio of 2:1 was employed. The gaseous and liquid products were analyzed by gas chromatography (GC). Different columns used in the product analysis were (1) permanent gases — combined column of Porapack-Q + MS 5 Å with TCD; (2) gaseous hydrocarbons — alumina with FID; (3) oxygenates and water — Porapack-Q with TCD; (4) hydrocarbons in organic liquid phase — SE-30 in FID. The overall percentages (v/v) of permanent gases and gaseous hydrocarbons were calculated by taking methane as an internal standard. 2.2. Catalyst characterization 2.2.1. Powder X-ray diffraction (XRD) XRD of these catalysts was done on a Rich Serfert MZ III diffractometer using Ni filter Cu K␣
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
radiation and scintillation counter detector with pulse height discriminator. 2.2.2. Surface area (SA) A Quantasorb SA analyzer using N2 as adsorbent gas at liquid nitrogen temperature was used to determine the SA of the catalysts by BET method. 2.2.3. SEM SEM studies were performed using a SEM, Leica Stereoscan 440 model manufactured by M/s. Leica Cambridge Ltd., UK. The powder samples were mounted on the standard specimen stubs with the help of double adhesive tape and silver paste. The samples were coated with a thin layer of gold in Polaran coating unit E-5000 to prevent the charging of the sample. The electron beam parameters were kept constant during the analysis of the entire sample. The micro-graph of the samples with 10 kV EHT and 25 pA beam current were recorded by a 35 mm camera attached on the high resolution recording unit. 2.2.4. XPS XPS measurements were carried out in VG. Scientific ESCA-3-MK-2 electron spectrometer fitted with an Al K␣ source. The anode was operated at 140 W (14 kV, 10 mA) power and the analyzer was operated
205
at a constant pass energy of 50 eV. All the spectra were recorded at similar spectrometer parameters. The details of the equipment, its calibration and other details are given in our earlier publication [25]. For the present setup, the full widths at half maximum (FWHM) for Au 4f7/2 is 1.7 eV. XPS peak intensities were calculated by integrating the peak areas after proper subtraction of the base line by linear background subtraction method. Peak synthesis were performed assuming that each peak in this study has a gaussian line shape; for all the samples, C 1s, Co 2p, Si 2p, Al 2p levels were recorded. We could not record any XPS peaks for the other additives, as their concentration is very negligible on the surface.
3. Results The results of Cat A under optimum reaction conditions, which led to maximum production of liquid hydrocarbons and other corresponding parameters in our earlier study, are given in Table 2. It was observed that on this catalyst higher amounts of methane and other gaseous hydrocarbon products were produced. It was also noted that some solid hydrocarbon formation occurred in the product. The details are given in our earlier publication [24]. To overcome these
Table 2 Experimental conditions, conversion levels and product analysis
Reactor conditions Pressure (kg/cm2 ) Space velocity (cc/cc cat/h) Reaction temperature (◦ C) % Conversion (v/v of CO + H2 ) Product analysis Hydrocarbonsa (total) Hydrocarbons (distribution)a CH4 C2 H6 C3 H8 C4 H10 C5 + -liquid hydrocarbons Oxygenates (including methanola ) CO2 a a b
Cat A
Cat B (run no.)
Optimum reaction condition
1
2
3
4
15 75 230 76.0
15 110 230 93.1
15 154 230 85.0
15 134 230 87.2
10 134 230 70.6
139.0
114.7
93.4
122.6
79.1
37.1 7.8 16.8 5.5 71.8 4.5 12.9
14.3 1.9 2.3 Neg.b 96.0 2.6 27.6
25.3 2.3 3.0 Neg. 62.8 2.9 11.6
18.6 2.0 3.0 Neg. 98.8 2.1 8.8
21.4 3.4 5.9 4.2 44.2 3.9 4.8
g/N m3 (gram per normal meter cube) of CO + H2 consumed. Neg.: present in negligible quantities.
206
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
Fig. 1. (a) SEM micrograph of unused Cat B. (b) SEM micrograph of used Cat B.
limitations, such as production of carbon dioxide, methane and other gaseous hydrocarbons (which are undesired products), and solid formation (high molecular weight hydrocarbons), we have modified the catalyst (Cat B) as described in the experimental portion. The Cat B (45 ml) was tested in the Catatest unit using syngas (H2 :CO = 2:1) as feed for the syn-
thesis of liquid hydrocarbons through modified FT synthesis. The results obtained on Cat B are also given in Table 2. The ideal conditions determined in our earlier study [24] were used as the starting point for the present study. It was observed that the CO2 formation and syngas conversion was very high. Hence, experiments were conducted at higher
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
space velocity, which resulted in a decrease in CO2 formation. The reaction conditions which led to the maximum production of liquid hydrocarbons C5 + fraction and lower production of CO2 , oxygenates and other gaseous hydrocarbons products were pressure −15 kg/cm2 , reaction temperature −230◦ C and space velocity 134 ml/ml cat/h. At such optimum conditions, the syngas conversion was 87.2% and total hydrocarbon formation was 122.6 g/N m3 of (CO + H2 ) consumed. The total hydrocarbon contained 98.8 g liquid hydrocarbons and 23.8 g C1 –C4 saturated hydrocarbons. Liquid hydrocarbons on analysis were found to contain 80% of C10 –C20 hydrocarbons, 5–7% of C20 –C30 hydrocarbons and the rest C6 –C9 hydrocarbons. Oxygenates formation and CO2 formation were 2.1 g and 8.8 g/N m3 of (CO + H2 ) consumed, respectively. Therefore, the modified Co/Al2 O3 -SiO2 catalyst significantly enhanced the formation of liquid hydrocarbons, i.e. C5 + hydrocarbon fractions, and also decreased methane, oxygenate and CO2 formations. The modified catalyst was further tested for its induction period at different intervals of time. The induction period (the time at which the liquid hydrocarbon formation starts) was found to be 6 h in this case. If we take this induction period into account, the formation of liquid hydrocarbons goes to 110.5 g/N m3 of (CO + H2 ) consumed at the optimum conditions of the synthesis. It was also found that the catalyst was not deactivated even after 1000 h of use Fig. 1. Surface areas of the catalyst were measured and are given in Table 3. It is evident from the SA data that
207
Table 3 Surface area of catalysts Catalysts
Surface area (m2 /g) BET
Cat Cat Cat Cat Cat
275.9 194.7 146.4 84.5 0.7
A A B B A
(unused) (used) (unused) (used) (heat treated in air above 1000◦ C)
in the case of modified catalyst, both the used and the unused conditions provided lower SAs than the unmodified ones (Cat A). 3.1. XRD The XRD patterns of heat-treated mass of Cat A (Fig. 2A), Cat B unused (Fig. 2B) and Cat B used (Fig. 2C) were found to be partially crystalline in nature. The X-ray diffraction pattern of heat-treated mass of Cat A sample indicates the presence of SiO2 in the form of crystobalite and two more phases, one corresponding to that of CoAl2 O4 and the other to that of Co2 SiO4 . The presence of Co2 O3 is also indicated in smaller percentage in the heat-treated Cat A sample. Crystobalite peaks were also noticed in the used and unused forms of Cat B sample; however, the peaks were less intense in Cat B (both used and unused form). From the XRD patterns in Fig. 2, the following changes in the peak intensity are observed:
Fig. 2. XRD pattern of (A) heat-treated Cat A, (B) Cat B unused, (C) Cat B used.
208
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
1. The peak position at 1.561 and 1.755◦ is decreasing in the order: heat-treated Cat A > Cat B unused > Cat B used. 1.755◦ is a characteristic peak for Co2 O3 in the present complex system. 2. The peak at 2.023◦ remains almost unchanged. However, some additional changes in the peaks in region 2.443–2.630◦ are noticed in the used Cat B. 2.023◦ peak is characteristic of Co3 O4 or Co2 Si O4 . 3. The peak at 2.852◦ is present in all the three systems; however, the broadening of the peak in the used Cat B reflects changes in particle size of Co3 O4 or Co2 O3 , or CoAl2 O4 . 3.2. SEM The SEM micrographs of fresh Cat B are shown in Fig. 1a. This showed inhomogeneous lump formations. Particle sizes were >2 m, whereas in the spent catalyst a morphological change was noticed along with aggregate formation, which led to 0.5–5 m sizes as shown in Fig. 1b. 3.3. XPS XPS were recorded on Cat A and Cat B (used and unused) samples. The data are presented in Tables 4 and 5. 1. In both the systems, no shift in the binding energy (BE) values of Al 2p level was observed. 2. The Si 2p level in both fresh and used Cat A resembles that of SiO2 . 3. In the used sample of Cat B, the Si 2p levels shift to lower BE values and the Si can be in the form of SiOx . Table 4 The binding energy values for various elements (± 0.2 eV) Atom
Al 2p Si 2p Co 2p3/2 C 1s
Cat A
Cat B
Fresh
Used
Fresh
Used
74.7 103.4 781.8 285.0
75.0 103.8 782.0 285.0 288.0
74.6 103.4 781.4 285.0
74.8 102.1 780.3 285.0 288.0
Table 5 Percentage composition (at.%) of the surface atoms Atom
Al Si C O Co
Cat A
Cat B
Fresh
Used
Fresh
Used
7.3 24.2 18.8 47.6 2.1
14.4 24.8 9.6 48.6 2.6
15.3 24.0 16.0 43.0 1.8
10.5 14.7 45.6 26.5 2.4
4. There is no change in the BE value of Co 2p3/2 for all the samples, except for the used Cat B sample (shifting to low BE side) which indicates the formation of a new Co-Al/Si bonding. 5. Though the Al concentration on the surface of Cat A after use show an increase, it was found to decrease in the case of used Cat B sample. In addition, the percentage composition of oxygen is also less in the used sample supporting our observation that the Si compound is deficient in oxygen. 6. The concentration of cobalt on the surface remains more or less the same in all the samples. However a small increase in concentration of Co is noticed in the case of the used catalyst of Cat B. 7. The used catalyst samples gives rise to two carbon species with BE values of 285 and 288 eV. 8. The former peak is attributed to hydrocarbon type and the latter can be assigned to a carbon–oxygen bonding.
4. Discussion The modification of the catalyst as described earlier has considerably influenced the selectivity towards the liquid hydrocarbon formation and the reduction in the formation of methane and other hydrocarbon gaseous products. It has also restricted the solid hydrocarbon and carbon dioxide formation. However, the total hydrocarbon formation, i.e. activity of the catalyst, was decreased. The water gas shift reaction affects the efficiency of the carbon monoxide utilization; hence the water gas shift reactivity of the catalysts determines the CO2 yield. The low water gas shift reactivity is one of the key objectives in the development of a superior catalyst in FT synthesis. The experimental results (Table 2) clearly indicate that, at the
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
optimum reaction conditions (Cat B) CO2 formation decreased. Therefore the modification of the catalyst helped in the decrease of water gas shift reactivity. The Boudouard reaction, i.e. 2CO → C + CO2 is the most favoured reaction for the coke formation. This not only deactivates the catalyst by covering the surface, but also destroys the mechanical properties of the catalyst structure. In the present case, no deactivation was observed even after 1000 h of experimental run. Thus, the deactivation of catalyst by carbon deposition was at a minimum. The enhanced life of the Cat B is very similar to that of Cat A and can be attributed to the addition of various promoters, such as oxides of alkali metals, magnesium and thorium. These promoters interact with the metal surface and facilitate the CO dissociation, which in turn reduces the carbon deposition. The role of promoter is well documented by various groups [10–16]. The values of percent carbon and hydrogen (after moisture correction) on the used form of (total reaction duration in different runs for both the system >1000 h) Cat A (C% 1.1; H% 0.77) and those of Cat B (C% 1.54; H% 0.25) determined by classical method indicate that the nature of carbon deposited on Cat B is different from that on Cat A. Similarly, higher value of surface carbon deposition is also observed from XPS result; however, no difference in the nature of surface carbon is indicated from the BE consideration. By deconvolution of C 1s peak in the XPS plots gave rise to the presence of two types of carbon species on the surface C1 at 285.0 eV and C2 at 288.0 eV. From the BE consideration, the low BE peak is assigned to the presence of hydrocarbon and the high BE peak to a carbon–oxygen bonding. The values of percentage carbon species on the surface of spent Cat A (C1 94.5%; C2 5.5%) and that of Cat B (C1 83.4%; C2 16.6%) determined from the area of these peaks also indicate that the formation of carbon oxygen bonding is higher on Cat B than that of Cat A. Loggenberg et al. [26] reported the formation of C2 Hx O complex and explained it as due to the interaction of CH2 and CO on a polycrystalline iron surface. A similar compound formation in the present study is very much possible due to a stabilization of surface intermediates on different surface sites of the catalysts during the synthesis. Our results are in good agreement with the mechanism proposed by Dry [5], which involves the formation of CH2 and CO as active surface inter-
209
mediates to explain the full product distribution in FT reactions. The data presented in Table 3 clearly indicate that the SA of the modified catalyst is much lower than that of the unmodified catalyst, both in fresh and in the spent catalysts. This decrease in the SA may help in creating the active sites on the catalyst surface as supported by available literature data [2,3,11]. The results on Cat B have clearly indicated that the decrease in the total hydrocarbons formation has been due to suppression in gaseous hydrocarbon yield in the product, i.e. C1 –C4 saturated hydrocarbons. It has also indicated remarkable increase in the C5 + liquid hydrocarbon yield, showing thereby the high selectivity towards higher hydrocarbons. The most interesting feature in our XPS study on Cat B is the shift in BE value of Co 2p3/2 and Si 2p to the lower side. The standard value of BE of Si 2p in SiO2 is 103.3 eV. The observed Si 2p BE value in used form of Cat B is 102.1 eV, which is lower than 1.2 eV, indicating charge transfer between interacting ions such as Si4+ and oxygen. XPS result further suggest the loss of oxygen and thereby the formation of a non-stoichiometric Si oxide (SiOx ). The used form of Co 2p3/2 BE values also shifts to lower BE side with respect to the fresh catalyst but still it is higher than metallic cobalt (Co 2p3/2 → 778.2). Though the binding energies for an element are characteristic to the chemical shift of that element in a systematic fashion as the oxidation state changes. However, in some cases the chemical shift depends also on other factors such as electron relaxation and extra inter-atomic forces, causing difference with the predicted values of chemical shift from zero order considerations. Especially in case of cobalt, the BE of Co-metal has the lowest BE for all cobalt chemical states. The binding energies of the Co 2p3/2 electrons increase from Co0 to Co2+ in a predictable manner. However, Co3 O4 has a lower BE than the CoO [27]. In the present study, the Co 2p3/2 matches neither with Co3 O4 nor with CoO; thus there is a possibility that on Cat B cobalt is interacting either with Al or Si oxides during the reaction. Chin and Hercules [19] studied the cobalt–alumina system in FT reaction and suggested the possibility of the formation of CoAl2 O4 at low concentration of cobalt. At higher concentrations, the formation of Co3 O4 is favoured. The BE value of used Cat B is very near to the matching BE values of CoAl2 O4 . However, our
210
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
XPS result does not show any change in the BE values of Al 2p and hence there is very little chance of cobalt aluminate formation during the FT reaction on Cat B. On silica-supported cobalt catalyst, Kent et al. [21] reported surface cobalt silicate in the FT reactions and suggested that the formation of an intermediate surface cobalt silicate under specific activation condition, maximizes the amount of reducible SA available for FT reactions. Experimental evidence from XRD patterns of Cat B (both used and unused) and also from XPS results, suggest the possibility of the presence of cobalt silicate intermediate on the surface in case Cat B is used. A second consequence is that Lewis acidity produced at the surface due to non-stoichiometry in SiO2 to SiOx makes the catalyst very selective towards liquid hydrocarbon formation and forming a non-stoichometric cobalt silicate, reducing thereby the activity of the catalyst for gaseous hydrocarbons. If cobalt replaces silicon isomorphously in the surface, crystal field stabilization energy will tend to preserve four-fold co-ordination of the cobalt, and this could occur by the transfer of oxygen from an adjacent SiO2 . The concentration of the cobalt migrating into the silica frame work depends on the cobalt dispersion and also on the number of available tetrahedral defects near the surface for incorporation of the Co2+ ions. Such situation may arise in the formation of a non-stoichometric cobalt silicate during the FT reaction on Cat B. This new compound or mixture of Co3 O4 and intermediate cobalt silicate on the surface may be the reason for the lower binding energies for Co 2p3/2 and Si 2p in XPS data. Yet another consequence which cannot be ruled out is that the presence of H2 and CO at elevated temperature and pressure of reaction may have reduced some of the cobalt state to the metallic state, which on cooling and after exposure to air produces mixture of Co3 O4 and intermediate cobalt silicate phases. Intermediate cobalt silicate could have also been formed during the reaction as surface intermediate, producing Lewis acidity on the surface. This finds support from XRD results, which show the absence of Co2 O3 in the used catalyst, and indicates the presence of new phases as reflected in the appearance of some new peaks. Formation of SiOx and its association with the active catalytic component and forming active surface intermediate cobalt silicate may be possible reasons of increased catalytic selectivity towards C5 +
liquid hydrocarbon formation during the reaction on Cat B.
5. Conclusions The modified catalysts considerably influenced the selectivity of C5 + liquid hydrocarbons and exhibit reasonable decrease in catalytic activity due to suppression of methane, other gaseous hydrocarbons and carbon dioxide formation compared to unmodified catalyst. The enhanced selectivity is attributed to a decrease in the SA, the loss of oxygen from SiO2 sites thereby creating reactive centers on the catalyst surface in the form of intermediate cobalt silicate.
Acknowledgements The authors express their sincere gratitude to the Department of Coal, Govt. of India, and CMPDIL, Ranchi, for financial help. We thank S. Mitra for recording the XRD of the samples. Further sincere thanks are due to the Director, CFRI, for kind permission to publish this paper. References [1] R.B. Anderson, The Fischer–Tropsch Synthesis, Academic Press, Orlando, 1984. [2] E. Iglesia, S.C. Reyes, R.J. Madon, S.L. Soled, Adv. Catal. 39 (1993) 221. [3] E. Iglesia, S.L. Soled, R.A. Fiato, J. Catal. 137 (1992) 212. [4] R. Qukaci, J.G. Goodwin, A. Singleton, Am. Chem. Soc. Div. Fuel Chem. 39 (1994) 1117. [5] M.E. Dry, in: G.J. Hutchings, M.S. Scurrell (Eds.), Synfuels, Catal. Today 6 (1990) 183. [6] S.H. Moon, Appl. Catal. 16 (1985) 289. [7] M. Blanchard, Appl. Catal. 9 (1984) 327. [8] J.G. Goodwin Jr., in: Proceedings of the Symposium on Methane Upgrading, ACS, Atlanta, 1991, p. 156. [9] L. Guezi, T. Hoffer, Z. Zsoldos, S. Zyode, G. Maire, F. Grain, J. Phys. Chem. 95 (1991) 802. [10] M.J. Ders, T. Shido, Y. Iwasawa, V. Ponec, J. Catal. 124 (1990) 350. [11] B.G. Johnson, C.H. Bartholomew, D.W. Goodman, J. Catal. 128 (1991) 231. [12] J. Ramirez, T. Klimera, Y. Huerta, J. Aracil, Appl. Catal. A 118 (1994) 73. [13] J. Venter, M. Kaminisky, G.L. Geoffrey, M.A. Vannice, J. Catal. 103 (1987) 450.
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211 [14] C.H. Bartholomew, in: L. Guezi (Ed.), Trends in Co Activation, Elsevier, Amsterdam, 1991 (Chapter 5). [15] H. Chen, A.A. Adesiner, Appl. Catal. A 112 (1994) 87. [16] G.V.D. Lec, V. Ponec, Catal. Rev. Sci. Eng. 29 (2/3) (1987) 183. [17] G. Dimitrova Penka, R. Mehandjiw Dimiter, J. Catal. 145 (1994) 356–363. [18] J. van de Loordrecht, M. van der Haar, A.M. van der Kraan, A.J. van Dillen, J.W. Geus, Appl. Catal. A 150 (1997) 365– 376. [19] R.L. Chin, D.M. Hercules, J. Phys. Chem. 86 (1982) 360. [20] H.M. Burce, G. Baker, Appl. Catal. A 123 (1995) 23–36. [21] E. Coutler Kent, A.G. Sault, J. Catal. 154 (1995) 56–64.
211
[22] A.A. Adesina, Appl. Catal. A 138 (1996) 345–367. [23] J.P. Hindermann, G.J. Hutchings, A. Kiennemann, Catal. Rev. Sci. Eng. 35 (1) (1993) 1–127. [24] B.K. Sharma, M.P. Sharma, S.K. Roy, S. Kumar, S.B. Tendulkar, S.S. Tambe, B.D. Kulkarni, Fuel 77 (1998) 1763– 1768. [25] S.B. Idage, S. Badrinarayanan, Langmuir 14 (1998) 2780. [26] P.M. Loggenberg, L. Carlton, R.G. Copperthwaite, G.J. Hutchings, J. Chem. Soc., Chem. Commun. (1987) 541. [27] D. Briggs, M.P. Seah, Practical Surface Analysis by Auger and X-ray Photoelectron Spectroscopy, Wiley, New York, 1984, p. 161.
Studies on cobalt-based Fischer–Tropsch catalyst and characterization using SEM and XPS techniques Bijay K. Sharma a,∗ , Mahendra P. Sharma a , Suresh Kumar a , Shyam K. Roy a , Sisir K. Roy a , Sundararajan Badrinarayanan b , Sudhakar R. Sainkar b , Anadarao B. Mandale b , Sadgopal K. Date b a
b
Central Fuel Research Institute, Dhanbad 828108, Bihar, India Special Instruments Lab, National Chemical Laboratory, Pune 411008, India
Received 29 May 2000; received in revised form 27 October 2000; accepted 30 October 2000
Abstract This paper reports the results of an experimental study involving Fischer–Tropsch (FT) synthesis on modified Co/SiO2 -Al2 O3 catalyst. The objectives are to increase the selectivity of C5 + liquid hydrocarbons (HC) and to decrease the formation of gaseous hydrocarbons during the reaction. The catalysts were characterized by X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and powder X-ray diffraction (XRD) techniques. The XPS results indicated the formation of a new surface species of cobalt complex on the surface of the catalyst. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Fischer–Tropsch synthesis; Catalyst modification; XPS and SEM studies
1. Introduction The catalytic conversion of coal into liquid and gaseous fuel by Fischer–Tropsch (FT) synthesis has been well known for the last 50 years. Iron and cobalt-based catalysts with promoters are most widely used for this reaction [1–9]. The major roles played by these metals or metal ions, in particular, cobalt, are (i) to hydrogenate the dissociating carbon species, (ii) to promote the chain growth and (iii) to restrict the de-activation by carbonaceous deposits. However, the efficiency of hydrogenation of carbon monoxide by cobalt ions is very much dependent upon their interactions with the support material, as well as upon the active species formed on the surface during the preparation and catalytic reaction. The formation of ∗ Corresponding author. Fax: +91-326-364350. E-mail address: [email protected] (B.K. Sharma).
active species on the surface of a cobalt-based catalyst depends on various processing parameters, such as method of impregnation, calcination temperature and time, pH of the precipitating medium, pre-treatment and the role of promoters [10–16]. Several research workers have also reported that the hydrocarbon formation in FT is not only dependent on support or on the cobalt dispersion, but also very much dependent on the amount of active site present in the catalyst system [2,3,11]. Different groups have reported the performance of Co/Al 2 O3 [17–19] and Co/SiO2 [20,21] catalysts in FT reaction. The catalytic activity was correlated with the physico-chemical properties that were obtained through SEM and X-ray photoelectron spectroscopy (XPS). In the Co/Al2 O3 system, XPS results suggested the interaction of cobalt with alumina (CoO/Al2 O3 ) on the surface. Also in case of cobalt on silica support, various complexes of CoO/SiO2
0926-860X/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 6 - 8 6 0 X ( 0 0 ) 0 0 8 6 0 - 7
204
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
species are formed on the surface of the catalysts as a reactive component. The design and development of catalysts, performance with respect to activity and selectivity has been reviewed by Adesina [22]. They suggested the application of multi-metallic catalyst as a promising substitute in FT synthesis. In a recent review, the mechanistic aspects of hydrogenation of carbon monoxide have been discussed [23] in detail for carbon bonding with respect to metal surfaces. The most accepted view implies that the tilted position of carbon monoxide configuration is a key one in CO dissociation responsible for hydrocarbon formation. However, in this model, the formation of a special surface species, which would affect the selectivity on the C5 + liquid hydrocarbon formation during the FT synthesis, is not well understood. In our recent publication [24], we have studied the performance of Co/SiO2 -Al2 O3 and determined the ideal conditions that led to better performance of the catalyst with respect to C5 + selectivity in FT reaction. We also reported a theoretical model, wherein the optimum selectivity with respect to C5 + liquid hydrocarbon has been discussed by a steady state modeling using artificial neural networks. However, this catalyst was found to produce some solid hydrocarbons and a higher percentage of methane. In the present communication, we report our study on the modification of the catalyst with a view to increase the yield of C5 + fraction, particularly C10 –C20 hydrocarbons (which is a major constituent of diesel fuel). The details of the experimental work on preparation of catalysts, catalytic activity as well as selectivity as regards to liquid hydrocarbon formation and also their physicochemical characterization have been discussed. It has been possible to identify the active surface species formed during the catalytic reaction on the SiO2 -Al2 O3 supported modified cobalt catalyst. 2. Experimental 2.1. Catalyst, apparatus and procedures The cobalt supported on SiO2 -Al2 O3 catalyst was prepared by co-precipitation and impregnation method, wherein Na, Mg, and thorium ions were doped as promoters, details were given in our earlier publication [24]. In the present study, this catalyst is referred to as Catalyst A (Cat A). The modification of
Table 1 Chemical composition of catalysta Samples
Cat A Cat B Heat-treated Cat A a
Chemical compositions (wt.%) SiO2
Al2 O3
Co
64.58 66.52 68.60
13.75 13.32 13.00
12.45 12.75 12.85
Rest → promoters.
the catalyst was carried out by heating Cat A above 1000◦ C for 4 h in a muffle furnace. It was then allowed to cool at ambient temperature. This heated mass was mixed with Cat A (Cat A/Cat A heated (3/2)), which is classified as modified catalyst (Cat B). The chemical compositions of the catalyst samples for the main constituent of support and the catalytic components, i.e. cobalt metal determined by conventional methods are shown in Table 1. The catalyst was calcined at 400◦ C for 4 h and packed in the Catatest reactor. The reduction of the catalyst was carried out at 450◦ C for 8 h before every run for in situ regeneration of the catalysts. The experimental runs were conducted in the Catatest unit made by Geomechanique, France. It consists of two units: basic unit and microprocessor-controlled unit. It has got a fixed bed reactor (ca. 140 ml) which can be operated at pressures up to 150 bar, temperatures up to 550◦ C and gas flows up to 300 l/h. The gas flow rate, pressure and temperature were maintained at desired levels by an electronic control system. The duration of each run was 50 h; the syngas feed with H2 :CO (v/v) ratio of 2:1 was employed. The gaseous and liquid products were analyzed by gas chromatography (GC). Different columns used in the product analysis were (1) permanent gases — combined column of Porapack-Q + MS 5 Å with TCD; (2) gaseous hydrocarbons — alumina with FID; (3) oxygenates and water — Porapack-Q with TCD; (4) hydrocarbons in organic liquid phase — SE-30 in FID. The overall percentages (v/v) of permanent gases and gaseous hydrocarbons were calculated by taking methane as an internal standard. 2.2. Catalyst characterization 2.2.1. Powder X-ray diffraction (XRD) XRD of these catalysts was done on a Rich Serfert MZ III diffractometer using Ni filter Cu K␣
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
radiation and scintillation counter detector with pulse height discriminator. 2.2.2. Surface area (SA) A Quantasorb SA analyzer using N2 as adsorbent gas at liquid nitrogen temperature was used to determine the SA of the catalysts by BET method. 2.2.3. SEM SEM studies were performed using a SEM, Leica Stereoscan 440 model manufactured by M/s. Leica Cambridge Ltd., UK. The powder samples were mounted on the standard specimen stubs with the help of double adhesive tape and silver paste. The samples were coated with a thin layer of gold in Polaran coating unit E-5000 to prevent the charging of the sample. The electron beam parameters were kept constant during the analysis of the entire sample. The micro-graph of the samples with 10 kV EHT and 25 pA beam current were recorded by a 35 mm camera attached on the high resolution recording unit. 2.2.4. XPS XPS measurements were carried out in VG. Scientific ESCA-3-MK-2 electron spectrometer fitted with an Al K␣ source. The anode was operated at 140 W (14 kV, 10 mA) power and the analyzer was operated
205
at a constant pass energy of 50 eV. All the spectra were recorded at similar spectrometer parameters. The details of the equipment, its calibration and other details are given in our earlier publication [25]. For the present setup, the full widths at half maximum (FWHM) for Au 4f7/2 is 1.7 eV. XPS peak intensities were calculated by integrating the peak areas after proper subtraction of the base line by linear background subtraction method. Peak synthesis were performed assuming that each peak in this study has a gaussian line shape; for all the samples, C 1s, Co 2p, Si 2p, Al 2p levels were recorded. We could not record any XPS peaks for the other additives, as their concentration is very negligible on the surface.
3. Results The results of Cat A under optimum reaction conditions, which led to maximum production of liquid hydrocarbons and other corresponding parameters in our earlier study, are given in Table 2. It was observed that on this catalyst higher amounts of methane and other gaseous hydrocarbon products were produced. It was also noted that some solid hydrocarbon formation occurred in the product. The details are given in our earlier publication [24]. To overcome these
Table 2 Experimental conditions, conversion levels and product analysis
Reactor conditions Pressure (kg/cm2 ) Space velocity (cc/cc cat/h) Reaction temperature (◦ C) % Conversion (v/v of CO + H2 ) Product analysis Hydrocarbonsa (total) Hydrocarbons (distribution)a CH4 C2 H6 C3 H8 C4 H10 C5 + -liquid hydrocarbons Oxygenates (including methanola ) CO2 a a b
Cat A
Cat B (run no.)
Optimum reaction condition
1
2
3
4
15 75 230 76.0
15 110 230 93.1
15 154 230 85.0
15 134 230 87.2
10 134 230 70.6
139.0
114.7
93.4
122.6
79.1
37.1 7.8 16.8 5.5 71.8 4.5 12.9
14.3 1.9 2.3 Neg.b 96.0 2.6 27.6
25.3 2.3 3.0 Neg. 62.8 2.9 11.6
18.6 2.0 3.0 Neg. 98.8 2.1 8.8
21.4 3.4 5.9 4.2 44.2 3.9 4.8
g/N m3 (gram per normal meter cube) of CO + H2 consumed. Neg.: present in negligible quantities.
206
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
Fig. 1. (a) SEM micrograph of unused Cat B. (b) SEM micrograph of used Cat B.
limitations, such as production of carbon dioxide, methane and other gaseous hydrocarbons (which are undesired products), and solid formation (high molecular weight hydrocarbons), we have modified the catalyst (Cat B) as described in the experimental portion. The Cat B (45 ml) was tested in the Catatest unit using syngas (H2 :CO = 2:1) as feed for the syn-
thesis of liquid hydrocarbons through modified FT synthesis. The results obtained on Cat B are also given in Table 2. The ideal conditions determined in our earlier study [24] were used as the starting point for the present study. It was observed that the CO2 formation and syngas conversion was very high. Hence, experiments were conducted at higher
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
space velocity, which resulted in a decrease in CO2 formation. The reaction conditions which led to the maximum production of liquid hydrocarbons C5 + fraction and lower production of CO2 , oxygenates and other gaseous hydrocarbons products were pressure −15 kg/cm2 , reaction temperature −230◦ C and space velocity 134 ml/ml cat/h. At such optimum conditions, the syngas conversion was 87.2% and total hydrocarbon formation was 122.6 g/N m3 of (CO + H2 ) consumed. The total hydrocarbon contained 98.8 g liquid hydrocarbons and 23.8 g C1 –C4 saturated hydrocarbons. Liquid hydrocarbons on analysis were found to contain 80% of C10 –C20 hydrocarbons, 5–7% of C20 –C30 hydrocarbons and the rest C6 –C9 hydrocarbons. Oxygenates formation and CO2 formation were 2.1 g and 8.8 g/N m3 of (CO + H2 ) consumed, respectively. Therefore, the modified Co/Al2 O3 -SiO2 catalyst significantly enhanced the formation of liquid hydrocarbons, i.e. C5 + hydrocarbon fractions, and also decreased methane, oxygenate and CO2 formations. The modified catalyst was further tested for its induction period at different intervals of time. The induction period (the time at which the liquid hydrocarbon formation starts) was found to be 6 h in this case. If we take this induction period into account, the formation of liquid hydrocarbons goes to 110.5 g/N m3 of (CO + H2 ) consumed at the optimum conditions of the synthesis. It was also found that the catalyst was not deactivated even after 1000 h of use Fig. 1. Surface areas of the catalyst were measured and are given in Table 3. It is evident from the SA data that
207
Table 3 Surface area of catalysts Catalysts
Surface area (m2 /g) BET
Cat Cat Cat Cat Cat
275.9 194.7 146.4 84.5 0.7
A A B B A
(unused) (used) (unused) (used) (heat treated in air above 1000◦ C)
in the case of modified catalyst, both the used and the unused conditions provided lower SAs than the unmodified ones (Cat A). 3.1. XRD The XRD patterns of heat-treated mass of Cat A (Fig. 2A), Cat B unused (Fig. 2B) and Cat B used (Fig. 2C) were found to be partially crystalline in nature. The X-ray diffraction pattern of heat-treated mass of Cat A sample indicates the presence of SiO2 in the form of crystobalite and two more phases, one corresponding to that of CoAl2 O4 and the other to that of Co2 SiO4 . The presence of Co2 O3 is also indicated in smaller percentage in the heat-treated Cat A sample. Crystobalite peaks were also noticed in the used and unused forms of Cat B sample; however, the peaks were less intense in Cat B (both used and unused form). From the XRD patterns in Fig. 2, the following changes in the peak intensity are observed:
Fig. 2. XRD pattern of (A) heat-treated Cat A, (B) Cat B unused, (C) Cat B used.
208
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
1. The peak position at 1.561 and 1.755◦ is decreasing in the order: heat-treated Cat A > Cat B unused > Cat B used. 1.755◦ is a characteristic peak for Co2 O3 in the present complex system. 2. The peak at 2.023◦ remains almost unchanged. However, some additional changes in the peaks in region 2.443–2.630◦ are noticed in the used Cat B. 2.023◦ peak is characteristic of Co3 O4 or Co2 Si O4 . 3. The peak at 2.852◦ is present in all the three systems; however, the broadening of the peak in the used Cat B reflects changes in particle size of Co3 O4 or Co2 O3 , or CoAl2 O4 . 3.2. SEM The SEM micrographs of fresh Cat B are shown in Fig. 1a. This showed inhomogeneous lump formations. Particle sizes were >2 m, whereas in the spent catalyst a morphological change was noticed along with aggregate formation, which led to 0.5–5 m sizes as shown in Fig. 1b. 3.3. XPS XPS were recorded on Cat A and Cat B (used and unused) samples. The data are presented in Tables 4 and 5. 1. In both the systems, no shift in the binding energy (BE) values of Al 2p level was observed. 2. The Si 2p level in both fresh and used Cat A resembles that of SiO2 . 3. In the used sample of Cat B, the Si 2p levels shift to lower BE values and the Si can be in the form of SiOx . Table 4 The binding energy values for various elements (± 0.2 eV) Atom
Al 2p Si 2p Co 2p3/2 C 1s
Cat A
Cat B
Fresh
Used
Fresh
Used
74.7 103.4 781.8 285.0
75.0 103.8 782.0 285.0 288.0
74.6 103.4 781.4 285.0
74.8 102.1 780.3 285.0 288.0
Table 5 Percentage composition (at.%) of the surface atoms Atom
Al Si C O Co
Cat A
Cat B
Fresh
Used
Fresh
Used
7.3 24.2 18.8 47.6 2.1
14.4 24.8 9.6 48.6 2.6
15.3 24.0 16.0 43.0 1.8
10.5 14.7 45.6 26.5 2.4
4. There is no change in the BE value of Co 2p3/2 for all the samples, except for the used Cat B sample (shifting to low BE side) which indicates the formation of a new Co-Al/Si bonding. 5. Though the Al concentration on the surface of Cat A after use show an increase, it was found to decrease in the case of used Cat B sample. In addition, the percentage composition of oxygen is also less in the used sample supporting our observation that the Si compound is deficient in oxygen. 6. The concentration of cobalt on the surface remains more or less the same in all the samples. However a small increase in concentration of Co is noticed in the case of the used catalyst of Cat B. 7. The used catalyst samples gives rise to two carbon species with BE values of 285 and 288 eV. 8. The former peak is attributed to hydrocarbon type and the latter can be assigned to a carbon–oxygen bonding.
4. Discussion The modification of the catalyst as described earlier has considerably influenced the selectivity towards the liquid hydrocarbon formation and the reduction in the formation of methane and other hydrocarbon gaseous products. It has also restricted the solid hydrocarbon and carbon dioxide formation. However, the total hydrocarbon formation, i.e. activity of the catalyst, was decreased. The water gas shift reaction affects the efficiency of the carbon monoxide utilization; hence the water gas shift reactivity of the catalysts determines the CO2 yield. The low water gas shift reactivity is one of the key objectives in the development of a superior catalyst in FT synthesis. The experimental results (Table 2) clearly indicate that, at the
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
optimum reaction conditions (Cat B) CO2 formation decreased. Therefore the modification of the catalyst helped in the decrease of water gas shift reactivity. The Boudouard reaction, i.e. 2CO → C + CO2 is the most favoured reaction for the coke formation. This not only deactivates the catalyst by covering the surface, but also destroys the mechanical properties of the catalyst structure. In the present case, no deactivation was observed even after 1000 h of experimental run. Thus, the deactivation of catalyst by carbon deposition was at a minimum. The enhanced life of the Cat B is very similar to that of Cat A and can be attributed to the addition of various promoters, such as oxides of alkali metals, magnesium and thorium. These promoters interact with the metal surface and facilitate the CO dissociation, which in turn reduces the carbon deposition. The role of promoter is well documented by various groups [10–16]. The values of percent carbon and hydrogen (after moisture correction) on the used form of (total reaction duration in different runs for both the system >1000 h) Cat A (C% 1.1; H% 0.77) and those of Cat B (C% 1.54; H% 0.25) determined by classical method indicate that the nature of carbon deposited on Cat B is different from that on Cat A. Similarly, higher value of surface carbon deposition is also observed from XPS result; however, no difference in the nature of surface carbon is indicated from the BE consideration. By deconvolution of C 1s peak in the XPS plots gave rise to the presence of two types of carbon species on the surface C1 at 285.0 eV and C2 at 288.0 eV. From the BE consideration, the low BE peak is assigned to the presence of hydrocarbon and the high BE peak to a carbon–oxygen bonding. The values of percentage carbon species on the surface of spent Cat A (C1 94.5%; C2 5.5%) and that of Cat B (C1 83.4%; C2 16.6%) determined from the area of these peaks also indicate that the formation of carbon oxygen bonding is higher on Cat B than that of Cat A. Loggenberg et al. [26] reported the formation of C2 Hx O complex and explained it as due to the interaction of CH2 and CO on a polycrystalline iron surface. A similar compound formation in the present study is very much possible due to a stabilization of surface intermediates on different surface sites of the catalysts during the synthesis. Our results are in good agreement with the mechanism proposed by Dry [5], which involves the formation of CH2 and CO as active surface inter-
209
mediates to explain the full product distribution in FT reactions. The data presented in Table 3 clearly indicate that the SA of the modified catalyst is much lower than that of the unmodified catalyst, both in fresh and in the spent catalysts. This decrease in the SA may help in creating the active sites on the catalyst surface as supported by available literature data [2,3,11]. The results on Cat B have clearly indicated that the decrease in the total hydrocarbons formation has been due to suppression in gaseous hydrocarbon yield in the product, i.e. C1 –C4 saturated hydrocarbons. It has also indicated remarkable increase in the C5 + liquid hydrocarbon yield, showing thereby the high selectivity towards higher hydrocarbons. The most interesting feature in our XPS study on Cat B is the shift in BE value of Co 2p3/2 and Si 2p to the lower side. The standard value of BE of Si 2p in SiO2 is 103.3 eV. The observed Si 2p BE value in used form of Cat B is 102.1 eV, which is lower than 1.2 eV, indicating charge transfer between interacting ions such as Si4+ and oxygen. XPS result further suggest the loss of oxygen and thereby the formation of a non-stoichiometric Si oxide (SiOx ). The used form of Co 2p3/2 BE values also shifts to lower BE side with respect to the fresh catalyst but still it is higher than metallic cobalt (Co 2p3/2 → 778.2). Though the binding energies for an element are characteristic to the chemical shift of that element in a systematic fashion as the oxidation state changes. However, in some cases the chemical shift depends also on other factors such as electron relaxation and extra inter-atomic forces, causing difference with the predicted values of chemical shift from zero order considerations. Especially in case of cobalt, the BE of Co-metal has the lowest BE for all cobalt chemical states. The binding energies of the Co 2p3/2 electrons increase from Co0 to Co2+ in a predictable manner. However, Co3 O4 has a lower BE than the CoO [27]. In the present study, the Co 2p3/2 matches neither with Co3 O4 nor with CoO; thus there is a possibility that on Cat B cobalt is interacting either with Al or Si oxides during the reaction. Chin and Hercules [19] studied the cobalt–alumina system in FT reaction and suggested the possibility of the formation of CoAl2 O4 at low concentration of cobalt. At higher concentrations, the formation of Co3 O4 is favoured. The BE value of used Cat B is very near to the matching BE values of CoAl2 O4 . However, our
210
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211
XPS result does not show any change in the BE values of Al 2p and hence there is very little chance of cobalt aluminate formation during the FT reaction on Cat B. On silica-supported cobalt catalyst, Kent et al. [21] reported surface cobalt silicate in the FT reactions and suggested that the formation of an intermediate surface cobalt silicate under specific activation condition, maximizes the amount of reducible SA available for FT reactions. Experimental evidence from XRD patterns of Cat B (both used and unused) and also from XPS results, suggest the possibility of the presence of cobalt silicate intermediate on the surface in case Cat B is used. A second consequence is that Lewis acidity produced at the surface due to non-stoichiometry in SiO2 to SiOx makes the catalyst very selective towards liquid hydrocarbon formation and forming a non-stoichometric cobalt silicate, reducing thereby the activity of the catalyst for gaseous hydrocarbons. If cobalt replaces silicon isomorphously in the surface, crystal field stabilization energy will tend to preserve four-fold co-ordination of the cobalt, and this could occur by the transfer of oxygen from an adjacent SiO2 . The concentration of the cobalt migrating into the silica frame work depends on the cobalt dispersion and also on the number of available tetrahedral defects near the surface for incorporation of the Co2+ ions. Such situation may arise in the formation of a non-stoichometric cobalt silicate during the FT reaction on Cat B. This new compound or mixture of Co3 O4 and intermediate cobalt silicate on the surface may be the reason for the lower binding energies for Co 2p3/2 and Si 2p in XPS data. Yet another consequence which cannot be ruled out is that the presence of H2 and CO at elevated temperature and pressure of reaction may have reduced some of the cobalt state to the metallic state, which on cooling and after exposure to air produces mixture of Co3 O4 and intermediate cobalt silicate phases. Intermediate cobalt silicate could have also been formed during the reaction as surface intermediate, producing Lewis acidity on the surface. This finds support from XRD results, which show the absence of Co2 O3 in the used catalyst, and indicates the presence of new phases as reflected in the appearance of some new peaks. Formation of SiOx and its association with the active catalytic component and forming active surface intermediate cobalt silicate may be possible reasons of increased catalytic selectivity towards C5 +
liquid hydrocarbon formation during the reaction on Cat B.
5. Conclusions The modified catalysts considerably influenced the selectivity of C5 + liquid hydrocarbons and exhibit reasonable decrease in catalytic activity due to suppression of methane, other gaseous hydrocarbons and carbon dioxide formation compared to unmodified catalyst. The enhanced selectivity is attributed to a decrease in the SA, the loss of oxygen from SiO2 sites thereby creating reactive centers on the catalyst surface in the form of intermediate cobalt silicate.
Acknowledgements The authors express their sincere gratitude to the Department of Coal, Govt. of India, and CMPDIL, Ranchi, for financial help. We thank S. Mitra for recording the XRD of the samples. Further sincere thanks are due to the Director, CFRI, for kind permission to publish this paper. References [1] R.B. Anderson, The Fischer–Tropsch Synthesis, Academic Press, Orlando, 1984. [2] E. Iglesia, S.C. Reyes, R.J. Madon, S.L. Soled, Adv. Catal. 39 (1993) 221. [3] E. Iglesia, S.L. Soled, R.A. Fiato, J. Catal. 137 (1992) 212. [4] R. Qukaci, J.G. Goodwin, A. Singleton, Am. Chem. Soc. Div. Fuel Chem. 39 (1994) 1117. [5] M.E. Dry, in: G.J. Hutchings, M.S. Scurrell (Eds.), Synfuels, Catal. Today 6 (1990) 183. [6] S.H. Moon, Appl. Catal. 16 (1985) 289. [7] M. Blanchard, Appl. Catal. 9 (1984) 327. [8] J.G. Goodwin Jr., in: Proceedings of the Symposium on Methane Upgrading, ACS, Atlanta, 1991, p. 156. [9] L. Guezi, T. Hoffer, Z. Zsoldos, S. Zyode, G. Maire, F. Grain, J. Phys. Chem. 95 (1991) 802. [10] M.J. Ders, T. Shido, Y. Iwasawa, V. Ponec, J. Catal. 124 (1990) 350. [11] B.G. Johnson, C.H. Bartholomew, D.W. Goodman, J. Catal. 128 (1991) 231. [12] J. Ramirez, T. Klimera, Y. Huerta, J. Aracil, Appl. Catal. A 118 (1994) 73. [13] J. Venter, M. Kaminisky, G.L. Geoffrey, M.A. Vannice, J. Catal. 103 (1987) 450.
B.K. Sharma et al. / Applied Catalysis A: General 211 (2001) 203–211 [14] C.H. Bartholomew, in: L. Guezi (Ed.), Trends in Co Activation, Elsevier, Amsterdam, 1991 (Chapter 5). [15] H. Chen, A.A. Adesiner, Appl. Catal. A 112 (1994) 87. [16] G.V.D. Lec, V. Ponec, Catal. Rev. Sci. Eng. 29 (2/3) (1987) 183. [17] G. Dimitrova Penka, R. Mehandjiw Dimiter, J. Catal. 145 (1994) 356–363. [18] J. van de Loordrecht, M. van der Haar, A.M. van der Kraan, A.J. van Dillen, J.W. Geus, Appl. Catal. A 150 (1997) 365– 376. [19] R.L. Chin, D.M. Hercules, J. Phys. Chem. 86 (1982) 360. [20] H.M. Burce, G. Baker, Appl. Catal. A 123 (1995) 23–36. [21] E. Coutler Kent, A.G. Sault, J. Catal. 154 (1995) 56–64.
211
[22] A.A. Adesina, Appl. Catal. A 138 (1996) 345–367. [23] J.P. Hindermann, G.J. Hutchings, A. Kiennemann, Catal. Rev. Sci. Eng. 35 (1) (1993) 1–127. [24] B.K. Sharma, M.P. Sharma, S.K. Roy, S. Kumar, S.B. Tendulkar, S.S. Tambe, B.D. Kulkarni, Fuel 77 (1998) 1763– 1768. [25] S.B. Idage, S. Badrinarayanan, Langmuir 14 (1998) 2780. [26] P.M. Loggenberg, L. Carlton, R.G. Copperthwaite, G.J. Hutchings, J. Chem. Soc., Chem. Commun. (1987) 541. [27] D. Briggs, M.P. Seah, Practical Surface Analysis by Auger and X-ray Photoelectron Spectroscopy, Wiley, New York, 1984, p. 161.