- Email: [email protected]
Studies on product distribution of nanostructured iron catalyst in Fischer–Tropsch synthesis: Effect of catalyst particle size
G Model
JIEC-1369; No. of Pages 6 Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec
Studies on product distribution of nanostructured iron catalyst in Fischer–Tropsch synthesis: Effect of catalyst particle size Ali Nakhaei Pour *, Mohammad Reza Housaindokht Department of Chemistry, Ferdowsi University of Mashhad, P.O. Box 9177948974, Mashhad, Iran
A R T I C L E I N F O
A B S T R A C T
Article history: Received 24 March 2013 Accepted 12 May 2013 Available online xxx
The dependencies of hydrocarbon product distributions of Fischer–Tropsch synthesis by iron catalysts on catalysts particle size are studied. The concept of two superimposed Anderson–Schulz–Flory distributions applied for represent size dependency of product distributions. A series of catalysts with different particle size are prepared by microemulsion method. It is found that the carbon number of produced hydrocarbon decreased with decreasing the catalyst particle size. These results indicate the H2 concentration on catalyst surface decreased by increasing the catalyst particle size. Thus the concentration of monomers that exhibited higher degree of hydrogenation (like CH2 species) on the surface of catalyst increased with decreasing the catalyst particle size. ß 2013 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
Keywords: Chain length distribution Fischer–Tropsch synthesis Iron catalyst Particle size
1. Introduction Strong interest in the Fischer–Tropsch synthesis (FTS) that converts synthesis gas into hydrocarbons is reappearing because it is one of the major routes that convert natural gas into liquid energy carriers [1–3]. The Fischer–Tropsch synthesis (FTS) consisting of a complex multi component mixture of linear and branched hydrocarbons and oxygenated products. Main products are linear paraffins and a-olefins. Herington first treated the molar distribution of hydrocarbons from FTS in terms of a polymerization mechanism [4–6]. The same formulation was rediscovered by Anderson et al. in 1951 and named the Anderson–Schulz–Flory (ASF) distribution [4,5,7,8]. In the ASF model, the formation of hydrocarbon chains was assumed as a stepwise polymerization procedure and the chain growth probability was assumed to be independent of the carbon number. For all FTS catalysts deviations from the ideal Anderson–Schulz–Flory (ASF) distributions are observed in many studies [9–11]. The usual deviations of the distribution of the linear hydrocarbons are a relatively higher selectivity to methane, a relatively lower selectivity to ethene, and an increase in the chain growth probability with increasing molecular size. Some authors interpreted the deviations from the standard ASF distribution by the superposition of two ASF distributions [10,12–14]. They suspected the existence of two sorts of sites for the chain growth on the catalyst surface and each
* Corresponding author. Tel.: +98 5118795457; fax: +98 5118795457. E-mail addresses: [email protected], [email protected] (A. Nakhaei Pour).
site might individually yield the ideal ASF distribution with different chain growth probabilities. Madon and Taylor [15] interpreted this bimodal distribution by differently sites causing different growth probabilities. Ko¨nig and Gaube [10] supposed that in the case of alkali promoted iron catalysts, all active sites are not influenced by alkali. Therefore the distribution of higher growth probably was attributed to products formed on promoted active sites while the other part of the product is formed on unpromoted sites with a similar growth probability as evaluated for the product formed on unprompted iron catalysts. Patzlaff et al. [11] studied the effects of 1-alkene readsorption and secondary chain growth on the product distribution of FTS on iron and cobalt catalyst by cofeeding 1-alkenes. They found that chain length distributions of products obtained on cobalt catalysts are slightly modified by secondary chain growth of readsorbed alkenes and hydrogenolysis of hydrocarbons. The FTS reaction is a surface phenomenon, therefore for optimum catalyst performance, maximum metal usage must be achieved. A rule of thumb in heterogeneous catalysis is that smaller metal crystallites provide largest surface area on which the reaction may happen [16–20]. In our previous work the effect of presence of the zeolite on product distributions of iron catalyst studied [21]. In the other our previous work the effect of Ca, Mg and La promoters on physico-chemical properties and the catalytic activity and selectivity during FTS performance studied [22,23]. In this experiment the effect of catalyst particle size on the carbon number distribution of Fischer–Tropsch products on iron catalysts are studied, using a modified Anderson–Schulz–Flory (ASF) distribution with two chain growth probabilities.
1226-086X/$ – see front matter ß 2013 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jiec.2013.05.019
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 2
A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
2. Experimental 2.1. Catalyst preparation and characterization Fe–Cu nanoparticles were prepared by coprecipitation in a water-in-oil microemulsion [24–28]. The precipitation was performed in the single-phase microemulsion operating region. In order to achieve a series of catalysts displaying different hematite particle size the surfactant-to-oil (S/O) weight ratio was set to a value of 0.3 the water-to-surfactant molar ratio (W/S) was varied from 4 to 12. 2.2. Catalyst characterization The surface area was calculated from the Brunauer–Emmett– Teller (BET) equation and pore volume, average pore diameter, and pore size distribution of the catalysts were determined by N2 physisorption using a Micromeritics ASAP 2010 automated system. A 0.5 g catalyst sample was degassed at 373 K for1 h and 573 K for 2 h prior to analysis. The analysis was done using N2 adsorption at 77 K. Both the pore volume and the average pore diameter were calculated by Barret–Joyner–Hallender (BJH) method from the adsorption isotherm. XRD was used to determine the phase composition of catalysts before and after pretreatments. The XRD spectrum of the catalysts were collected by a X-ray diffractometer, Philips PW1840 X-ray diffractometer, using monochromatized Cu/Ka radiation (40 kV, 40 mA) and a step scan mode at a scan rate of 0.028 (2u) per second from 108 to 808. XRD peak identification was recognized by comparison to the JCPDS database software. The average crystallite size of samples, dXRD, can be estimated from XRD patterns by applying full-width half-maximum (FWHM) of characteristic peak (1 0 4) Fe2O3 located at 2u = 33.38 peak to Scherrer equation: dXRD ¼
0:9 l FWHM cos u
with the same mesh size range. The catalyst samples were activated by a 5% (v/v) H2/N2 gas mixture with space velocity equal to 15.1 nl/gcat/h at 1 bar and 1800 rpm. The reactor temperature increased to 673 K with a heating rate of 5 K/min, maintained for 1 h at this temperature, and then reduced to 543 K. The activation is followed by the synthesis gas stream with H2/CO ratio of 1 and space velocity equal to 3.07 nl/gcat/h for 24 h at 1 bar and 543 K. After catalyst activation, synthesis gas was fed to the reactor at conditions operated at 563 K, 17 bar, H2/CO ratio of 1 and a space velocity equal to 4.9 nl/gcat/h. After reaching steady state, the FTS reaction rate was measured. The out gas was analyzed by a gas chromatograph (Varian CP3800) equipped with TCD and FID detectors. The CO, CO2, N2, and O2 were analyzed through two packed column in series (Molecular sieve13 CP 81025 with 2 m length, and 3 mm OD, and Hayesep Q CP1069 with 4 m length, and 3 mm OD) connected to TCD detector. The C1–C5 hydrocarbons were analyzed via a capillary column (CP fused silica with 25 m 0.25 mm 0.2 mm film thickness) connected to FID detector. Hydrogen was analyzed through Shimatzu, GC PTF 4C, equipped with TCD detector and two column in series (Propack-Q with 2 m length, and 3 mm OD for CO2, C2H4 and C2H6 separation and molecular sieve-5A with 2 m length, and 3 mmOD for CO, N2, CH4 and O2 separation), which were connected to each other via a three way valve. The collected liquid (including hydrocarbons and oxygenates) analyzed offline with Varian CP-3800 gas chromatograph equipped with capillary column (TM DH fused silica capillary column, PETRO COL 100 m 0.25 mm 0.5 mm film thickness) connected to FID detector. Total mass balances were performed with the carbon material balance closed between 97 and 103%. This criterion was adopted since compounds containing carbon and hydrogen might accumulate in the reactor, in the form of high molecular weight hydrocarbons.
(1)
where l is the X-ray wavelength (1.5406 A˚ in this study) and u is the diffraction angle for the (1 0 4) plane. The H2-TPD experiments were performed by means of the temperature-programmed desorption (TPD) of H2 on the catalyst (0.5 g), which was packed in a shallow-bed quartz reactor with a low dead volume from 300 K to 1100 K at a linear heating rate of 5 K/min while Ar was used as a carrier gas. A thermal conductivity detector (TCD) was used to measure the H2 desorbed in the TPD quantitatively. The catalyst was first reduced with H2 at 673 K and 1 bar for 11 h. Then the sample was heated in Ar from 323 K to 673 K, held at 673 K until the baseline leveled off (ensuring complete removal of adsorbed species on the reduced catalyst surface has been achieved), and finally cooled to 323 K for TPD tests. In the subsequent steps, H2 adsorption on the catalyst was performed at 323 K for 30 min, and then the sample was purged with Ar in order that weakly adsorbed species could be removed until the baseline leveled off. Following this, H2-TPD was being carried out while the temperature increased to 1050 K. H2 chemisorptions uptakes determined by integrating the area of H2-TPD curves as compared to the certain amounts of gas passed through the TCD. 2.3. Experimental apparatus and procedure Steady-state FTS reaction rates measured in a continuous spinning basket reactor. A detailed description of the experimental setup and procedures has been provided in our previous works [29]. The fresh catalyst is crushed and sieved to particles with the diameter of 0.25–0.36 mm (40–60 ASTM mesh). The weight of the catalyst loaded was 2.5 g and diluted by 30 cm3 inert silica sand
3. Data analysis Hydrocarbon products of the Fischer–Tropsch synthesis are generally taken to follow the ASF distribution. For carbon number i, the mole fraction of product xi as determined by a single chain growth probability a is given by: xi ¼ ð1 aÞai1
(2)
In this work the carbon number distribution of Fischer–Tropsch products on iron catalysts studied by modified Anderson–Schulz– Flory (ASF) distribution with two chain growth probabilities. This method proposed by Donnelly et al. [14] to characterize the carbon number distribution of FTS hydrocarbon products by ASF distributions with different chain growth probabilities. þ B ai1 xi ¼ A ai1 1 2
(3)
Note that at the break point on ASF diagram, the contributions of each term in Eq. (3) are equal: ¼ B ai1 A ai1 1 2
i¼z
(4)
In break point, i illustrate as z and necessarily is not an integral carbon number. In this model A and B were not correspond directly to the fractions of products produced from a1 and a2, respectively. Instead the sum of the mole fractions over all carbon numbers is unity: 1 1 X X xi ¼ ½A a1 ði1Þ þ B a2 ði1Þ ¼ 1 i¼1
(5)
i¼1
Methane and ethene do not obey the ASF equation and after removing C1 and C2 products to fit theoretical distributions to data
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
3
leads to: 1 1 X X ði1Þ ði1Þ xi ¼ ½A a1 þ B a2 Að1 þ a1 Þ Bð1 þ a2 Þ i¼3
i¼1
xexp ¼ 1 xexp 1 2
(6)
Therefore, the determination of growth probabilities of the bimodal ASF distribution is based on hydrocarbons with carbon numbers i > 2. The bimodal distribution is characterized by two growth probabilities (a1 and a2) and the point of intersection j of both ASF distributions. The mathematical procedure given by Donnelly et al. [14] arrives at: " A ¼ ð1 xexp xexp 1 2 Þ
a12 a1 þ 1 a1 a2
j1 #1 a22 1 a2
(7)
and finally at: "
xi 1 xexp xexp 1 2
¼ ai1 þ 1
a1 a2
"
j1
a21 a1 þ 1 a1 a2
# ai1 2
j1 #1 a22 1 a2
(8)
The three parameters a1, a2 and j are evaluated by minimizing the least squares:
F¼
I X 2 ðln x1 ln xexp Þ ¼ min l
(9)
i¼3
in which I is the upper limit of experimental data. The advantage of Donnelly method is its’ independence of this upper limit I. However, the intersection point j is a formal parameter without any physicochemical relevance. Also this method has been extended by the calculation of m1 the total mole fraction of hydrocarbons formed with the growth probability a1. 1 X A ai1 1
m1 ¼
i¼3
1 xexp xexp 1 2
1 X A ai1 Að1 þ a1 Þ 1
¼
i¼1
1 xexp xexp 1 2
(10)
Using the limiting value of the geometric series 1 X
ai1 ¼ ð1 aÞ1
(11)
i¼1
gives
m1 ¼
A a21 ð1 a1 Þð1 xexp xexp 1 2 Þ
(12)
and with Eq. (7)
m1 ¼
ð1 a2 Þ=ð1 a1 Þa21 ½ð1 a2 Þ=ð1 a1 Þa21 þ ðða1 Þ=ða2 ÞÞj1 a22
(13)
Since two independent ASF distributions are superimposed the fraction of hydrocarbons formed with a chain growth probability a2 equals:
m2 ¼ 1 m1
(14)
4. Results and discussion 4.1. Catalyst characterization Nanostructure iron catalysts characterized by X-ray diffraction (XRD) measurement after calcination. Fig. 1 shows the XRD patterns of iron catalysts, the characteristic peaks corresponding to
Fig. 1. XRD pattern of the prepared catalysts after calcinations. (a) W/S = 4, (b) W/ S = 6, (c) W/S = 8 and (d) W/S = 12.
(0 1 2), (1 0 4), (1 1 0), (1 1 3), (0 2 4), (1 1 6), (0 1 8), (2 1 4), (3 0 0) planes are located at 2u = 24.38, 33.38, 35.88, 40.88, 49.68, 54.18, 57.68, 64.18 and 65.68, respectively [30–34]. They show very close to the ones with cubic hematite structured Fe2O3 crystal in JCPDS database. It shows that the hematite structure once formed remains stable during subsequent aqueous impregnation and thermal treatment. This strongly infers that microemulsion system modules only physical properties of reaction medium without changing the reaction paths and arrangements of crystal structure. The characteristic peak at 2u = 33.38 corresponds to the hematite 104 plane was used to calculate the average metal particle size by the Scherrer equation. The calculated dXRD for the samples listed in Table 1. The H2-TPD results, BET surface area and textural properties and pore size of the fresh iron-based catalysts are shown in Table 1. As shown in Table 1, the average particle sizes of hematite are linearly correlated with the water-to-surfactant ratio used during the microemulsion catalyst preparation route. In fact, nanoparticles are formed in the internal structure of the microemulsion, which is determined by the ratio of water-to surfactant. At high oil concentration, the bicontinuous phase is transformed into a structure of small water droplets within a continuous oil phase (reverse micelles) when surfactant is added. Thus, the results show that the size of different droplets determines the cobalt’s particle size, depending on the amount of surfactant [35,36]. As shown in this table, surface area and particle size of catalysts are changed with W/S ratio in microemulsion system. When the total surfactants-to-oil phase weight ratio was constant, the particle size decreased and surface area increased slightly with the increase in the water content. Also Table 1 shows the H2 adsorbed on surface of catalysts increased by decreasing of catalyst particle size.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
4
Table 1 H2 chemisorption, BET surface area and textural properties of prepared catalysts. W/S
Catalyst particle size (nm)a
BET surface area (m2/g)
Average pore size (nm)b
Pore volume (cm3/g)b
H2 uptake (102 mmol H2/mmol Fe)
4 6 8 12
16 21 24 28
68.2 56.4 48.2 43.5
16.2 19.6 22.3 25.9
0.14 0.16 0.19 0.21
7.8 7.1 6.2 5.7
a b
Calculated from XRD results. These values were calculated by BJH method from desorption isotherm.
4.2. Catalyst activity and products distribution The FTS reaction yields organic compounds and the water and/ or carbon dioxide as by-products. Thus, carbon monoxide can be consumed for the formation of organic compounds or carbon dioxide. Therefore, the rate of FTS reaction must be written as: r FTS ¼ r CO r CO2
(15)
where r CO2 is the rate of CO2 formation (water–gas shift reaction) and rCO is the rate of CO consumption. The FTS rate equals the rate of formation of organic compounds on carbon basis. The experimental results of the effects of catalyst particle size on FTS reaction rates were listed in Table 2. Table 2 shows the change of syngas conversions, FTS reaction rates and hydrocarbons production rates (gHC/gcat/h), with catalyst particle size. As shown in this table, the FTS reaction rates passed from a maximum by decreasing the catalyst particle size. Also, Table 2 shows that the CO and H2 conversion increase by decreasing the catalyst particle size and passed from a maximum. It is found that the carbon number of produced hydrocarbon decreased with decreasing the catalyst particle size. Many of previous results reported the size dependence of FTS activity for iron and cobalt catalysts [35,37]. Table 3 lists the chain growth probabilities a1 and a2, the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) and hydrocarbons selectivity. As shown in this table, the chain growth probabilities a1 and a2 and the selectivity to heavier hydrocarbons decreased with decreasing the catalyst particle size. Also, the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) and the selectivity to lighter hydrocarbons increased with decreasing the catalyst particle size. Fig. 2 illustrates the variation of products distribution calculated by Donnelly model as a function of the catalyst particle size. This figure shows that by increasing the catalyst particle size,
the average carbon number of products and the break in the product distribution curve increased. The origin of the break is not established clearly. Ko¨nig and Gaube [10] and Huff and Satterfield [12] proposed that the two branches observed with potassium promoted iron catalysts are due to the synthesis over two groups of active sites differing in their promotion level. This interpretation may not be the only one since Huff and Satterfield [12] have shown that some unpromoted iron catalysts can also produce ASF plots, which consist of two branches. Dictor and Bell [38] suggested generally that deviations from uniform ASF distributions might result from the chemistry of the synthesis. The most accepted mechanism is CH2 insertion, which leads to conclude the carbide theory of Fischer–Tropsch [8]. In this mechanism, chain termination occurs mainly with hydrogen b-elimination and 1-alkenes being desorbed as primary products. In the following reactions, readsorbed 1-alkenes are hydrogenated to form intermediates for surface intermediates chain growth with C1 surface species. These C1 species have various hydrogenation degrees are as following: CO, HCO, HCOH, CH, and CH2 [39–42]. Patzlaff et al. [11] have discussed the distribution which is characterized with a1 is built up by monomers that exhibited higher degree of hydrogenation than those attributed with a2 ˙ Therefore, CH2 is assumed as C1 intermediate attributed to the distribution of growth probability a1 ˙ Obviously, hydrocarbon chains growth probability a2 are built up due to monomers except CH2 species with lower degree of hydrogenation. So, at higher hydrogen concentration on surface of catalyst the concentration of monomers with higher degree of hydrogenation increased. As shown in Table 1, the H2 concentration on catalyst surface decreased by increasing the catalyst particle size. Thus the concentration of monomers that exhibited higher degree of hydrogenation (like CH2 species) on the surface of catalyst increased with decreasing the catalyst particle size. The distribution which is characterized with a1, which is built up by monomers
Table 2 Feed conversion, RFTS and hydrocarbons production rates of catalysts. Catalyst particle size (nm)
28 24 21 16
Conversion (%)
RFTS (mol/gcat/h)
CO
H2
70.1 75.2 79.7 73.1
68.0 74.1 78.0 71.7
0.050 0.055 0.059 0.055
Hydrocarbons production rates (gHC/gcat/h) CH4
C2–C10
C11–C20
C21+
0.112 0.134 0.158 0.155
0.503 0.580 0.628 0.602
0.063 0.046 0.032 0.018
0.020 0.010 0.004 0.001
Reaction conditions: 563 K, H2/CO = 1, 17 bar, space velocity = 4.9 nl/gcat/h and time-on-stream 40 h.
Table 3 The chain growth probabilities a1 and a2, the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) and hydrocarbons selectivity. Catalyst particle size (nm)
a1
a2
m1
Hydrocarbons selectivity (%) CH4
C2–C4
C5–C10
C11–C19
C19+
28 24 21 16
0.72 0.68 0.65 0.61
0.93 0.90 0.85 0.82
0.71 0.80 0.83 0.87
14.7 16.0 17.7 18.4
44.5 49.4 52.5 56.1
31.8 29.4 27.1 24.0
5.6 3.5 2.1 1.2
3.4 1.6 0.6 0.3
Reaction conditions: 563 K, H2/CO = 1, 17 bar, space velocity = 4.9 nl/gcat/h and time-on-stream 40 h.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
5
Fig. 2. The variation of products distribution calculated by Donnelly model as a function of the catalyst particle size. Reaction conditions: 563 K, H2/CO = 1, 17 bar, space velocity = 4.9 nl/gcat./h and time-on-stream 40 h.
that exhibited higher degree of hydrogenation become more important and the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) increased. As shown in Table 3, m1 and the selectivity of lighter hydrocarbons increased with decreasing the catalyst particle size. 5. Conclusion The experimental results showed the FTS reaction rates on iron catalysts passed from a maximum by decreasing the catalyst particle size. It is found that the hydrocarbon produced by the catalysts becomes lighter with decreasing the catalyst particle size. Studies on chain length distributions of Fischer–Tropsch products have lead to the assumptions of two distinctive mechanism of chain growth that cause a superposition of two ASF distributions. The results indicate the H2 concentration on catalyst surface decreased by increasing the catalyst particle size. Thus the concentration of monomers that exhibited higher degree of hydrogenation (like CH2 species) on the surface of catalyst increased with decreasing the catalyst particle size. The total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) increased by decreasing of catalyst particle size. References [1] [2] [3] [4] [5]
B. Davis, Topics in Catalysis 32 (2005) 143. R. de Deugd, F. Kapteijn, J. Moulijn, Topics in Catalysis 26 (2003) 29. M.E. Dry, Catalysis Today 71 (2002) 227. M. Dry, Catalysis Letters 7 (1990) 241. M.E. Dry, in: S. Andre´, D. Mark (Eds.), FT Catalysts, Elsevier, 2004, p. 533 (Chapter 7). [6] A.P. Steynberg, M.E. Dry, B.H. Davis, B.B. Breman, in: S. Andre´, D. Mark (Eds.), Fischer–Tropsch Reactors, Elsevier, 2004, p. 64 (Chapter 2).
[7] Y.-N. Wang, W.-P. Ma, Y.-J. Lu, J. Yang, Y.-Y. Xu, H.-W. Xiang, Y.-W. Li, Y.-L. Zhao, B.J. Zhang, Fuel 82 (2003) 195. [8] G.P. Van Der Laan, A.A.C.M. Beenackers, Catalysis Reviews 41 (1999) 255. [9] J. Gaube, K. Herzog, L. Ko¨nig, B. Schliebs, Chemie Ingenieur Technik 58 (1986) 682. [10] L. Ko¨nig, J. Gaube, Chemie Ingenieur Technik 55 (1983) 14. [11] J. Patzlaff, Y. Liu, C. Graffmann, J. Gaube, Applied Catalysis A: General 186 (1999) 109. [12] G.A. Huff Jr., C.N. Satterfield, Journal of Catalysis 85 (1984) 370. [13] C.N. Satterfield, G.A. Huff Jr., R. Summerhayes, Journal of Catalysis 80 (1983) 486. [14] T.J. Donnelly, I.C. Yates, C.N. Satterfield, Energy and Fuels 2 (1988) 734. [15] R.J. Madon, W.F. Taylor, Journal of Catalysis 69 (1981) 32. [16] J. Wang, M.S. Gudiksen, X. Duan, Y. Cui, C.M. Lieber, Science 293 (2001) 1455. [17] A.T. Bell, Science 299 (2003) 1688. [18] T. Bernhardt, U. Heiz, U. Landman, in: U. Heiz, U. Landman (Eds.), Chemical and Catalytic Properties of Size-Selected Free and Supported Clusters, Springer, Berlin Heidelberg, 2007, p. 1. [19] C.Q. Sun, Progress in Solid State Chemistry 35 (2007) 1. [20] A.L. Dantas Ramos, P.D.S. Alves, D.A.G. Aranda, M. Schmal, Applied Catalysis A: General 277 (2004) 71. [21] A.N. Pour, Y. Zamani, A. Tavasoli, S.M. Kamali Shahri, S.A. Taheri, Fuel 87 (2008) 2004. [22] A. Nakhaei Pour, S.M.K. Shahri, H.R. Bozorgzadeh, Y. Zamani, A. Tavasoli, M.A. Marvast, Applied Catalysis A: General 348 (2008) 201. [23] A.N. Pour, M. Zare, Y. Zamani, Journal of Natural Gas Chemistry 19 (2010) 31. [24] A.N. Pour, S. Taghipoor, M. Shekarriz, S.M.K. Shahri, Y. Zamani, Journal of Nanoscience and Nanotechnology 9 (2009) 4425. [25] A. Nakhaei Pour, M.R. Housaindokht, S.F. Tayyari, J. Zarkesh, M.R. Alaei, Journal of Molecular Catalysis A: Chemical 330 (2010) 112. [26] A.N. Pour, M.R. Housaindokht, S.M.K. Shahri, E.G. Babakhani, M. Irani, Journal of Industrial and Engineering Chemistry 17 (2011) 596. [27] A.N. Pour, M.R. Housaindokht, S.F. Tayyari, J. Zarkesh, Journal of Natural Gas Chemistry 19 (2010) 284. [28] A.N. Pour, M.R. Housaindokht, S.F. Tayyari, J. Zarkesh, Journal of Natural Gas Chemistry 19 (2010) 333. [29] A.N. Pour, M.R. Housaindokht, J. Zarkesh, M. Irani, E.G. Babakhani, Journal of Industrial and Engineering Chemistry 18 (2012) 597. [30] M. Ding, Y. Yang, B. Wu, T. Wang, H. Xiang, Y. Li, Fuel Processing Technology 92 (2011) 2353. [31] J.W. Niemantsverdriet, A.M. Van der Kraan, W.L. Van Dijk, H.S. Van der Baan, Journal of Physical Chemistry 84 (1980) 3363. [32] S. Li, A. Li, S. Krishnamoorthy, E. Iglesia, Catalysis Letters 77 (2001) 197.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 6
A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
[33] Z. Tao, Y. Yang, C. Zhang, T. Li, J. Wang, H. Wan, H. Xiang, Y. Li, Catalysis Communications 7 (2006) 1061. [34] J. Yang, Y. Sun, Y. Tang, Y. Liu, H. Wang, L. Tian, H. Wang, Z. Zhang, H. Xiang, Y. Li, Journal of Molecular Catalysis A: Chemical 245 (2006) 26. [35] M. Tre´panier, A.K. Dalai, N. Abatzoglou, Applied Catalysis A: General 374 (2010) 79. [36] C.-C. Wang, D.-H. Chen, T.-C. Huang, Colloids and Surfaces A: Physicochemical and Engineering Aspects 189 (2001) 145.
[37] D. Murzin, Reaction Kinetics and Catalysis Letters 97 (2009) 165. [38] R.A. Dictor, A.T. Bell, Journal of Catalysis 97 (1986) 121. [39] M. Ojeda, R. Nabar, A.U. Nilekar, A. Ishikawa, M. Mavrikakis, E. Iglesia, Journal of Catalysis 272 (2010) 287. [40] B.H. Davis, Catalysis Today 141 (2009) 25. [41] J. Gaube, H.F. Klein, Applied Catalysis A: General 374 (2010) 120. [42] J. Gaube, H.F. Klein, Journal of Molecular Catalysis A: Chemical 283 (2008) 60.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
JIEC-1369; No. of Pages 6 Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Journal of Industrial and Engineering Chemistry journal homepage: www.elsevier.com/locate/jiec
Studies on product distribution of nanostructured iron catalyst in Fischer–Tropsch synthesis: Effect of catalyst particle size Ali Nakhaei Pour *, Mohammad Reza Housaindokht Department of Chemistry, Ferdowsi University of Mashhad, P.O. Box 9177948974, Mashhad, Iran
A R T I C L E I N F O
A B S T R A C T
Article history: Received 24 March 2013 Accepted 12 May 2013 Available online xxx
The dependencies of hydrocarbon product distributions of Fischer–Tropsch synthesis by iron catalysts on catalysts particle size are studied. The concept of two superimposed Anderson–Schulz–Flory distributions applied for represent size dependency of product distributions. A series of catalysts with different particle size are prepared by microemulsion method. It is found that the carbon number of produced hydrocarbon decreased with decreasing the catalyst particle size. These results indicate the H2 concentration on catalyst surface decreased by increasing the catalyst particle size. Thus the concentration of monomers that exhibited higher degree of hydrogenation (like CH2 species) on the surface of catalyst increased with decreasing the catalyst particle size. ß 2013 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
Keywords: Chain length distribution Fischer–Tropsch synthesis Iron catalyst Particle size
1. Introduction Strong interest in the Fischer–Tropsch synthesis (FTS) that converts synthesis gas into hydrocarbons is reappearing because it is one of the major routes that convert natural gas into liquid energy carriers [1–3]. The Fischer–Tropsch synthesis (FTS) consisting of a complex multi component mixture of linear and branched hydrocarbons and oxygenated products. Main products are linear paraffins and a-olefins. Herington first treated the molar distribution of hydrocarbons from FTS in terms of a polymerization mechanism [4–6]. The same formulation was rediscovered by Anderson et al. in 1951 and named the Anderson–Schulz–Flory (ASF) distribution [4,5,7,8]. In the ASF model, the formation of hydrocarbon chains was assumed as a stepwise polymerization procedure and the chain growth probability was assumed to be independent of the carbon number. For all FTS catalysts deviations from the ideal Anderson–Schulz–Flory (ASF) distributions are observed in many studies [9–11]. The usual deviations of the distribution of the linear hydrocarbons are a relatively higher selectivity to methane, a relatively lower selectivity to ethene, and an increase in the chain growth probability with increasing molecular size. Some authors interpreted the deviations from the standard ASF distribution by the superposition of two ASF distributions [10,12–14]. They suspected the existence of two sorts of sites for the chain growth on the catalyst surface and each
* Corresponding author. Tel.: +98 5118795457; fax: +98 5118795457. E-mail addresses: [email protected], [email protected] (A. Nakhaei Pour).
site might individually yield the ideal ASF distribution with different chain growth probabilities. Madon and Taylor [15] interpreted this bimodal distribution by differently sites causing different growth probabilities. Ko¨nig and Gaube [10] supposed that in the case of alkali promoted iron catalysts, all active sites are not influenced by alkali. Therefore the distribution of higher growth probably was attributed to products formed on promoted active sites while the other part of the product is formed on unpromoted sites with a similar growth probability as evaluated for the product formed on unprompted iron catalysts. Patzlaff et al. [11] studied the effects of 1-alkene readsorption and secondary chain growth on the product distribution of FTS on iron and cobalt catalyst by cofeeding 1-alkenes. They found that chain length distributions of products obtained on cobalt catalysts are slightly modified by secondary chain growth of readsorbed alkenes and hydrogenolysis of hydrocarbons. The FTS reaction is a surface phenomenon, therefore for optimum catalyst performance, maximum metal usage must be achieved. A rule of thumb in heterogeneous catalysis is that smaller metal crystallites provide largest surface area on which the reaction may happen [16–20]. In our previous work the effect of presence of the zeolite on product distributions of iron catalyst studied [21]. In the other our previous work the effect of Ca, Mg and La promoters on physico-chemical properties and the catalytic activity and selectivity during FTS performance studied [22,23]. In this experiment the effect of catalyst particle size on the carbon number distribution of Fischer–Tropsch products on iron catalysts are studied, using a modified Anderson–Schulz–Flory (ASF) distribution with two chain growth probabilities.
1226-086X/$ – see front matter ß 2013 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jiec.2013.05.019
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 2
A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
2. Experimental 2.1. Catalyst preparation and characterization Fe–Cu nanoparticles were prepared by coprecipitation in a water-in-oil microemulsion [24–28]. The precipitation was performed in the single-phase microemulsion operating region. In order to achieve a series of catalysts displaying different hematite particle size the surfactant-to-oil (S/O) weight ratio was set to a value of 0.3 the water-to-surfactant molar ratio (W/S) was varied from 4 to 12. 2.2. Catalyst characterization The surface area was calculated from the Brunauer–Emmett– Teller (BET) equation and pore volume, average pore diameter, and pore size distribution of the catalysts were determined by N2 physisorption using a Micromeritics ASAP 2010 automated system. A 0.5 g catalyst sample was degassed at 373 K for1 h and 573 K for 2 h prior to analysis. The analysis was done using N2 adsorption at 77 K. Both the pore volume and the average pore diameter were calculated by Barret–Joyner–Hallender (BJH) method from the adsorption isotherm. XRD was used to determine the phase composition of catalysts before and after pretreatments. The XRD spectrum of the catalysts were collected by a X-ray diffractometer, Philips PW1840 X-ray diffractometer, using monochromatized Cu/Ka radiation (40 kV, 40 mA) and a step scan mode at a scan rate of 0.028 (2u) per second from 108 to 808. XRD peak identification was recognized by comparison to the JCPDS database software. The average crystallite size of samples, dXRD, can be estimated from XRD patterns by applying full-width half-maximum (FWHM) of characteristic peak (1 0 4) Fe2O3 located at 2u = 33.38 peak to Scherrer equation: dXRD ¼
0:9 l FWHM cos u
with the same mesh size range. The catalyst samples were activated by a 5% (v/v) H2/N2 gas mixture with space velocity equal to 15.1 nl/gcat/h at 1 bar and 1800 rpm. The reactor temperature increased to 673 K with a heating rate of 5 K/min, maintained for 1 h at this temperature, and then reduced to 543 K. The activation is followed by the synthesis gas stream with H2/CO ratio of 1 and space velocity equal to 3.07 nl/gcat/h for 24 h at 1 bar and 543 K. After catalyst activation, synthesis gas was fed to the reactor at conditions operated at 563 K, 17 bar, H2/CO ratio of 1 and a space velocity equal to 4.9 nl/gcat/h. After reaching steady state, the FTS reaction rate was measured. The out gas was analyzed by a gas chromatograph (Varian CP3800) equipped with TCD and FID detectors. The CO, CO2, N2, and O2 were analyzed through two packed column in series (Molecular sieve13 CP 81025 with 2 m length, and 3 mm OD, and Hayesep Q CP1069 with 4 m length, and 3 mm OD) connected to TCD detector. The C1–C5 hydrocarbons were analyzed via a capillary column (CP fused silica with 25 m 0.25 mm 0.2 mm film thickness) connected to FID detector. Hydrogen was analyzed through Shimatzu, GC PTF 4C, equipped with TCD detector and two column in series (Propack-Q with 2 m length, and 3 mm OD for CO2, C2H4 and C2H6 separation and molecular sieve-5A with 2 m length, and 3 mmOD for CO, N2, CH4 and O2 separation), which were connected to each other via a three way valve. The collected liquid (including hydrocarbons and oxygenates) analyzed offline with Varian CP-3800 gas chromatograph equipped with capillary column (TM DH fused silica capillary column, PETRO COL 100 m 0.25 mm 0.5 mm film thickness) connected to FID detector. Total mass balances were performed with the carbon material balance closed between 97 and 103%. This criterion was adopted since compounds containing carbon and hydrogen might accumulate in the reactor, in the form of high molecular weight hydrocarbons.
(1)
where l is the X-ray wavelength (1.5406 A˚ in this study) and u is the diffraction angle for the (1 0 4) plane. The H2-TPD experiments were performed by means of the temperature-programmed desorption (TPD) of H2 on the catalyst (0.5 g), which was packed in a shallow-bed quartz reactor with a low dead volume from 300 K to 1100 K at a linear heating rate of 5 K/min while Ar was used as a carrier gas. A thermal conductivity detector (TCD) was used to measure the H2 desorbed in the TPD quantitatively. The catalyst was first reduced with H2 at 673 K and 1 bar for 11 h. Then the sample was heated in Ar from 323 K to 673 K, held at 673 K until the baseline leveled off (ensuring complete removal of adsorbed species on the reduced catalyst surface has been achieved), and finally cooled to 323 K for TPD tests. In the subsequent steps, H2 adsorption on the catalyst was performed at 323 K for 30 min, and then the sample was purged with Ar in order that weakly adsorbed species could be removed until the baseline leveled off. Following this, H2-TPD was being carried out while the temperature increased to 1050 K. H2 chemisorptions uptakes determined by integrating the area of H2-TPD curves as compared to the certain amounts of gas passed through the TCD. 2.3. Experimental apparatus and procedure Steady-state FTS reaction rates measured in a continuous spinning basket reactor. A detailed description of the experimental setup and procedures has been provided in our previous works [29]. The fresh catalyst is crushed and sieved to particles with the diameter of 0.25–0.36 mm (40–60 ASTM mesh). The weight of the catalyst loaded was 2.5 g and diluted by 30 cm3 inert silica sand
3. Data analysis Hydrocarbon products of the Fischer–Tropsch synthesis are generally taken to follow the ASF distribution. For carbon number i, the mole fraction of product xi as determined by a single chain growth probability a is given by: xi ¼ ð1 aÞai1
(2)
In this work the carbon number distribution of Fischer–Tropsch products on iron catalysts studied by modified Anderson–Schulz– Flory (ASF) distribution with two chain growth probabilities. This method proposed by Donnelly et al. [14] to characterize the carbon number distribution of FTS hydrocarbon products by ASF distributions with different chain growth probabilities. þ B ai1 xi ¼ A ai1 1 2
(3)
Note that at the break point on ASF diagram, the contributions of each term in Eq. (3) are equal: ¼ B ai1 A ai1 1 2
i¼z
(4)
In break point, i illustrate as z and necessarily is not an integral carbon number. In this model A and B were not correspond directly to the fractions of products produced from a1 and a2, respectively. Instead the sum of the mole fractions over all carbon numbers is unity: 1 1 X X xi ¼ ½A a1 ði1Þ þ B a2 ði1Þ ¼ 1 i¼1
(5)
i¼1
Methane and ethene do not obey the ASF equation and after removing C1 and C2 products to fit theoretical distributions to data
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
3
leads to: 1 1 X X ði1Þ ði1Þ xi ¼ ½A a1 þ B a2 Að1 þ a1 Þ Bð1 þ a2 Þ i¼3
i¼1
xexp ¼ 1 xexp 1 2
(6)
Therefore, the determination of growth probabilities of the bimodal ASF distribution is based on hydrocarbons with carbon numbers i > 2. The bimodal distribution is characterized by two growth probabilities (a1 and a2) and the point of intersection j of both ASF distributions. The mathematical procedure given by Donnelly et al. [14] arrives at: " A ¼ ð1 xexp xexp 1 2 Þ
a12 a1 þ 1 a1 a2
j1 #1 a22 1 a2
(7)
and finally at: "
xi 1 xexp xexp 1 2
¼ ai1 þ 1
a1 a2
"
j1
a21 a1 þ 1 a1 a2
# ai1 2
j1 #1 a22 1 a2
(8)
The three parameters a1, a2 and j are evaluated by minimizing the least squares:
F¼
I X 2 ðln x1 ln xexp Þ ¼ min l
(9)
i¼3
in which I is the upper limit of experimental data. The advantage of Donnelly method is its’ independence of this upper limit I. However, the intersection point j is a formal parameter without any physicochemical relevance. Also this method has been extended by the calculation of m1 the total mole fraction of hydrocarbons formed with the growth probability a1. 1 X A ai1 1
m1 ¼
i¼3
1 xexp xexp 1 2
1 X A ai1 Að1 þ a1 Þ 1
¼
i¼1
1 xexp xexp 1 2
(10)
Using the limiting value of the geometric series 1 X
ai1 ¼ ð1 aÞ1
(11)
i¼1
gives
m1 ¼
A a21 ð1 a1 Þð1 xexp xexp 1 2 Þ
(12)
and with Eq. (7)
m1 ¼
ð1 a2 Þ=ð1 a1 Þa21 ½ð1 a2 Þ=ð1 a1 Þa21 þ ðða1 Þ=ða2 ÞÞj1 a22
(13)
Since two independent ASF distributions are superimposed the fraction of hydrocarbons formed with a chain growth probability a2 equals:
m2 ¼ 1 m1
(14)
4. Results and discussion 4.1. Catalyst characterization Nanostructure iron catalysts characterized by X-ray diffraction (XRD) measurement after calcination. Fig. 1 shows the XRD patterns of iron catalysts, the characteristic peaks corresponding to
Fig. 1. XRD pattern of the prepared catalysts after calcinations. (a) W/S = 4, (b) W/ S = 6, (c) W/S = 8 and (d) W/S = 12.
(0 1 2), (1 0 4), (1 1 0), (1 1 3), (0 2 4), (1 1 6), (0 1 8), (2 1 4), (3 0 0) planes are located at 2u = 24.38, 33.38, 35.88, 40.88, 49.68, 54.18, 57.68, 64.18 and 65.68, respectively [30–34]. They show very close to the ones with cubic hematite structured Fe2O3 crystal in JCPDS database. It shows that the hematite structure once formed remains stable during subsequent aqueous impregnation and thermal treatment. This strongly infers that microemulsion system modules only physical properties of reaction medium without changing the reaction paths and arrangements of crystal structure. The characteristic peak at 2u = 33.38 corresponds to the hematite 104 plane was used to calculate the average metal particle size by the Scherrer equation. The calculated dXRD for the samples listed in Table 1. The H2-TPD results, BET surface area and textural properties and pore size of the fresh iron-based catalysts are shown in Table 1. As shown in Table 1, the average particle sizes of hematite are linearly correlated with the water-to-surfactant ratio used during the microemulsion catalyst preparation route. In fact, nanoparticles are formed in the internal structure of the microemulsion, which is determined by the ratio of water-to surfactant. At high oil concentration, the bicontinuous phase is transformed into a structure of small water droplets within a continuous oil phase (reverse micelles) when surfactant is added. Thus, the results show that the size of different droplets determines the cobalt’s particle size, depending on the amount of surfactant [35,36]. As shown in this table, surface area and particle size of catalysts are changed with W/S ratio in microemulsion system. When the total surfactants-to-oil phase weight ratio was constant, the particle size decreased and surface area increased slightly with the increase in the water content. Also Table 1 shows the H2 adsorbed on surface of catalysts increased by decreasing of catalyst particle size.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
4
Table 1 H2 chemisorption, BET surface area and textural properties of prepared catalysts. W/S
Catalyst particle size (nm)a
BET surface area (m2/g)
Average pore size (nm)b
Pore volume (cm3/g)b
H2 uptake (102 mmol H2/mmol Fe)
4 6 8 12
16 21 24 28
68.2 56.4 48.2 43.5
16.2 19.6 22.3 25.9
0.14 0.16 0.19 0.21
7.8 7.1 6.2 5.7
a b
Calculated from XRD results. These values were calculated by BJH method from desorption isotherm.
4.2. Catalyst activity and products distribution The FTS reaction yields organic compounds and the water and/ or carbon dioxide as by-products. Thus, carbon monoxide can be consumed for the formation of organic compounds or carbon dioxide. Therefore, the rate of FTS reaction must be written as: r FTS ¼ r CO r CO2
(15)
where r CO2 is the rate of CO2 formation (water–gas shift reaction) and rCO is the rate of CO consumption. The FTS rate equals the rate of formation of organic compounds on carbon basis. The experimental results of the effects of catalyst particle size on FTS reaction rates were listed in Table 2. Table 2 shows the change of syngas conversions, FTS reaction rates and hydrocarbons production rates (gHC/gcat/h), with catalyst particle size. As shown in this table, the FTS reaction rates passed from a maximum by decreasing the catalyst particle size. Also, Table 2 shows that the CO and H2 conversion increase by decreasing the catalyst particle size and passed from a maximum. It is found that the carbon number of produced hydrocarbon decreased with decreasing the catalyst particle size. Many of previous results reported the size dependence of FTS activity for iron and cobalt catalysts [35,37]. Table 3 lists the chain growth probabilities a1 and a2, the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) and hydrocarbons selectivity. As shown in this table, the chain growth probabilities a1 and a2 and the selectivity to heavier hydrocarbons decreased with decreasing the catalyst particle size. Also, the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) and the selectivity to lighter hydrocarbons increased with decreasing the catalyst particle size. Fig. 2 illustrates the variation of products distribution calculated by Donnelly model as a function of the catalyst particle size. This figure shows that by increasing the catalyst particle size,
the average carbon number of products and the break in the product distribution curve increased. The origin of the break is not established clearly. Ko¨nig and Gaube [10] and Huff and Satterfield [12] proposed that the two branches observed with potassium promoted iron catalysts are due to the synthesis over two groups of active sites differing in their promotion level. This interpretation may not be the only one since Huff and Satterfield [12] have shown that some unpromoted iron catalysts can also produce ASF plots, which consist of two branches. Dictor and Bell [38] suggested generally that deviations from uniform ASF distributions might result from the chemistry of the synthesis. The most accepted mechanism is CH2 insertion, which leads to conclude the carbide theory of Fischer–Tropsch [8]. In this mechanism, chain termination occurs mainly with hydrogen b-elimination and 1-alkenes being desorbed as primary products. In the following reactions, readsorbed 1-alkenes are hydrogenated to form intermediates for surface intermediates chain growth with C1 surface species. These C1 species have various hydrogenation degrees are as following: CO, HCO, HCOH, CH, and CH2 [39–42]. Patzlaff et al. [11] have discussed the distribution which is characterized with a1 is built up by monomers that exhibited higher degree of hydrogenation than those attributed with a2 ˙ Therefore, CH2 is assumed as C1 intermediate attributed to the distribution of growth probability a1 ˙ Obviously, hydrocarbon chains growth probability a2 are built up due to monomers except CH2 species with lower degree of hydrogenation. So, at higher hydrogen concentration on surface of catalyst the concentration of monomers with higher degree of hydrogenation increased. As shown in Table 1, the H2 concentration on catalyst surface decreased by increasing the catalyst particle size. Thus the concentration of monomers that exhibited higher degree of hydrogenation (like CH2 species) on the surface of catalyst increased with decreasing the catalyst particle size. The distribution which is characterized with a1, which is built up by monomers
Table 2 Feed conversion, RFTS and hydrocarbons production rates of catalysts. Catalyst particle size (nm)
28 24 21 16
Conversion (%)
RFTS (mol/gcat/h)
CO
H2
70.1 75.2 79.7 73.1
68.0 74.1 78.0 71.7
0.050 0.055 0.059 0.055
Hydrocarbons production rates (gHC/gcat/h) CH4
C2–C10
C11–C20
C21+
0.112 0.134 0.158 0.155
0.503 0.580 0.628 0.602
0.063 0.046 0.032 0.018
0.020 0.010 0.004 0.001
Reaction conditions: 563 K, H2/CO = 1, 17 bar, space velocity = 4.9 nl/gcat/h and time-on-stream 40 h.
Table 3 The chain growth probabilities a1 and a2, the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) and hydrocarbons selectivity. Catalyst particle size (nm)
a1
a2
m1
Hydrocarbons selectivity (%) CH4
C2–C4
C5–C10
C11–C19
C19+
28 24 21 16
0.72 0.68 0.65 0.61
0.93 0.90 0.85 0.82
0.71 0.80 0.83 0.87
14.7 16.0 17.7 18.4
44.5 49.4 52.5 56.1
31.8 29.4 27.1 24.0
5.6 3.5 2.1 1.2
3.4 1.6 0.6 0.3
Reaction conditions: 563 K, H2/CO = 1, 17 bar, space velocity = 4.9 nl/gcat/h and time-on-stream 40 h.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
5
Fig. 2. The variation of products distribution calculated by Donnelly model as a function of the catalyst particle size. Reaction conditions: 563 K, H2/CO = 1, 17 bar, space velocity = 4.9 nl/gcat./h and time-on-stream 40 h.
that exhibited higher degree of hydrogenation become more important and the total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) increased. As shown in Table 3, m1 and the selectivity of lighter hydrocarbons increased with decreasing the catalyst particle size. 5. Conclusion The experimental results showed the FTS reaction rates on iron catalysts passed from a maximum by decreasing the catalyst particle size. It is found that the hydrocarbon produced by the catalysts becomes lighter with decreasing the catalyst particle size. Studies on chain length distributions of Fischer–Tropsch products have lead to the assumptions of two distinctive mechanism of chain growth that cause a superposition of two ASF distributions. The results indicate the H2 concentration on catalyst surface decreased by increasing the catalyst particle size. Thus the concentration of monomers that exhibited higher degree of hydrogenation (like CH2 species) on the surface of catalyst increased with decreasing the catalyst particle size. The total mole fraction of hydrocarbons formed with the chain growth probability a1 (m1) increased by decreasing of catalyst particle size. References [1] [2] [3] [4] [5]
B. Davis, Topics in Catalysis 32 (2005) 143. R. de Deugd, F. Kapteijn, J. Moulijn, Topics in Catalysis 26 (2003) 29. M.E. Dry, Catalysis Today 71 (2002) 227. M. Dry, Catalysis Letters 7 (1990) 241. M.E. Dry, in: S. Andre´, D. Mark (Eds.), FT Catalysts, Elsevier, 2004, p. 533 (Chapter 7). [6] A.P. Steynberg, M.E. Dry, B.H. Davis, B.B. Breman, in: S. Andre´, D. Mark (Eds.), Fischer–Tropsch Reactors, Elsevier, 2004, p. 64 (Chapter 2).
[7] Y.-N. Wang, W.-P. Ma, Y.-J. Lu, J. Yang, Y.-Y. Xu, H.-W. Xiang, Y.-W. Li, Y.-L. Zhao, B.J. Zhang, Fuel 82 (2003) 195. [8] G.P. Van Der Laan, A.A.C.M. Beenackers, Catalysis Reviews 41 (1999) 255. [9] J. Gaube, K. Herzog, L. Ko¨nig, B. Schliebs, Chemie Ingenieur Technik 58 (1986) 682. [10] L. Ko¨nig, J. Gaube, Chemie Ingenieur Technik 55 (1983) 14. [11] J. Patzlaff, Y. Liu, C. Graffmann, J. Gaube, Applied Catalysis A: General 186 (1999) 109. [12] G.A. Huff Jr., C.N. Satterfield, Journal of Catalysis 85 (1984) 370. [13] C.N. Satterfield, G.A. Huff Jr., R. Summerhayes, Journal of Catalysis 80 (1983) 486. [14] T.J. Donnelly, I.C. Yates, C.N. Satterfield, Energy and Fuels 2 (1988) 734. [15] R.J. Madon, W.F. Taylor, Journal of Catalysis 69 (1981) 32. [16] J. Wang, M.S. Gudiksen, X. Duan, Y. Cui, C.M. Lieber, Science 293 (2001) 1455. [17] A.T. Bell, Science 299 (2003) 1688. [18] T. Bernhardt, U. Heiz, U. Landman, in: U. Heiz, U. Landman (Eds.), Chemical and Catalytic Properties of Size-Selected Free and Supported Clusters, Springer, Berlin Heidelberg, 2007, p. 1. [19] C.Q. Sun, Progress in Solid State Chemistry 35 (2007) 1. [20] A.L. Dantas Ramos, P.D.S. Alves, D.A.G. Aranda, M. Schmal, Applied Catalysis A: General 277 (2004) 71. [21] A.N. Pour, Y. Zamani, A. Tavasoli, S.M. Kamali Shahri, S.A. Taheri, Fuel 87 (2008) 2004. [22] A. Nakhaei Pour, S.M.K. Shahri, H.R. Bozorgzadeh, Y. Zamani, A. Tavasoli, M.A. Marvast, Applied Catalysis A: General 348 (2008) 201. [23] A.N. Pour, M. Zare, Y. Zamani, Journal of Natural Gas Chemistry 19 (2010) 31. [24] A.N. Pour, S. Taghipoor, M. Shekarriz, S.M.K. Shahri, Y. Zamani, Journal of Nanoscience and Nanotechnology 9 (2009) 4425. [25] A. Nakhaei Pour, M.R. Housaindokht, S.F. Tayyari, J. Zarkesh, M.R. Alaei, Journal of Molecular Catalysis A: Chemical 330 (2010) 112. [26] A.N. Pour, M.R. Housaindokht, S.M.K. Shahri, E.G. Babakhani, M. Irani, Journal of Industrial and Engineering Chemistry 17 (2011) 596. [27] A.N. Pour, M.R. Housaindokht, S.F. Tayyari, J. Zarkesh, Journal of Natural Gas Chemistry 19 (2010) 284. [28] A.N. Pour, M.R. Housaindokht, S.F. Tayyari, J. Zarkesh, Journal of Natural Gas Chemistry 19 (2010) 333. [29] A.N. Pour, M.R. Housaindokht, J. Zarkesh, M. Irani, E.G. Babakhani, Journal of Industrial and Engineering Chemistry 18 (2012) 597. [30] M. Ding, Y. Yang, B. Wu, T. Wang, H. Xiang, Y. Li, Fuel Processing Technology 92 (2011) 2353. [31] J.W. Niemantsverdriet, A.M. Van der Kraan, W.L. Van Dijk, H.S. Van der Baan, Journal of Physical Chemistry 84 (1980) 3363. [32] S. Li, A. Li, S. Krishnamoorthy, E. Iglesia, Catalysis Letters 77 (2001) 197.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019
G Model
JIEC-1369; No. of Pages 6 6
A. Nakhaei Pour, M.R. Housaindokht / Journal of Industrial and Engineering Chemistry xxx (2013) xxx–xxx
[33] Z. Tao, Y. Yang, C. Zhang, T. Li, J. Wang, H. Wan, H. Xiang, Y. Li, Catalysis Communications 7 (2006) 1061. [34] J. Yang, Y. Sun, Y. Tang, Y. Liu, H. Wang, L. Tian, H. Wang, Z. Zhang, H. Xiang, Y. Li, Journal of Molecular Catalysis A: Chemical 245 (2006) 26. [35] M. Tre´panier, A.K. Dalai, N. Abatzoglou, Applied Catalysis A: General 374 (2010) 79. [36] C.-C. Wang, D.-H. Chen, T.-C. Huang, Colloids and Surfaces A: Physicochemical and Engineering Aspects 189 (2001) 145.
[37] D. Murzin, Reaction Kinetics and Catalysis Letters 97 (2009) 165. [38] R.A. Dictor, A.T. Bell, Journal of Catalysis 97 (1986) 121. [39] M. Ojeda, R. Nabar, A.U. Nilekar, A. Ishikawa, M. Mavrikakis, E. Iglesia, Journal of Catalysis 272 (2010) 287. [40] B.H. Davis, Catalysis Today 141 (2009) 25. [41] J. Gaube, H.F. Klein, Applied Catalysis A: General 374 (2010) 120. [42] J. Gaube, H.F. Klein, Journal of Molecular Catalysis A: Chemical 283 (2008) 60.
Please cite this article in press as: A. Nakhaei Pour, M.R. Housaindokht, J. Ind. Eng. Chem. (2013), http://dx.doi.org/10.1016/ j.jiec.2013.05.019