Morphology and deactivation behaviour of Co-Ru/[gamma] [AL.sub.2][O.sub.3] Fischer-Tropsch synthesis catalyst.
Supported cobalt catalysts for the Fischer-Tropsch (FT) synthesis are receiving commercial applications for gas to liquid (GTL) technologies. Cobalt FT synthesis catalysts have many advantages over iron such as high conversion per pass, long lifetime, low water gas shift activity, and selectivity towards higher hydrocarbons. However, cobalt based catalysts used in FT synthesis are relatively expensive as compared to iron and require highly dispersed cobalt in conjunction with stable catalytic activity over an extended period of operation to remain economically attractive. Hence, the important challenges of development of cobalt catalysts have focused on decreasing the deactivation rate of supported catalysts [Krishnamurthy et al., 2002; Das et al., 2003; Jacobs et al., 2004; Tavasoli et al., 2007a].
The deactivation of cobalt catalysts is mainly due to the following reasons: oxidation of cobalt metal, cobalt support interactions, metal migration into the support lattice resulting in the formation of the inactive compounds (e.g. aluminate), the aggregation and growth of metal cobalt on the surface of catalyst, refractory coke formation, and the loss of metal cobalt because of attrition (especially for the three-phase slurry bed reactors) (Van Berge et al., 2000; Jacobs et al., 2002, 2004; Krishnamurthy et al., 2002; Das et al., 2003; Kiss et al., 2003; Tavasoli et al., 2007a). Water is the main FT by-product and accounts for ca. 50 wt% of all products. The amount of water depends on the synthesis gas conversion, reaction system and type of catalyst. When cobalt is deactivated, it means it is converted into either CoO or [Co.sub.3][O.sub.4] or mixed oxides of the form xCoO x y[Al.sub.2][O.sub.3] or Co[Al.sub.2][O.sub.4]. This is not thermodynamically possible unless this is affected by water (Jongsomjit and Goodwin, 2002; Tavasoli et al., 2007b).
Despite that the deactivation of cobalt catalysts has been studied extensively, literature review reveals that mechanism of the deactivation of cobalt catalyst along the reactor has not been studied. In other words, it is not well known whether factors such as coke formation, oxidation, sintering, etc. have their major influence on deactivation of catalysts located at different parts in a fixed bed reactor. This study has focused on illustrating the deactivation of ruthenium promoted alumina supported cobalt catalyst along the catalytic bed. The catalytic bed was divided into four parts and the deactivation of the catalyst from each bed was studied separately.
Ruthenium promoted (Ru/Co = 0.010 by weight) cobalt catalyst were prepared with 27 wt% cobalt on Conndea Vista Catalox B [gamma]-alumina support. The support was calcined at 500[degrees]C for 10 h prior to the impregnation. The catalyst was prepared by sequential aqueous impregnation of the support with the appropriate solutions of the cobalt nitrate (Co[(N[O.sub.3]).sub.2] x 6[H.sub.2]O 99.0% Merck) and Ruthenium(III) nitrosylnitrate (Ru(NO)[(N[O.sub.3]).sub.3]). After each step, the catalyst sample was dried at 120[degrees]C and after impregnation of metal, it was calcined at 450[degrees]C for 3.5 h.
Reaction Set-Up and Experimental Procedure
The catalyst was evaluated in terms of its Fischer-Tropsch synthesis (FTS) activity (g HC produced/g cat/min) and selectivity (the percentage of the converted CO that appears as a hydrocarbon product) in a tubular fixed-bed micro-reactor. The catalytic bed was divided into four parts in series. Typically, 0.5 g of the catalyst was charged into each catalytic bed (total amount of catalyst was 2 g and the length of catalytic bed was 60 cm). The reactor was placed in a molten salt bath with a stirrer to ensure a uniform temperature along the catalyst beds. The temperature of the bath was controlled via a proportional-integral-differential (PID) temperature controller. Separate Brooks 5850 mass flow controllers were used to add [H.sub.2] and CO at the desired rate to a mixing vessel that was preceded by a lead oxide-alumina containing vessel to remove carbonyls before entering to the reactor. Prior to the activity tests, the temperature was raised to 400[degrees]C with a heating rate of 1[degrees]C/min and the catalyst was reduced in a flow of [H.sub.2] for 12 h. The FTS tests were carried out at 220[degrees]C, 20 bar, and a [H.sub.2]/CO ratio of 2 for a period of 850 h. Also, a low feed flow rate of 60 mL/min was used to achieve high conversions and high water partial pressures that further accelerated the deactivation of the catalyst. Products were continuously passed through two traps, one maintained at 100[degrees]C (hot trap) and the other at 0[degrees]C (cold trap). The uncondensed vapour stream was reduced to atmospheric pressure through a back pressure valve. The flow was measured with a bubble flow meter and composition quantified using an on-line Varian 3800 gas chromatograph. The contents from the hot and cold traps were removed every 24 h, the hydrocarbon and water fractions separated, and then analyzed by GC. CO conversion and different product selectivities (the percentage of the converted CO that appears as a given product) were calculated based on the GC analyses. An Anderson-Schultz-Floury (ASF) distribution line was plotted for [C.sup.+.sub.3] products to determine the chain growth probability, [alpha].
After 850 h of FT synthesis, the flow of CO and [H.sub.2] were switched off and the used catalysts of the each bed were discharged. The physico-chemical properties of the used catalysts of each bed and that of the fresh catalyst were extensively studied as follows.
The cobalt loadings of the calcined fresh and used catalysts were verified by an inductively coupled plasma (ICP) atomic emission spectrometry (AES) system.
Brunauer, Emmett, and Teller (BET) Surface Area Measurements/Barrett-Joyner-Halanda (BJH) Pore Size Distributions
The surface area, pore volume, and average pore radius of the support, fresh and used calcined catalysts were measured by an ASAP-2010 Micromeritics system. The samples were degassed at 200[degrees]C for 4 h under 50 mTorr vacuum and their BET area, pore volume, and average pore radius were determined.
X-ray diffraction (XRD) measurements of the fresh and used catalysts were conducted with a Philips PW1840 X-ray diffractometer with monochromatized Cu/K[alpha] radiation. Using the Scherrer equation, the average size of the [Co.sub.3][O.sub.4] crystallites in the calcined catalysts was estimated from the line broadening of a [Co.sub.3][O.sub.4] peak at 2[theta] of 36.8[degrees]C.
Temperature Programmed Reduction
Temperature programmed reduction (TPR) spectra of the fresh and used catalysts were recorded using a Micromeritics TPD-TPR 290 system, equipped with a thermal conductivity detector. The catalyst samples were first purged in a flow of argon at 150[degrees]C, to remove traces of water, and then cooled to 40[degrees]C. The TPR of 50 mg of each sample was performed using 5.1 mol% hydrogen in argon with a flow rate of 40 [cm.sup.3]/min. The samples were heated from 40 to 900[degrees]C with a heating rate of 10[degrees]C/min.
Hydrogen Chemisorption and Re-Oxidation
The amount of chemisorbed hydrogen was measured using the Micromeritics TPD-TPR 290 system. 0.22 g of the calcined fresh and calcined used catalysts were reduced at 400[degrees]C for 12 h and then cooled to 100[degrees]C under hydrogen flow. Then the flow of hydrogen was switched to argon at the same temperature, which lasted for about 30 min in order to remove the weakly adsorbed hydrogen. Afterwards the temperature programmed desorption (TPD) of the samples was obtained by increasing the temperature of the samples, with a ramp rate of 10[degrees]C/min, to 400[degrees]C under the argon flow. The TPD spectrum was used to determine the cobalt dispersion and its surface average crystallite size. After the TPD of hydrogen, the sample was re-oxidized at 400[degrees]C by pulses of 10% oxygen in helium to determine the extent of reduction. It is assumed that [Co.sup.0] is oxidized to [Co.sub.3][O.sub.4]. The calculations are summarized below (Jacobs et al., 2002, 2004)
calibration value (L gas/area units) = loop volume x %analitical gas / mean calibration area x 100 (1)
[H.sub.2] uptake (mol/[g.sub.cat]) = analitical area from TPD x calibration value / sample weight x 24.5 (2)
%[D.sub.Total Co] = [H.sub.2] uptake x atomic weight x stoichiometry / %metal = number of [Co.sup.0] atoms on the surface x 100 / total number of [Co.sup.0] atom (3)
%[D.sub.Total Co] = [H.sub.2] uptake x atomic weight x stoichiometry / %metal = number of [Co.sup.0] atoms on the surface x 100 / total number of [Co.sup.0] atom (4)
[O.sub.2] uptake(mol/[g.sub.cat]) = sum of consumed pulses areas x calibration value / sample weight x 24.5 (5)
Fraction reduced = [O.sub.2] uptake (mol/g x cat) x 2/3 x atomic weight / Percentage metal (6)
Diameter [(nm).sub.Total Co] = 6000 / density x maximum area x dispersion (7)
Diameter [(nm).sub.reduced Co] = 6000 / density x maximum area x dispersion x Fraction reduced (8)
Measurement of Coke Deposition
The amount of coke deposition on the used catalysts was measured by UOP 703 standard method (Tavasoli et al., 2005). The sample was pretreated for the coke measurements as follows: After FT synthesis, the sample was treated under hydrogen flow of 120 mL/min starting from 220[degrees]C and by a ramp rate of 10[degrees]C/min raised to 400[degrees]C. The exit flow was monitored for C[H.sub.4] until its concentration was zero. The reactor was cooled to 40[degrees]C and the catalyst was then discharged and characterized.
RESULTS AND DISCUSSION
Activity and Selectivity Studies
Table 1 presents the FT synthesis rate (g CH/g cat/h), CO conversion, chain growth probability and different product selectivity during first 24 h of time-on-stream. Also Figure 1 shows the selectivity of all hydrocarbons produced during the first 24 h of FT reaction. It should be mentioned that for determination of the percentage of heavy alcohols the produced water was also analyzed. The FT synthesis productivity of the catalyst (0.295 g CH/g cat/h) and its products distribution (5.7% C[H.sub.4] selectivity and 90.21% [C.sub.5.sup.+] selectivity) are comparable to that of commercial cobalt based FT synthesis catalysts. It is to note that the FT synthesis productivity of the commercial catalysts are 0.25-0.32 g CH/g cat/h and the C[H.sub.4] and [C.sub.5.sup.+] selectivities of the commercial catalysts are 4-6% and 88-94%, respectively, indicating that the data are comparable (Tavasoli et al., 2007a).
Figure 2 presents the %CO conversion with time-on-stream during FT synthesis. It can be seen that, during 850 h FT synthesis, the FTS rate sharply decreases in the first 200 h, and then levels off. Two different steps are distinguishable. The FTS rate drops by 7% in the first 200 h while in the second 650 h the FTS rate only drops by 2.2%. The slope of the lines for the first and the second deactivation steps are different. The deactivation curve sloped steeply at first and then moderately when the time on stream exceeded 200 h. The loss of activity for the first deactivation step can be simulated with following linear correlation
[X.sub.CO] = -0.0316[T.sub.(h)] + 77.642 (9)
The linear deactivation mode suggests that the deactivation rate is zero order to CO conversion. This reveals that during first 200 h FT synthesis the deactivation rate is not related to the number of the catalyst active sites and the deactivation is caused by exterior factors such as partial pressure of water. It has been suggested that in FT synthesis on alumina-supported cobalt catalysts at high conversions, the loss of active sites is caused by water-induced oxidation of cobalt and formation of more refractory forms of oxidized cobalt generated by cobalt-alumina interactions (Krishnamurthy et al., 2002; Das et al., 2003; Kiss et al., 2003; Jacobs et al., 2004; Tavasoli et al., 2005, 2007b). When time on stream exceeded 200 h the catalyst deactivation could be simulated with a power law expression:
[X.sub.CO] = 77.9[T.sup.-0.0363.sub.(h)] (10)
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Assuming the deactivation rate is:
- dX / dt = k[X.sup.n] (11)
After integration and data reduction by least square fit, the power order (n) and the rate constant (k) can be determined as 28.5 and 3.88E-54 [[L.sup.(n-1)] / [(mol.sup.(n-1)] S)], respectively. The value is in the range that ordinary metal catalysts would experience during sintering (Bartholomew, 2001).
The uncondensed vapour stream of cold trap was reduced to atmospheric pressure through a pressure letdown valve. The composition of this stream was quantified using an on-line Varian 3800 gas chromatograph. Also, the contents in hot and cold traps were removed every 24 h and the hydrocarbon and water fractions separated. The contents of these traps were then analyzed using a Varian CP 3800 GC equipped with a Petrocol Tmdh fused silica capillary column and a flame ionization detector (FID) equipped in an offline GC. Figure 3 shows the methane and [C.sub.5.sup.+] liquid hydrocarbon selectivity variations with reaction time. This Figure displays that C[H.sub.4] selectivity decreases with time-on-stream during 850 h FT synthesis at 220[degrees]C and 20 bar. Also, Figure 3 shows that the [C.sub.5.sup.+] selectivity increases during 850 h FT synthesis. It has been shown (Tavasoli et al., 2007c) that the larger cobalt particles are more selective to higher molecular weight hydrocarbons and smaller cobalt particles are selective to methane and light gaseous hydrocarbons. The results shown in Figure 3 provide evidence that the deactivation of ruthenium promoted cobalt catalyst supported on [gamma]-alumina is dependent on the size of particles. It can be concluded that the smaller particles, selective for methane, are deactivated first, leading to enhancement of [C.sub.5.sup.+] selectivity and suppression of C[H.sub.4] production with time on stream. The higher C[H.sub.4] and [C.sub.5.sup.+] selectivity changes with time on stream during the first 200 h of FT synthesis is due to a higher rate of deactivation of smaller cobalt particles.
[FIGURE 4 OMITTED]
Figure 4 shows the carbon dioxide selectivity variations with reaction time. As shown in this figure, C[O.sub.2] selectivity increases with time-on-stream during 850 h FT synthesis at 220[degrees]C and 20 bar. It has been shown that cobalt oxides and also mixed oxides of cobalt and alumina of the form xCoO x y[Al.sub.2][O.sub.3] are selective to production of C[O.sub.2] (Li et al., 2002). Increasing the C[O.sub.2] selectivity with the TOS of FT synthesis in Figure 4 confirms that some surface cobalt atoms have been reoxidized to form cobalt oxide or some other form of cobalt which are active for the water-gas shift reaction. Also Figure 4 shows that a plateau region is reached after 450-500 h which indicates that the loss of active sites by water-induced oxidation of cobalt and formation of more refractory forms of oxidized cobalt generated by cobalt-alumina interactions decreases significantly after 500 h of continues FT synthesis.
Structural Changes of the Catalysts
The results of ICP, BET and porosity tests for support, fresh catalyst and used catalysts of different beds are listed in Table 2. This table shows that the metal loss during FT synthesis is zero. The BET surface area and pore volume of fresh catalyst is much lower than that of the [gamma]-[Al.sub.2][O.sub.3] which indicates pore blockage due to cobalt loading on the support. Also, it could be seen that the pore volume, BET surface area and average pore radius for the used catalysts showed a small difference. The difference is higher for the catalyst from the last two beds (Bed # 3 and 4). BET surface of the catalyst of the 4th bed decreased from 168 to 152 which may be assigned to the pore blockage due to coke formation or sintering. At the same time, the average pore radius of the catalyst of the fourth bed is increased from 4.30 to 4.80 nm and its pore volume is decreased from 0.45 to 0.43 [cm.sup.3]/g which can confirm the blockage of smaller pores with refractory coke. This reveals that the skeleton of the catalysts of the first beds were more stable than those from beds 3 and 4 during FT synthesis.
Figures 5 and 6 show the XRD patterns of the fresh and used catalysts of different beds. On the spectra of the fresh catalyst the peaks at 46.1 and 66.5[degrees]C correspond to [gamma]-alumina while the other peaks relate to the different crystal planes of [Co.sub.3][O.sub.4]. The peak of 36.8[degrees]C is the most intense peak of [Co.sub.3][O.sub.4]. No peak was observed indicating formation of cobalt support compounds.
XRD patterns of the used catalysts on Figure 6 show that, important structural changes occurred during reaction to the catalysts in comparison to the calcined fresh catalyst. In the spectra of all the used catalyst samples support peaks appeared at 46.1 and 66.5[degrees]C. Also, the peaks at 2[theta] values of 36.8 and 60[degrees]C can be attributed to different crystal planes of cobalt oxides (Jacobs et al., 2002; Jongsomjit and Goodwin, 2002). These peaks show that a fraction of cobalt clusters in all catalyst beds may oxidize in presence of significant amount of water formed during FT synthesis. It is to note that, some amount of the cobalt oxide in the used samples is probably formed during the discharge step at room temperature. It is proposed that bulk oxidation of large Co metal crystallites to CoO or [Co.sub.3][O.sub.4] is not thermodynamically favoured at typical FT synthesis conditions, but metal-oxygen bonds at metal surfaces are stronger than in the bulk oxides, making the oxidation of Co surfaces possible even when bulk oxidation is unfavourable (Jacobs et al., 2002; Jongsomjit and Goodwin, 2002). Peaks corresponding to cobalt metal appeared at 2[theta] values of 43.8[degrees], 51.5[degrees], and 75.6[degrees] for the used catalysts of beds 1 to 4. These peaks well correlate with the reported literature values for a cubic cobalt structure (Jacobs et al., 2002; Jongsomjit and Goodwin, 2002).
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
The peak at 2[theta] value of 56.5[degrees]C in the spectra of the catalyst of bed 3 and the two peaks at 2[theta] values of 41.3[degrees] and 56.5[degrees] in the spectra of the catalyst of bed 4 correlate well with [Co.sub.2]C (Tavasoli et al., 2005). These two peaks did not appear on the spectra of the catalysts of bed 1 and 2. It seems that a small amount of [Co.sub.2]C is formed during the FT synthesis only in the last beds of the catalytic beds. These results show that cobalt carbide formation mainly occurs at hydrogen deficient regions of the catalytic beds.
[FIGURE 7 OMITTED]
The small peak at 2[theta] value of 49 in the spectra of the catalyst of bed 3 and the peaks at 2[theta] values of 29.6, 49, and 78[degrees]C in the spectra of the catalyst of bed 4 correspond to Co[Al.sub.2][O.sub.4] spinel. Significant amount of water formed in the last catalytic beds during the FT synthesis is responsible for the formation of the small amounts of cobalt aluminates in bed 3 and noteworthy amounts of cobalt aluminates in bed 4 (Jacobs et al., 2002; Jongsomjit and Goodwin, 2002; Tavasoli et al., 2005).
These results suggest that during FT synthesis, different cobalt species can be formed, some of which can be regenerated to cobalt during next hydrogen treatment steps. Bulk [Co.sub.3][O.sub.4] became completely reduced at temperatures below 450[degrees]C while Co[Al.sub.2][O.sub.4] only can be reduced at temperatures above 800[degrees]C and is therefore recognized as an irreducible phase. Also, [Co.sub.2]C cannot be reduced at temperatures below 600[degrees]C (Tavasoli et al., 2005, 2007b). XRD results also reveal that, morphology change and as well as the rate of catalyst deactivation in the last catalytic beds are much higher than that of the earlier beds and the catalyst deactivation of the earlier beds (i.e.1 and 2) is mostly reversible but the catalyst deactivation of the last two beds (3 and mainly 4) is not completely reversible.
The TPR spectrum of the calcined fresh catalyst is shown in Figure 7. In this Figure, the first peak at 345[degrees]C is typically assigned to the reduction of [Co.sub.3][O.sub.4] to CoO, although a fraction of the peak likely comprises of the reduction of the larger, bulk-like [Co.sup.0] species to CoO. The second peak at 470[degrees]C, with a small shoulder to about 570[degrees]C is mainly assigned to the second reduction step, which is mainly reduction of CoO to [Co.sup.0]. This peak also includes the reduction of small amounts of cobalt species that interact with the support, which extends the TPR spectra to higher temperatures of about 570[degrees]C. Small tailing of the second TPR peak also reveals that in ruthenium promoted cobalt catalyst the degree of interaction between active metal and support is not significant (Tavasoli et al., 2005, 2007b).
Figure 8 shows the TPR spectra of the used catalysts of all beds. This figure demonstrates that two TPR peaks in case of the fresh calcined catalyst have been transferred to a single peak in case of all used catalyst at lower temperatures with a broad tailing extended to higher temperatures. The temperature of this peak is 300, 305, 310, and 320[degrees]C for the used catalysts of beds 1, 2, 3, and 4, respectively. In case of the used catalyst of bed 1, the tailing of TPR peak is extended to about 660[degrees]C. This broad tailing was mainly assigned to the reduction of cobalt species that interact with the alumina support and make mixed oxides of the form xCoO x y[Al.sub.2][O.sub.3], which hinders their reduction to [Co.sup.0] and extend the TPR spectra to higher temperatures (Tavasoli et al., 2005). This indicates that FT synthesis increases the extent of interaction of cobalt with alumina. In case of the used catalysts of bed 2 and 3, the tailing of TPR peak is extended to about 680 and 700[degrees]C, respectively. This reveals that interaction of cobalt with alumina and formation of mixed oxides of the form xCoO x y[Al.sub.2][O.sub.3] is increased along the catalytic bed. It is noted the water can enhance the interaction of cobalt and alumina and shifts the TPR peaks to higher temperatures. Increasing the amount of water partial pressure along the catalytic bed increases the interaction of active metal and support as well as catalyst deactivation along the catalytic bed.
In the case of the used catalyst of bed 4, in addition to a very broad peak, another peak has appeared on the TPR profile of the catalyst at 804[degrees]C (see Figure 8), which can be attributed to the cobalt aluminate. This peak was not present for the used catalysts of beds 1, 2, and 3. It has been shown for alumina-supported cobalt catalysts that the high partial pressure of water promotes the formation of Co-phases with reduction properties similar to cobalt aluminates (Tavasoli et al., 2005). It seems that high amounts of water in the 4th catalytic bed increases the mobility of ions on the support and therefore increases the diffusion of [Co.sup.2.sub.+] (and [Al.sup.3.sub.+]) which in turn leads to formation of [Co.sup.2.sub.+] with O-Al ligands or cobalt aluminates spinel, and also exchange of [Co.sup.2.sub.+] in [Co.sub.3][O.sub.4] with [Al.sup.3.sub.+] (Arnoldy and Moulijin, 1995). It should be mentioned that, the broadening of the first peak in this bed is suggested to be due to different levels of interaction with the support which delays their reduction to [Co.sup.0]. All these features suggest part of the cobalt in the high water partial pressures strongly interacts with the support, as also evidenced from XRD patterns (Figure 6).
[FIGURE 8 OMITTED]
Figure 9 shows the hydrogen uptake and percentage reduction for the fresh and used catalysts of beds 1, 2, 3, and 4 determined by hydrogen chemisorption and oxygen titration tests. It is to note that the hydrocarbon products from the used catalysts were extracted with an organic solvent and dried and calcined in an oven before the hydrogen chemisorption and re-oxidation test. As shown in Figure 9, for the used catalysts, hydrogen uptake and the percentage reduction decrease, which reveals that regeneration of the used catalysts at the reduction conditions similar to reduction step for fresh catalyst, cannot recover the catalyst reducibility as well as catalyst activity by 100% at all locations of the reactor. Hydrogen uptake decreases from 94 for the fresh catalyst to 91, 89.4, 79.5, and 66.1 ([micro]mol [H.sub.2] desorbed/g cat.) for the used catalysts of beds 1-4, respectively. At the same time, the percentage reduction which is calculated from the amount of [O.sub.2] consumed, decreases from 57.7% for the fresh catalyst to 56.3%, 55.6%, 50.2%, and 43% for the used catalysts of beds 1-4, respectively. It gives the impression that some parts of cobalt interacted strongly with support or made irreducible cobalt aluminates spinel that cannot be reduced at the reduction conditions similar to that used for fresh catalyst. The reducibility of the catalysts decreases by 2.4%, 3.6%, 13%, and 25.5% for the used catalysts of beds 1-4, respectively. These results clearly show that the catalyst of last bed is deactivated 10 times more than the catalyst of the first bed. It reveals that the amounts of these irreducible cobalt species are significantly higher at the last 25% of the catalytic bed compared to the other parts of the catalytic bed.
Figure 10 shows the percentage dispersion and particle diameters calculated based on the reduced cobalt in fresh and used catalysts of beds 1, 2, 3, and 4. As shown in this figure, the %dispersion of cobalt for all the used catalysts is lower than that of the fresh catalyst. In contrast to the %dispersion, particle diameter in the used catalysts shows a small increase. Percentage dispersion decreased by 0.5%, 1.7%, 2.8%, and 8.8% for the used catalysts of beds 1-4, respectively, while the particle diameter increased by 0.8%, 1.3%, 2.8%, and 6.1% for the used catalysts of beds 1-4, respectively. This suggests a metal growth or sintering that occurred during FT synthesis. So in accordance to the deactivation due to the cobalt support interaction and cobalt aluminates formation, the amount of deactivation due to sintering is higher at the last parts (3 and 4) of the catalytic beds. FT synthesis temperature is low to boost the cluster growth at the catalyst surface. But it is suggested that water produced during FT synthesis increases the oxidation-reduction cycles on the catalyst surface and therefore enhances sintering rate of supported metals (Jacobs et al., 2004). Higher amount of water present in the last parts of the catalytic bed is responsible for higher sintering rate for catalysts located in the parts 3 and 4 in the reactor.
The amount of coke deposition on the used catalysts was measured using UOP 703 standard method (Tavasoli et al., 2005). The quantities of coke deposited on the used catalysts of beds 1, 2, 3, and 4 are shown in Figure 11. The coke deposition in all the beds seems to be low. It is highest (0.5%) for the bed # 4. The amount of coke formation in the 4th catalytic bed is 6 times more than that of the 1st catalytic bed. Production of heavy unsaturated hydrocarbons during the FT synthesis can be the main source of refractory coke formation on the cobalt crystallite sites, which led to catalyst deactivation. It has been shown that the amount of heavy unsaturated hydrocarbons in FT products increases with increasing the feed residence time in FT synthesis reactors (Tavasoli et al., 2007a). Increasing the amount of these hydrocarbons as a source of coke formation in the last catalytic beds is the cause for higher amount of coke formation in the last catalytic beds.
[FIGURE 9 OMITTED]
In Fischer-Tropsch synthesis, high water partial pressures cause rapid catalyst deactivation. At high water partial pressures, catalyst deactivation rate is not dependent on the number of the catalyst active sites. It is zero order to CO conversion and independent of the size of active sites. At low water partial pressures, catalyst deactivation could be simulated with a power law expression: -d[X.sub.CO]/dt = k[X2.sup.8.5.sub.CO]. The physical properties of the catalysts in the first half of the catalytic bed did not changed significantly compared to that of the 2nd half during FTS. Interaction of cobalt with alumina and formation of mixed oxides of the form xCoO x y[Al.sub.2][O.sub.3], [Co.sub.2]C and Co[Al.sub.2][O.sub.4] increased along the catalytic bed. Also, the rate of sintering of cobalt clusters and refractory coke formation increases along the catalytic bed.
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
NOMENCLATURE P pressure (bar) T temperature ([degrees]C) Greek Symbol [alpha] chain growth probability Abbreviations FTS Fischer-Tropsch synthesis ASF Anderson-Schultz-Floury GTL gas to liquid TOS time on stream TPR temperature programmed reduction GC gas chromatograph CH hydrocarbon S selectivity Subscript n carbon number
Manuscript received April 23, 2008; revised manuscript received July 15, 2008; accepted for publication August 27, 2008
Arnoldy, P. and J. A. Moulijin, "Temperature-Programmed Reduction of CoO/[Al.sub.2][O.sub.3] Catalysts," J. Catal. 93, 38-54 (1995).
Bartholomew, C. H., "Mechanisms of Catalyst Deactivation," Appl. Catal. A 212, 17-60 (2001).
Das, T., G. Jacobs, P. M. Patterson, W. A. Conner, J. Li and B. H. Davis, "Fischer-Tropsch Synthesis: Characterization and Catalytic Properties of Rhenium Promoted Cobalt Alumina Catalysts," Fuel 82, 805-815 (2003).
Jacobs, G., P. M. Patterson, Y. Zhang, T. Das, J. Li and B. H. Davis, "Fischer-Tropsch Synthesis: Deactivation of Noble Metal-Promoted Co/[Al.sub.2][O.sub.3] Catalysts," Appl. Catal. A 233, 215-226 (2002).
Jacobs, G., P. M. Patterson, T. Das, M. Luo and B. H. Davis, "Fischer-Tropsch Synthesis: Effect of Water on Co/[Al.sub.2][O.sub.3] Catalysts and XAFS Characterization of Reoxidation Phenomena," Appl. Catal. A 270, 65-76 (2004).
Jongsomjit, B. and J. G. Goodwin, "Co-Support Compound Formation in Co/[Al.sub.2][O.sub.3] Catalysts: Effect of Reduction Gas Containing CO," Catal. Today 77, 191-204 (2002).
Kiss, G., C. E. Kliewer, G. J. DeMartin, C. C. Culross and J. E. Baumgartner, "Hydrothermal Deactivation of Silica-Supported Cobalt Catalysts in Fischer-Tropsch Synthesis," J. Catal. 217, 127-140 (2003).
Krishnamurthy, S., M. Tu, M. P. Ojeda, D. Pinna and E. Iglesia, "An Investigation of the Effects of Water on Rate and Selectivity for the Fischer-Tropsch Synthesis on Cobalt-Based Catalysts," J. Catal. 211, 422-433 (2002).
Li, J., Y. Zhang, G. Jacobs, T. Das and B. H. Davis, "Fischer-Tropsch Synthesis: Effect of Water on the Deactivation of Pt Promoted Co/[Al.sub.2][O.sub.3] Catalysts," Appl. Catal. A 228, 203-212 (2002).
Tavasoli, A., A. Karimi, A. Khodadadi, Y. Mortazavi and M. Mousavian, "Accelerated Deactivation and Activity Recovery Studies of Ruthenium and Rhenium Promoted Cobalt Catalysts in Fischer-Tropsch Synthesis," Iranian J. Chem. Chem. Eng. 24, 25-36 (2005).
Tavasoli, A., A. Karimi, K. Sadaghianizadeh, A. Khodadadi, Y. Mortazavi and S. Sharifnia, "Integrated process and apparatus for producing liquid fuels," US Patent No. 7226954 B2 (2007a).
Tavasoli, A., A. Nakhaeipour and K. Sadaghiani, "Raising Co/[Al.sub.2][O.sub.3] Catalyst Lifetime in Fischer-Tropsch Synthesis by Using a Novel Dual-Bed Reactor," Fuel Processing Tech. 88, 461-469 (2007b).
Tavasoli, A., K. Sadaghiani, A. Nakhaeipour and M. G. Ahangari, "Cobalt Loading Effects on the Structure and Activity for Fischer-Tropsch and Water-Gas Shift Reactions of Co/[Al.sub.2][O.sub.3] Catalysts," Iranian J. Chem. Chem. Eng. 26, 9-16 (2007c).
Van Berge, P. J., J. van de Loosdrecht, S. Barradas and A. M. Van der Kraan, "Oxidation of Cobalt Based Fischer-Tropsch Catalysts as a Deactivation Mechanism," Catal. Today 58, 321-334 (2000).
Ahmad Tavasoli, (1,2) Mohammad Irani, (2) Reza M. Malek Abbaslou, (1) Mariane Trepanier (1) and Ajay K. Dalai (1) *
(1.) Catalysis & Chemical Reaction Engineering Laboratories, Department of Chemical Engineering, University of Saskatchewan, Saskatoon, SK, Canada S7N5C5
(2.) Research Institute of Petroleum Industry, P.O. Box 14665-1998, Tehran, Iran
* Author to whom correspondence may be addressed. E-mail address: email@example.com
Table 1. FT synthesis results during first 24 h FT synthesis rate (g CH/g cat/h) 0.295 Percentage CO Conversion 78.06 Chain growth probability ([alpha]) 0.92 C[0.sub.2] selectivity 0.94 C[H.sub.4] selectivity 5.70 [C.sub.2]-[C.sub.4] selectivity 3.15 [C.sub.5.sup.+] selectivity 90.21 Percentage parafins 91.45 Percentage olefins 5.65 Percentage alcohols 2.90 Table 2. ICF; BET and porosity data Support/catalyst Amount of Co BET ([m.sup.2]/g) wt.% [gamma]-[A1.sub.2][0.sub.3] -- 270 Fresh catalyst 27 168 Used catalyst, bed # 1 27 167 Used catalyst, bed # 2 27 166 Used catalyst, bed # 3 27 157 Used catalyst, bed # 4 27 152 Support/catalyst Pore volume Average pore (single point) radius (nm) ([cm.sup.3]/g) [gamma]-[A1.sub.2][0.sub.3] 0.639 4.72 Fresh catalyst 0.45 4.30 Used catalyst, bed # 1 0.45 4.30 Used catalyst, bed # 2 0.45 4.30 Used catalyst, bed # 3 0.44 4.50 Used catalyst, bed # 4 0.43 4.80