ARTICLE
Genesis of Acid-Base Support Properties with Variations of Preparation Conditions: Cumene Cracking and Its Kinetics Fernando Trejo,*,† Mohan S. Rana,‡ and Jorge Ancheyta§ †
Centro de Investigacion en Ciencia Aplicada y Tecnología Avanzada Unidad Legaria del Instituto Politecnico Nacional (CICATA-IPN), Legaria 694, Col. Irrigacion, Mexico DF 11500, Mexico ‡ Petroleum Refining Department (PRD), Kuwait Institute for Scientific Research, P.O. Box 24885, Safat 13109, Kuwait § Instituto Mexicano del Petroleo, Eje Central Lazaro Cardenas Norte 152, Col. San Bartolo Atepehuacan, Mexico DF 07730, Mexico ABSTRACT: Mixed oxides (Al-Si, Zr-Si, Mg-Si) were synthesized by the sol-gel method at different gelation times. The synthesized supports were impregnated by the incipient wetness method to obtain Ni-Mo, Co-Mo, and Ni-W supported catalysts. Both the supports and catalysts were characterized by N2 physisorption for determining textural properties, and their catalytic activities were tested with isopropyl alcohol (2-propanol) and cumene decomposition in order to evaluate their acid-base properties. According to the obtained products, it was possible to assess qualitatively that Al-Si and Zr-Si supports and their respective catalysts were mainly acidic. In the case of NiMo/AlSi catalyst, high conversion during cumene cracking particularly at higher temperatures was observed. Cracking activity results indicated that surface acidity predominated due to Brønsted acid sites, which are mainly dependent on the support composition as well as their preparation methods. Conversion of isopropyl alcohol leads to the dehydration products, typically propylene with a very low amount of di-isopropyl ether or acetone. NiMo/AlSi catalyst was evaluated for cumene cracking and its kinetic parameters were obtained by applying the Langmuir-Hinshelwood model. The activation energy was found to be 27.75 kcal mol-1 while the heat of adsorption calculated with the Van’t Hoff equation was -14.15 kcal mol-1 taking into account a single-site surface reaction.
1. INTRODUCTION The demand for stringent environmental regulations of clean fuels is focusing the attention of researchers as well as refiners to the hydrotreatment of various petroleum fractions using new catalytic materials and the development of new and more efficient processes.1-3 The use of mixed oxides (SiO2-Al2O3, SiO2ZrO2, SiO2-TiO2) as supports has become very attractive with respect to their textural properties and interaction with active metals.4,5 Oxides used as catalysts have the ability to take part in the exchange of electrons or protons in both redox and acidbase catalysis.6 In metal oxides, coordinative unsaturation is principally responsible for the adsorption and catalysis of various reactions. The exposed cations and anions of the metal oxide surfaces form acid and basic sites as well as acid-base pairs. The catalytic activity of acidic solids is not only related to the surface concentration of acid sites, but also depends on their nature, that is, being Brønsted or Lewis sites.7 Amorphous silica-based mixed metal oxides such as silicaalumina and silica-zirconia are widely used as acidic catalysts, and their properties depend on the preparation procedure. Silicabased amorphous mixed oxides supports have been studied, and their preparation methods are reported in the literature.8-10 Usually, oxide properties depend on the preparation procedure which affects textural properties mainly. The sol-gel method is one of the most important applications available in heterogeneous catalysis. Generally, supports obtained by using the solgel method have high porosity, large specific surface area, and acid sites that make them very attractive.6 This method has attracted great interest in the past few years since it is convenient to prepare supports where heteropoly condensation of different metals is responsible for the formation of metal-O-Si (M-O-Si) r 2010 American Chemical Society
strong bonds, which control the support particle size and distribution of metal particles in the solid matrix.11 This method involves hydrolysis of an alkoxide followed by a condensation reaction to form a gel,12 which is obtained by aging, drying, and calcination. It has been reported that bimodal catalysts (meso and macroporous) can be obtained with the use of tetraethyl orthosilicate (TEOS) and polyethylene oxide (PEO) by spinodal decomposition.13 A common way to evaluate acidic properties of supports/ catalysts is during cumene cracking which has been reported in the literature to be a reaction for the simultaneous determination of Brønsted as well as Lewis acidity; however, it has been also stated that acidity of catalysts is modified during sulfidation which may also influence the catalyst activity. The cracking of alkylbenzenes is catalyzed by Brønsted acid sites, particularly during conversion of cumene into benzene and propylene which is a well-known reaction.14 Cumene cracking is enhanced by the acidity of mixed-oxide supports where the Mo-supported catalysts are more active than the support only. This is due to the presence of -SH groups on the catalysts and the exchange protons from the molybdenum sulfide phase as reported by Topsøe.15 Sarback16 studied the cumene decomposition over alumina modified with fluoride ions obtaining benzene and propylene as products, whereas alumina modified with sodium ions yielded R-methyl Special Issue: IMCCRE 2010 Received: April 2, 2010 Accepted: June 15, 2010 Revised: June 14, 2010 Published: July 8, 2010 2715
dx.doi.org/10.1021/ie1008037 | Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
Figure 1. General scheme of the cumene cracking reaction.
styrene. Alumina by itself is not active toward cumene dealkylation. The addition of fluoride on alumina supports and its effect on cobalt molybdenum catalysts have been well documented;17,18 however, in this case, only oxidic catalysts were studied. It has been suggested that cumene dealkylation is carried out on Brønsted acid sites where the molecule gains protons to form π complexes to be then transformed into σ complexes.19 The deposition of molybdenum on the support increases the amount of Brønsted acid sites associated with the sulfided molybdenum species. The major reactions that occur during cumene decomposition may be classified as dealkylation (cracking) and dehydrogenation. Dealkylation of cumene yields benzene and propene, whereas dehydrogenation gives R-methyl styrene.20-25 Cumene reaction can be schematically observed in Figure 1. In addition, a series of studies have reported experiments performed to calculate activation energy and heat of adsorption during cumene cracking.26-28 The activity for dehydration of isopropyl alcohol has been reported to be a good measurement of acidity in metal oxide catalysts. Fairly good correlations have been found between the acidity quantified by adsorption of ammonia or pyridine and the activity for dehydration of isopropyl alcohol (IPA). Catalytic activity for dehydration of IPA to propene is proportional to the catalyst acidity, whereas dehydrogenation toward acetone is assumed to be related to catalyst basicity because dehydrogenation is considered to proceed by a concerted mechanism. Therefore, 2-propanol dehydrogenation products (ketones) are preferentially formed on basic catalysts, and dehydration products (olefins or/and ethers) are favored when acidic sites are present. In this work different silica-based mixed oxides were synthesized by the sol-gel method with variations of preparation conditions. Specific surface area (SSA), pore volume (PV), and average pore diameter (APD) were analyzed both in supports and catalysts. Further acid-base properties of supports and catalysts were qualitatively analyzed during IPA conversion. Silicaalumina support and silica-alumina supported NiMo catalyst were evaluated at the microplant scale under atmospheric pressure for cumene cracking conversion, which exhibited high values due to the stronger catalyst acidity. Activation energy was obtained by fitting experimental data to a Langmuir-Hinshelwood model, and heat of adsorption was calculated with the Van’t Hoff equation.
Figure 2. Support preparation from different precursor salts. R* is Al(NO3)3 3 9H2O, Cl2OZr 3 8H2O, Ti[OCH(CH3)2]4, or Mg(NO3)2 3 6H2O.
2. EXPERIMENTAL SECTION 2.1. Support and Catalysts Synthesis. Silica-alumina supports were prepared from tetraethyl ortosilicate (TEOS), aluminum nitrate, and polyethylene oxide (PEO) with an average molecular weight of 8500 g/mol. Aluminum nitrate and PEO were previously dissolved in water and nitric acid (1 mol/dm3), whereas TEOS was dropped into the acid solution with vigorous stirring. Experimental methodology is shown in Figure 2. However, the original procedure13 was modified in order to synthesize different supports, that is, silica-titania, silica-zirconia, and silicamagnesia by changing the precursor salts, amount of water, and gelation time. After the solution had become homogeneous, it was sealed and kept at 50 °C for 24, 48, and 72 h for gelation time. In the case of AlSi supports, gelation time was extended up to 120 h. The amount of PEO was 1.15 g, and the Si/R molar ratio of 7.58 was kept as constant (where R is a different metal, e.g., Al, Ti, Zr, Mg), whereas the water/TEOS molar ratio was varied to study the influence of water amount on support synthesis. Supports were then calcined at 500 °C during 5 h and impregnated with active metals and promoters by using the incipient wetness method. Salts used in this work were cobalt nitrate, ammonium heptamolibdate (AHM), hydrate ammonium tungstate, and nickel nitrate. Calcination of catalysts was carried out at 450 °C during 4 h in the presence of air. Table 1 shows the preparation and compositions for supports and catalysts, respectively. As can be observed, in the case of NiMo and CoMo catalysts the amount of Mo was 12 wt % whereas for NiW catalyst the amount of W was 21 wt %. Nomenclature indicates the chemical symbol of metal precursors used in the support preparation; that is, AlSi-24 corresponds to Al and Si, and the number is the gelation time, 24 corresponds to 24 h of gelation time. 2.2. Support and Catalysts Characterization. Textural properties were analyzed in a Quantachrome Instruments NOVA 4000. A sample of 0.15 g was previously degassed at 300 °C 2716
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
Table 1. Supports and Catalysts Preparation support
salt type
Si/Al (or Zr, Mg, Ti)
gelation
molar ratio
time, h
ZrSi-24
Cl2OZr 3 8H2O
7.58
24
ZrSi-48
Cl2OZr 3 8H2O
7.58
48
ZrSi-72
Cl2OZr 3 8H2O
7.58
72
C12H28O4Zr
7.58
24
Al(NO3)3 3 9H2O Al(NO3)3 3 9H2O
7.58
24
AlSi-48
7.58
48
AlSi-72 AlSi-120
Al(NO3)3 3 9H2O Al(NO3)3 3 9H2O
7.58 7.58
72 120
MgSi-24
Mg(NO3)2 3 6H2O
7.58
24
TiSi-24
Ti[OCH(CH3)2]4
7.58
24
ZrSi-24pa AlSi-24
catalyst type component
NiMo
CoMo
NiW
85
85
74
3
3
5
12
12
21
support wt % promoter wt % active metal wt % a
ZrSi-24p corresponds to a support where the precursor salt was zirconium isopropoxide.
during 3 h before starting the nitrogen physisorption. SSA was calculated from N2 physisorption at -196 °C using the BrunauerEmmet-Teller (BET) equation. 2.3. Evaluation of Supports and Catalysts during Isopropyl Alcohol (IPA) Decomposition. It is well-known that IPA dehydration is carried out on acidic sites giving propene as product, whereas dehydrogenation occurs in basic sites favoring the acetone formation.29 Different AlSi and ZrSi supports were evaluated for determining its acidic or basic nature with isopropyl alcohol. Supports (200 mg) were previously dried at 400 °C for 1 h with a nitrogen flow of 50 mL/min. After drying, the acid-base tests were performed in a plug flow glass reactor at 125, 150, and 175 °C. Evaluation was completed at atmospheric pressure by bubbling nitrogen through a saturator at 15 °C containing isopropyl alcohol to ensure a low molar fraction of alcohol. The following modified Antoine’s equation for IPA was used to calculate the saturation pressure and estimate the molar fraction at the reactor inlet which was calculated to be 3.9 mol % under these conditions:
were tested at 100, 125, and 150 °C because at higher temperature the conversion is very high and reaction was controlled by temperature. Reaction products were analyzed by GC as well. The internal diameter of the glass reactor is 7 mm and its length is 10 cm; however, the amount of solid was ∼200 mg which corresponds to a catalytic bed height of 2-3 mm. 2.4. Evaluation of Supports and Catalysts during Cumene Cracking. Cumene cracking was carried out on the prepared systems for characterizing the supports and catalysts acidity. For this purpose, 200 mg of synthesized support were evaluated in a fixed bed microreactor under atmospheric pressure. Supports were dried by passing a nitrogen flow for 1 h at 400 °C. Catalysts were sulfided in situ with a mixture of H2 and CS2 over 3 h at 400 °C. In both cases, the gas flow rate was 50 mL/ min. After sulfidation, H2 was passed through a saturator at 5 °C containing cumene which gave a molar concentration of 0.20 mol % with a flow rate of 100 mL/min at the reactor inlet. Three different reaction temperatures were evaluated (300, 350, and 400 °C) to analyze the catalysts behavior. Reaction time was 4 h, and products were analyzed by on line GC every 30 min once reaction began. 2.5. Calculation of Activation Energy and Heat of Adsorption during Cumene Cracking Reaction. The conversions of cumene were kept below 10% to operate under differential regime and obtain kinetic data. Catalytic dealkylation of cumene has been employed for testing the acidic nature of catalysts, and kinetic studies have been reported in the literature.23 It has been accepted that reaction can be schematically represented by eq 2, where the reversible reaction is negligible as reported by Take et al.:26 C þ S T C 3 S þ H2 f products ð2Þ where C is cumene, S is an active site and C 3 S is cumene adsorbed onto an active site. The stage which controls the overall reaction rate is considered to be the surface reaction where cumene is adsorbed onto a single active site. Reaction between cumene and hydrogen takes place and products are immediately formed without desorption stage according to kK A PC ð3Þ - rc ¼ 1 þ K A PC
ð1Þ
where -rc is the cumene reaction rate (mol gcat-1 h-1), k is the kinetic constant, KA is the cumene adsorption constant, and PC is the cumene partial pressure. Heat of adsorption was calculated by the Van’t Hoff equation according to ΔH ΔS ð4Þ þ ln K A ¼ RT R
Equation 1 was obtained from Hysys Process Simulator Ver. 3.1 and it is valid in the temperature range of -23 to 235 °C. P is in kPa and T in absolute degrees. Products were analyzed by GC in line with the reaction system. On the other hand, the activity of supported catalysts (Ni, Co, Mo, and W) was tested in the sulfided state. Sulfidation was carried out in situ at 400 °C for 3 h by passing a gaseous mixture of H2 þ CS2 through the catalyst (200 mg). After sulfidation, the gas was changed to N2 which was also bubbled into a saturator at 15 °C containing IPA with a flow rate of 50 mL/min. Only NiMo/ AlSi-72 and NiW/AlSi-72 were evaluated in the IPA decomposition during 4 h. Since catalysts enhance the IPA conversion, they
NiMo/AlSi-72 catalyst was employed to obtain kinetic parameters at low conversion using the Langmuir-Hinshelwood kinetic model, whereas heat of adsorption was calculated with the Van’t Hoff equation. To employ eq 3 it is necessary to have different values of partial pressure of cumene during the reaction that is fed to the reactor at different initial concentrations. A saturator operating at three different temperatures (5, 10, and 15 °C) was used to control the initial concentration of cumene at the reactor inlet (0.2, 0.3, and 0.4 mol % at the aforementioned saturator temperatures, respectively) which directly impacts on the initial partial pressure of cumene. Equation 5 shows the modified Antoine’s equation
ln P ¼ 83:837 -
8249:01 - 9:5152 ln T T
þ 2:00272 10-6 T 2
2717
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research used to calculate the saturation pressure of cumene at each temperature. This equation was also taken from Hysys Process Simulator Ver. 3.1.
ARTICLE
Table 2. Textural Properties of Synthesized Supports and Porous Distribution pore size distribution, %
9687:7 - 19:305 ln T þ 0:0177 T ð5Þ T
SSA,
PV,
APD,
support
m2/g
cc/g
nm
micropores
mesopores
macropores
where P is pressure in kPa and T is absolute temperature in Kelvin. The validity range of the equation is from -96 to 360 °C. To calculate activation energy and heat of adsorption during cumene cracking, the reaction temperature was varied (250, 275, and 300 °C) because at higher temperatures, conversion was above 10%. Therefore, to ensure the reactor operated in differential regime, by using 20 mg of catalyst which is equivalent to less than 1 mm of catalytic bed height.
ZrSi-24
439
0.25
2.32
42.47
56.96
0.57
ZrSi-48
472
0.27
2.26
43.52
55.99
0.49
ZrSi-72
592
0.33
2.23
44.62
54.91
0.48
AlSi-24 AlSi-48
586 521
0.36 0.28
2.43 2.14
29.43 27.58
60.08 62.34
10.50 10.08
AlSi-72
516
0.45
3.45
21.79
75.19
3.03
AlSi-120
490
0.85
5.79
11.95
58.15
29.90
MgSi-24
621
0.40
2.56
30.36
69.09
0.56
3. RESULTS AND DISCUSSION
TiSi-24
744
0.43
2.29
42.38
57.05
0.57
ln P ¼ 136:712 -
3.1. Support and Catalysts Synthesis. It is known that different organic polymers can induce the phase separation when they are involved in sol-gel reactions using TEOS. When PEO is used, phase separation is carried out by repulsion forces between a polar solvent and PEO which has a hydrophobic nature. A common phenomenon is phase separation to give two immiscible phases, and sometimes these phases are stable. Phase separation occurs when limits between miscibility and spinodal regions are reached and separation is imminently enhanced by nucleation and growth. Composition within the spinodal region is unstable, and phases are separated easily by a process known as spinodal decomposition. Sometimes the mechanism of separation is determined by the morphology of each phase because small droplets are formed by nucleation and growth, whereas spinodal decomposition leads both phases to be interconnected. However, morphology is not always a confident parameter because all continuum phases can be broken and droplets can adopt any shape.30-32 A possible reaction scheme by which silica interacts with PEO has been reported in the literature33 where (1) PEO and small oligomers of silica coexist in the solution without specific interaction just after hydrolysis of the silicon alkoxide. (2) Then, with the progress of condensation, a ramified aggregated complex between PEO and silica oligomers is formed, which is characterized by a larger apparent value of radius of gyration and a smaller fractal dimension than in the PEO-free system. (3) After gelation, the fractal dimension of scatterers remains to be smaller than in the PEO-free system, because PEO associated with the silica network inhibits aggregation within the gel networks. Furthermore, PEO inhibits the condensation during the aging and drying process to create a less strongly cross-linked dry gel. Table 2 summarizes the synthesized supports and their textural properties as well as their pore size distribution. In the case of ZrSi supports, SSA and PV increased as gelation time increased, whereas in the AlSi supports there was a reduction of SSA but significantly APD and PV increased with gelation time. The macropores content is also higher at 120 h of gelation time as a product of cross-linking. The TiSi support also has a high SSA; however, PV and APD are very similar compared to the MgSi support. In this work, we found that PEO can be used to control the macropores morphology by phase separation in the system TEOS-Al(NO3)3 3 9H2O (or Cl2OZr 3 8H2O; Ti[OCH(CH3)2]4; Mg(NO3)2 3 6H2O). Figure 3a shows the relationship between textural properties and gelation time for AlSi supports. It can be observed graphically that the tendency followed by APD and PV
Figure 3. (a) Textural properties of AlSi supports as function of gelation time; (b) influence of water amount on textural properties of AlSi supports: (O) specific surface area, (b) average pore diameter, (0) total pore volume.
is to increase as gelation time increases. On the other hand, SSA showed a decrease in its value likely due to spinodal decomposition which causes a reduction of the sample volume during phase separation in the sol-gel medium. However, pores are not affected by this volume contraction, instead showing a tendency to increase as a consequence of higher time of gelation, making the AlSi network denser because of better M-O-Si cross-linking. In this way, the transitory phase is easily removed by calcination causing a bigger pore cavity. The higher the molar ratio of H2O/ TEOS, the higher the average pore diameter up to 1.5 mol of water as shown in Figure 3b. Beyond this value, average pore diameter and pore volume tend to decrease. It is stated that the pore size is dependent on the initial compositions and preparation conditions. AlSi supports which were exposed to the highest gelation time with a lesser amount of water were meso and macroporous. 2718
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
Figure 4a shows different trends of textural properties of ZrSi supports. It can be observed that pore volume and average pore diameter are almost constants. It is also seen that SSA diminishes
as the water amount is increased as shown in Figure 4b. It was also observed that only for ZrSi supports did the stirring become difficult at a low water amount after adding TEOS. A probable explanation for this is that Si and Zr formed networks and condensed almost immediately in the presence of small amounts of water creating a porous structure which was kept in the successive steps. In this way, only a reduction of support volume is observed with respect to the amount of water present, and at the same time there is a reduction in SSA, but pore size is kept constant. Once supports were synthesized and characterized by nitrogen physisorption they were impregnated with different precursor salts and promoters as shown in Table 1. Catalysts obtained were NiMo, CoMo, and NiW supported on AlSi, ZrSi, and MgSi. Table 3 shows a comparison among the catalysts synthesized. There is a drastic reduction of SSA, APD, and PV in all catalysts due to the blockage that pores undergo after impregnation. Particularly, in the case of the NiMo/MgSi-24 catalyst, an almost 6-fold severe reduction of SSA after impregnation was observed. A possible explanation for this behavior is that MgO in the support easily is hydrated and converted into Mg(OH)2 which possesses typical behavior in the presence of water as reported in our previous study.34 For NiMo catalysts supported on AlSi, it is observed that at lower gelation time the micropores were more abundant. However, the catalyst prepared at 72 h of gelation time almost kept its porous distribution without significant variation with respect to its support, but the SSA and PV were reduced. This is due to the wide amount of mesopores that remains unaffected. In a comparison of the AlSi-24 and MgSi-24 supports for the NiMo catalysts, it is observed that NiMo/MgSi-24 is extremely mesoporous compared with NiMo/AlSi-24 but the SSA of NiMo/ MgSi-24 is relatively very low, mainly due to the large pore diameter (i.e., 6.8 nm). In the case of CoMo/AlSi catalysts it is
Figure 4. (a) Textural properties of ZrSi supports as function of gelation time; (b) influence of water amount on textural properties of ZrSi supports: (O) specific surface area, (b) average pore diameter, (0) total pore volume.
Table 3. Textural Properties of Synthesized Catalyst and Porous Distribution porous distribution, % 2
SSA, m /g
PV, cc/g
NiMo/AlSi-24
327.3
0.23
2.75
NiMo/AlSi-72
335.3
0.29
3.47
CoMo/AlSi-72
351.6
0.46
CoMo/AlSi-120
389.6
0.55
NiMo/AlSi-24
327.3
0.23
2.75
54.66
40.94
4.39
NiW/AlSi-24
242.3
0.34
5.61
7.51
80.77
11.71
CoMo/AlSi-72
351.6
0.46
5.28
10.12
80.14
9.74
NiMo/AlSi-72
335.3
0.29
3.47
19.63
78.21
2.16
NiW/AlSi-72
238.7
0.23
3.79
32.35
64.51
3.14
NiW/ZrSi-24
279.6
0.16
2.31
41.13
57.75
1.12
NiMo/ZrSi-72
338.6
0.21
2.48
44.92
51.65
3.44
support
APD, nm
micropores
mesopores
macropores
54.66
40.94
4.39
19.63
78.21
2.16
5.28
10.12
80.14
9.74
5.62
9.76
68.39
21.85
Different Gelation Time on NiMo Catalysts
Different Gelation Time on CoMo Catalysts
Different Active Metals at the Same Gelation Time on NiMo Catalysts
Different Active Metals at the Same Gelation Time on CoMo, NiMo, and NiW Catalysts
Different Active Metals and Different Gelation Time on NiW and NiMo Catalysts
Different Support at the Same Gelation Time on NiMo Catalysts NiMo/MgSi-24 NiMo/AlSi-24
73.35 327.3
0.13
6.83
10.96
85.34
3.71
0.23
2.75
54.66
40.94
4.39
2719
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research observed that at higher gelation time the catalyst has more macropores which could be due to the bigger cavities left by water after evaporation during the calcination of supports. These cavities did not undergo a drastic collapse after impregnation. When AlSi-24 supports were impregnated with a different active metal (Mo or W), it is seen that the NiW catalyst possesses a higher amount of meso- and macropores. On the other hand, NiMo/AlSi-24 is richer in micro- and mesopores. However, the
Figure 5. Pore size distribution of (a) AlSi-supported catalysts having 72 h of gelation time, (9) NiMo/AlSi-72, (—) CoMo/AlSi-72, ([) NiW/AlSi-72, and (b) NiMo catalysts varying the gelation time, (]) 24 h, (0) 72 h.
ARTICLE
higher amount of precursor salts used to prepare the NiW catalyst reduced its SSA as compared with the NiMo catalyst SSA. A comparison of the AlSi-72 supports with different active metals and precursors shows that the CoMo catalyst is meso- and macroporous mainly whereas the NiW is micro- and mesoporous in nature. In the case of the ZrSi support prepared at different gelation times and impregnated with Ni and Mo, it is seen that the NiMo/ ZrSi-72 catalyst has a higher SSA compared with the NiW/ZrSi-24 SSA. Both catalysts were similar in their pore size distribution. Figure 5 shows the deviation in pore volume as a function of pore diameter of supported catalysts. Figure 5a shows a comparison between the CoMo, NiMo, and NiW catalysts on AlSi-72, and it is observed that CoMo/AlSi-72 has bigger pores while NiMo/AlSi-72 has smaller pores indicating that many of them are plugged by the NiMo metal deposition. Figure 5b shows the pore diameter distribution for NiMo/AlSi-24 and NiMo/AlSi-72 catalyst where the aging effect of the support is obvious and significantly remained after catalyst preparation also. According to Table 3, the catalyst that exhibits the best textural properties in terms of macropores is CoMo/AlSi-120. This property is important during reactions since diffusion gradients are avoided. Hence, it is worth noting the importance of the sol-gel method to synthesize supports and catalysts having meso- and macropores. 3.2. IPA Reaction for Determining Acid-Base Properties. The acid-base properties of support can be evaluated by measuring the dehydration of 2-propanol which is carried out on both the Lewis and Brønsted acid sites. IPA decomposition products are mainly propene and acetone. Propene represents the acidic character of the material, whereas acetone corresponds to the basic nature. The reaction is also sensitive to the temperature; therefore, the reaction temperature was kept low during the estimation of selective products yield. 3.2.1. IPA Reaction on AlSi and ZrSi Supports. To estimate acid-base properties of supports (AlSi, ZrSi, TiSi, and MgSi), IPA decomposition was carried out by using selective yield of products. Figure 6 shows the IPA conversion on these supports
Figure 6. Conversion of isopropyl alcohol over different supports. 2720
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research prepared at the same gelation time. The inset shows the IPA conversion for AlSi supports as a function of gelation time during its preparation at constant reaction temperature (150 °C). In all cases different conversion degrees are obtained. For ZrSi-24 and MgSi-24 supports conversion is higher than 60%, whereas for the AlSi-24 support conversion is around 30%. It is important to keep in mind that IPA decomposition is an indirect method to test acid-base properties of supports/catalysts, and conversion results are only qualitative because of the contribution of both (acid/base) sites. The higher conversion for ZrSi and MgSi further indicates the contribution of both sites, where basic sites of Zr and Mg correspond to acetone formation while their acid sites (mixed oxide sites, Brønsted sites) are responsible for the dehydration reaction. On the other hand, AlSi (Figure 6 inset) indicated that the reaction conversion is selectively carried out on acidic sites and the formation of Brønsted acid sites is the maximum at 72 h of gelation time. The amount and nature of active sites can be estimated by calculating the selective yield of reaction products such as propene, ether, and acetone. Hence 2-propanol decomposition on basic sites proceeds through an elimination reaction yielding acetone, while in acid sites 2-propanol dehydrates to propylene and di-isopropyl ether. The ratio of propylene and di-isopropyl ether may further depend on the strength of acid sites where the reaction takes place. In addition, the reaction temperature was varied to study the acid-base behavior of supports because it is known that IPA decomposition follows two routes: dehydration and dehydrogenation. Both of them may occur simultaneously. Dehydration occurs on Brønsted acid sites where the propylene is produced, whereas di-isopropyl ether and propanaldehyde are formed on Lewis acid sites. Dehydrogenation is carried out in basic sites to yield acetone as product. It has been reported that reaction conversion not only depends on acid-base properties of support/catalyst surface but also is strongly dependent on reaction temperature and IPA partial pressure.35 Figure 7 shows the acid-base behavior of AlSi-120 and ZrSi-24 with two different precursors of zirconium, that is, Cl2OZr 3 8H2O that was the most used precursor during the synthesis of ZrSi supports in this work and C12H28O4Zr that was used to study the influence of changing the zirconium salt in the support properties. In both cases, gelation time was the same (24 h). AlSi-120 support in Figure 7a shows high yield of acetone at lower temperatures but as temperature is increased the propene yield is higher and acetone yield is decreased. Comparison of the two ZrSi supports indicated that by using C12H28O4Zr the support behavior turns more acidic as compared with using Cl2OZr 3 8H2O as precursor, and a possible explanation for the higher acidity of support when using zirconium isopropoxide could be found in its high oxygen content that influences the formation of bonds. ZrSi supports generally exhibit acid and basic properties depending on the reaction temperature. However, the acidic tendency is more significant at higher temperature which could be due to a change in the reaction mechanism on the surface electronic state by interactions among zirconium and silicon oxide. After support calcination only acid properties prevail, and the basic character remains at lower temperatures as reported by Yamaguchi et al.36 It is to be expected that AlSi-120 support enhances the activity toward cracking reactions due to its acid properties which are improved at higher temperatures in spite of having lower IPA conversion compared with other supports. In all cases propanaldehyde, which is formed as well in Lewis acid sites, tends to disappear as temperature increases and its yield is around 2% by
ARTICLE
Figure 7. Yield of different products during IPA decomposition over supports. (a) AlSi-120; (b) ZrSi-24 with Cl2OZr 3 8H2O as precursor salt; (c) ZrSi-24p with C12H28O4Zr as precursor salt: propene (black columns); di-isopropyl ether (white columns); propanaldehyde (gray columns); acetone (dotted columns).
far. From this observation it is concluded that activity of Lewis acid sites is kept almost constant during reaction but the activity of Brønsted acid sites is increased linearly which is a consequence of the employed method of synthesis. 3.2.2. IPA Reaction on NiMo/AlSi-72 and NiW/AlSi-72 Catalyst. IPA decomposition is useful to determine the acid-base properties of catalysts as well. NiMo and NiW on AlSi-72 support were tested during IPA decomposition. Textural properties for both catalysts were similar as can be seen in Table 3, especially PV and APD, but SSA is higher for NiMo catalyst compared with that for NiW catalyst, being predominantly mesoporous. Figure 8a shows the reaction rate, and it is observed that it is very similar for both catalysts. Conversion was slightly higher in catalysts compared with bare supports which indicated that sulfidation enhanced the formation of Brønsted acid sites.15 WO3 has been reported37 to generate strong Brønsted acid sites that can be rapidly deactivated depending on the tungsten amount added. Figure 8b shows the yield of propene, di-isopropyl ether, propanaldehyde, and acetone obtained as a function of temperature. It is observed that at low temperatures the acetone yield is higher indicating that the 2721
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
Figure 9. Yield of different products during cumene cracking over supports: propene (black columns); propane (white columns); benzene (gray columns); styrenes (dotted columns).
Figure 8. (a) Reaction rate during IPA conversion over catalysts: NiMo/AlSi-72 (9); NiW/AlSi-72 (0); (b) yield of different products during IPA decomposition: propene (black columns); di-isopropyl ether (white columns); propanaldehyde (gray columns); acetone (dotted columns).
reaction is thermally controlled by the dehydration (removal H2O) reaction mechanism. However, as temperature increases, the propene yield also increases which means a higher influence of acid sites during the reaction and a linear increment of propene yield at higher temperature is observed. 3.3. Catalytic Activity of Supports during Cumene Cracking. Cumene cracking was carried out on mixed-oxides supports and supported catalysts. Since silica-alumina support has more acid sites the cumene cracking can be enhanced on these sites. It is to be expected that conversion is almost 100% on an AlSi support due to its strong acidity especially at high temperatures. Smaller conversions were exhibited with the remaining supports; however, in the case of silica-titania and silica-magnesia supports, conversion was negligible. It is thought that the synthesis method of this support is responsible for its behavior in spite of having good textural properties as shown in Table 2. Regardless of insignificant conversion in cumene cracking, TiSi supports have shown high conversion on thiophene HDS and other important reactions related with petroleum upgrading.38,39 Silica-alumina prepared by the sol-gel process is able to produce a larger amount of Brønsted acid sites most likely because of the higher homogeneity of the Al atoms in the silica network by which PEO has the ability to generate more Brønsted acid sites on the surface of silica-alumina by forming hydrogen bonds with silica during the sol-gel reactions. The interaction between PEO and silica not only induces the phase separation, but also affects the condensation pathways of silica polymers and the mesoporous structure in the calcined gel. It has been reported that the oxygen atoms in PEO can coordinate with Al3þ cations and the coordination probably affects the homogeneity in silicaalumina and the local structure of aluminum in silica networks
promoting the generation of larger number of Brønsted acid sites.40 Conversion on silica-magnesia support was practically zero due to its basic nature which is a consequence of the role that Mg2þ and O2- ions play in the Si network.41 This property is very important on a HDS reaction due to its tendency to reduce coke formation. According to the present results, AlSi support enhances the hydrogenolysis reaction during cumene cracking on Brønsted acid sites produced during support preparation which forms a strong metal-oxygen-silicon bond. It was found that acid sites in catalysts are formed from negative or positive charges in excess, and their formation depends on the preparation method.11 Evaluation of these mixed oxides allowed us to establish the nature of active sites in a qualitative way. Several authors have stated the importance of support acidity during cumene reaction,17,18 paying special attention to acid sites and their role in the reaction. However, it is well-known that acid sites are strong promoters for coke deposition deactivating the catalyst. In this case the short reaction time kept the conversion almost constant indicating that probably the preparation method of supports and linkage between atoms retards the catalyst aging under our experimental conditions, but we did not quantify in any case the carbon deposition nor did we analyze the spent catalyst after experiments. Figure 9 shows the yield of different products which were mainly propene, propane, benzene, and styrenes, except when MgSi and TiSi supports were used which did not exhibit any conversion. Propene and propane are present in smaller amounts. The most abundant product was benzene followed by styrenes. It has been reported26 that during cumene decomposition there are two main reactions, that is, dealkylation and disproportionation. Both reactions are carried out on strong and weak acid sites, respectively. The main product of dealkylation is benzene, whereas styrene-derivatives are the main product of disproportionation. Rana et al.20 have established that benzene, propene, and R-methyl-styrene are the main products of cumene cracking where R-methyl-styrene is a product of formation of π-complexes that are transformed into σ-complexes in a sequential reaction. 3.4. Catalytic Activity of Different Catalysts in Cumene Cracking. Cumene decomposition is a common way to verify the nature of catalysts because the reaction takes place on Brønsted acid sites as reported but preparation method, metals impregnation, and activation play an important role. Figure 10 shows the reaction rate during cumene cracking with different catalysts. Most of synthesized catalysts are acidic in nature, except those supported on magnesia because this is basic as shown with 2722
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
Figure 10. Catalysts evaluated during the cumene cracking reaction at different temperatures: 300 °C (black columns); 350 °C (white columns); 400 °C (dotted columns).
their respective cumene reaction rates. It can be observed that NiMo/AlSi-72 was the most active catalyst at 400 °C, and for this reason it was employed for determining kinetics of cumene cracking. Since the cumene cracking reaction is carried out on supported (NiMo, CoMo, NiW) catalysts in the sulfided phase, the contribution of Brønsted acid sites by -H groups cannot be excluded clearly from this study. NiMo/AlSi catalysts were more active than AlSi-supported CoMo and NiW catalysts for cumene cracking. On the contrary, it is seen that NiMo/ZrSi and NiMo/MgSi exhibited the lowest reaction rates (results not shown). NiW/AlSi-72 catalyst was more active than CoMo/AlSi-72 catalyst which evidences the influence of preparation method on catalytic activity. CoMoS (NiMoS) phase contributes to improve the cumene cracking, and the reason for this behavior can be found in sulfhydryl groups which are present in sulfided catalysts. It is considered that anion vacancies are converted to sulfhydryl groups, which is a reversible reaction depending on the H2S concentration as well as the total number of vacancy sites (anion vacancies). Vacancy sites are thought to be involved in the formation of Brønsted acid sites38 by which sulfided catalysts contain both anion vacancies and sulfhydryl groups. This is in line with literature reports where it is stated that the number of Brønsted acid sites increases with an increase of H2S concentration through dissociative adsorption on CoMo supported catalysts.42 A similar type of results was reported by Rana et al.43 who showed that Mo, CoMo, and NiMo catalysts supported on SiO2-ZrO2 exhibited higher cracking activity. They also found an uneven relationship between cracking function and oxygen chemisorption, which measures the general state of MoS2 dispersion (anion vacancies). This suggests that the cracking functionality is a property of the supported MoS2 phase and that sulfide anion vacancies are related to the cracking sites. 3.5. Kinetics and Heat of Adsorption of Cumene Cracking on NiMo/AlSi-72. Cumene cracking (dealkylation) was carried out on NiMo/AlSi-72 catalyst to evaluate kinetic parameters and an adsorption constant. Experimental data were well fitted to eq 3 where a mechanism based on cumene adsorption on a single site is carried out. For the dealkylation, the adsorbed cumene has been considered to be σ-complex with the proton attached to the substituted position. Accordingly, the rate-determining step in the kinetic model corresponds to C(phenyl)-C(isopropyl) cleavage
Figure 11. (a) Calculation of kinetic constants for cumene cracking at (O) 250, (b) 275, and (0) 300 °C. (b) Determination of activation energy (O) and heat of adsorption (b).
in the σ-complex, which has been considered as heterolytic and demonstrated in earlier studies that dealkylation of monoalkylbenzenes in the rate-determining step is heterolytic.44,45 By linearization of eq 3, it is possible to obtain kinetic constants at different reaction temperatures whose determination by linear regression is shown in Figure 11a. It is observed that all data are well fitted to the Langmuir-Hinshelwood model and good correlation coefficients were obtained at each temperature (0.980, 0.999, and 0.999 for a reaction temperature of 250, 275, and 300 °C, respectively). Once kinetic constants were obtained, the activation energy was also calculated as illustrated in Figure 11b where a linear fitting is observed with a correlation coefficient of 0.988. Heat of adsorption and entropy were calculated from eq 4 and presented in Table 4 together with activation energy obtained with the catalyst employed. The correlation coefficient for heat of adsorption was 0.976 and showed a linear tendency as observed in Figure 11b. The value of heat of adsorption (-14.15 kcal mol-1) means that cumene is strongly adsorbed because of catalyst acidity and supported by the entropy loss (-25.20 cal mol-1 K-1). It has been reported that values of ca. -30 cal mol-1 K-1 are typical for immobile adsorption of cumene onto active sites indicating strong adsorption of molecules;26 however, reported values in this work are slightly lower which could be a result of the catalysts preparation method and reaction conditions in both studies. Conversion during cumene cracking is dependent not only on the acid nature of catalyst but also on pore structure, that is, when the free mean pathway of reactant molecules is higher than the 2723
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
Table 4. Adsorption and Kinetic Parameters for Cumene Cracking temperature, °C parameter
250
275
300
k
0.00022
0.00101
0.00223
KA
2.7621
1.1796
0.8480
activation energy, kcal mol-1
27.75
frequency factor, h-1
9.49 107 -1
heat of adsorption, kcal mol
-14.15
entropy of adsorption, cal mol-1 K-1
-25.20
mean pore diameter, it is probable that molecules adsorb onto catalytic surface and further react. For this reason, pore structure also plays an important role in kinetics together with external diffusion gradients that must be minimized in order to obtain intrinsic kinetics.
’ CONCLUSIONS Different supports and catalysts were synthesized by the solgel method, and it was observed that this method is able to give meso- and even macroporous materials with a good combination of acid-base properties. The acid-base properties of the support is determined by the selectivity during the IPA reaction, which indicated that acidic supports (AlSi) are selectively good for dehydration reaction, while MgSi and ZrSi have a combination of both (acid and basic) sites. It was also observed that gelation time significantly influences textural properties as well as genesis of acid sites. In addition, the acidic reaction function of IPA was considerably increased as reaction temperature increased indicating that not only the support surface properties influence the activity but also reaction conditions play a vital role. In general, mixed oxides had good catalytic activity in cumene and IPA reactions. Acid sites favored the cumene cracking reaction especially when AlSi supports were used which was attributable to Brønsted acid sites, and, at 400 °C conversion, the AlSi-supported catalysts exhibited very high values indicating a large number of Brønsted sites. Experimental data were well fitted to a LangmuirHinshelwood kinetic model, and heat of adsorption was obtained by using the Van’t Hoff equation. The stage that controls the overall kinetics was the surface reaction, and kinetic modeling is in agreement with adsorption of cumene over a single active site. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT The authors thank Instituto Mexicano del Petroleo for financial support. Fernando Trejo also thanks CONACYT and Instituto Politecnico Nacional. ’ REFERENCES (1) Speight, J. G. The Chemistry and Technology of Petroleum; Marcel Dekker: New York, 1999. (2) Ramírez, J.; Rana, M. S.; Ancheyta, J. In Hydroprocessing of Heavy Oils and Residua; Ancheyta, J., Speight, J. G., Eds.; Taylor & Francis: New York, 2007.
(3) Topsøe, H.; Clausen, S. B.; Franklin, M. E. Hydrotreating Catalysts, Science and Technology; Springer: Germany, 1996. (4) Howard, F. R. Handbook of Commercial Catalysts. Heterogeneous Catalysts; CRC Press: New York, 2000. (5) Scherzer, J.; Gruia, A. J. Hydrocracking Science and Technology; Marcel Dekker: New York, 1996. (6) Shali, N. B.; Sugunan, S. Influence of transition metals on the surface acidic properties of titania prepared by sol-gel route. Mater. Res. Bull. 2007, 42, 1777–1783. (7) Selli, E.; Forni, L. Comparison between the surface acidity of solid catalysts determined by TPD and FTIR analysis of pre-adsorbed pyridine. Microporous Mesoporous Mater. 1999, 31, 129–140. (8) Klein, J.; Lettmann, C.; Maier, W. F. Thermally stable, silicabased amorphous porous mixed oxides prepared by sol-gel procedures. J. Non-cryst. Solids 2001, 282, 203–220. (9) Murrell, L. L. Sols and mixtures of sols as precursors of unique oxides. Catal. Today 1997, 35, 225–245. (10) Klein, S.; Thorimbert, S.; Maier, W. F. Amorphous microporous titania-silica mixed oxides: preparation, characterization, and catalytic redox properties. J. Catal. 1996, 163, 476–488. (11) Yabuki, M.; Takahashi, R.; Sato, S.; Sodesawa, T.; Ogura, K. Silica-alumina catalysts prepared in sol-gel process of TEOS with organic additives. Phys. Chem. Chem. Phys. 2002, 4, 4830–4837. (12) Rodríguez, O.; Gonzalez, F.; Bosch, P.; Portilla, M.; Viveros, T. Physical characterization of TiO2 and Al2O3 prepared by precipitation and sol-gel methods. Catal. Today 1992, 14, 243–252. (13) Takahashi, R.; Sato, S.; Sodesawa, T.; Yabuki, M. Silicaalumina catalyst with bimodal pore structure prepared by phase separation in sol-gel process. J. Catal. 2001, 200, 197–202. (14) Hino, M.; Kurashige, M.; Matsuhashi, H.; Arata, K. A solid acid of tungsta-niobia more active than aluminosilicates for decompositions of cumene, ethylbenzene, and toluene. Appl. Catal., A 2006, 310, 190– 193. (15) Topsøe, N. Y.; Topsøe, H. FTIR studies of Mo/Al2O3-based catalysts. I. Morphology and structure of calcined and sulfided catalysts. J. Catal. 1993, 139, 631–640. (16) Sarback, Z. Acidity and catalytic activity of cobalt-molybdena catalyst supported on alumina. The effect of incorporation sequence of sodium and fluoride ions. Appl. Catal., A 1997, 164, 13–19. (17) Boorman, P. M.; Kydd, R. A.; Sarbak, Z.; Somogyvari, A. Surface acidity and cumene conversion. I: A study of γ-alumina containing fluoride, cobalt, and molybdenum additives. J. Catal. 1985, 96, 115– 121. (18) Boorman, P. M.; Kydd, R. A.; Sarbak, Z.; Somogyvari, A. Surface acidity and cumene conversion. III: A study of γ-alumina containing fluoride, cobalt, and molybdenum additives: The effect of sulfidation. J. Catal. 1987, 106, 544–548. (19) Bautista, F. M.; Campelo, J. M.; García, A.; Luna, D.; Marinas, J. M.; Romero, A. A.; Navío, J. A.; Macías, M. Fluoride and sulfate treatment of AlPO4-Al2O3 catalysts. I: Structure, texture, surface acidity, and catalytic performance in cyclohexene conversion and cumene cracking. J. Catal. 1994, 145, 107–125. (20) Rana, M. S.; Maity, S. K.; Ancheyta, J.; Murali Dhar, G.; Prasada Rao, T. S. R. Cumene cracking functionalities on sulfide Co(Ni)Mo/ TiO2-SiO2 catalysts. Appl. Catal., A 2004, 258, 215–225. (21) Malecka, A. Cumene cracking on dodecatungstosilicic acid catalyst. J. Catal. 1997, 165, 121–128. (22) Belhakem, A.; Bengueddach, A. Cumene cracking on modified mesoporous material type MCM-41. Turk. J. Chem. 2006, 30, 287–295. (23) Goncharuk, V. V. Relationship between kinetic and thermodynamic characteristics of alumosilicate catalysts. React. Kinet. Catal. Lett. 1979, 10, 259–262. (24) Chiranjeevi, T.; Muthu Kumaran, G.; Gupta, J. K.; Murali Dhar, G. Synthesis and characterization of acidic properties of Al-HMS materials of varying Si/Al ratios. Thermochim. Acta 2009, 443, 87–92. (25) Bradley, S. M.; Kydd, R. A. Ga13, Al13, GaAl12, and chromiumpillared montmorillonites: Acidity and reactivity for cumene conversion. J. Catal. 1993, 141, 239–249. 2724
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
Industrial & Engineering Chemistry Research
ARTICLE
(26) Take, J.; Tozawa, Y.; Yoneda, Y. A kinetic study of the simultaneous dealkylation and disproportionation of cumene over silica-alumina. Bull. Chem. Soc. Jpn. 1979, 52, 302–306. (27) Pansing, W. F.; Malloy, J. B. Characterizing cracking catalysts by the kinetics of cumene cracking. Ind. Eng. Chem. Process Des. Dev. 1965, 4, 181–187. (28) Gambaro, L. A.; Briand, L. E. In situ quantification of the active acid sites of H6P2W18O62 3 nH2O heteropoly-acid through chemisorption and temperature programmed surface reaction of isopropanol. Appl. Catal., A 2004, 264, 151–159. (29) Seo, K. T.; Kang, S. C.; Kim, H. J.; Moon, S. K. Isopropyl alcohol decomposition over molybdena-alumina catalyst. Korean J. Chem. Eng. 1985, 2, 163–171. (30) Seward, T. P. Phase Diagrams, Materials Science and Technology; A.M. Alper Academic Press: New York, 1970. (31) Nakanishi, K.; Tanaka, N. Sol-gel with phase separation. Hierarchically porous materials optimized for high-performance liquid chromatography separations. Acc. Chem. Res. 2007, 40, 863–873. (32) Sagui, C.; Grant, M. Theory of nucleation and growth during phase separation. Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys. 1999, 4, 4175–4187. (33) Takahashi, R.; Nakanishi, K.; Soga, N. Aggregation behavior of alkoxide-derived silica in sol-gel process in presence of poly(ethylene oxide). J. Sol Gel Sci. Technol. 2000, 17, 7–18. (34) Caloch, B.; Rana, M. S.; Ancheyta, J. Improved hydrogenolysis (C-S, C-M) function with basic supported hydrodesulfurization catalysts. Catal. Today 2004, 98, 91–98. (35) Ortiz, I. E.; Lopez, N. J.; Bokhimi, X.; Gomez, R. High selectivity to isopropyl ether over sulfated titania in the isopropanol decomposition. J. Mol. Catal. A 2005, 228, 345–350. (36) Yamaguchi, T.; Morita, T.; Salama, T. M.; Tanabe, K. Surface properties of ZrO2 dispersed on SiO2. Catal. Lett. 1990, 4, 1–6. (37) Herrera, J. E.; Kwak, J. H.; Hu, J. Z.; Wang, Y.; Peden, C. H. F.; Macht, J.; Iglesia, E. Synthesis, characterization, and catalytic function of novel highly dispersed tungsten oxide catalysts on mesoporous silica. J. Catal. 2006, 239, 200–211. (38) Rana, M. S.; Maity, S. K.; Ancheyta, J.; Murali Dhar, G.; Prasada Rao, T. S. R. TiO2-SiO2 supported hydrotreating catalysts: Physicochemical characterization and activities. Appl. Catal., A 2003, 253, 165– 176. (39) Maity, S. K.; Rana, M. S.; Bej, S. K.; Ancheyta, J.; Murali Dhar, G.; Prasada Rao, T. S. R. Studies on physico-chemical characterization and catalysis on high surface area titania supported molybdenum hydrotreating catalysts. Appl. Catal., A 2001, 205, 215–225. (40) Takahashi, R.; Sato, S.; Sodesawa, T.; Suzuki, M.; Ogura, K. Preparation of microporous silica gel by sol-gel process in the presence of ethylene glycol oligomers. Bull. Chem. Soc. Jpn. 2000, 73, 765–774. (41) Chizallet, C.; Bailly, M. L.; Costentin, G.; Pernot, H. L.; Krafft, J. M.; Bazin, P.; Saussey, J.; Che, M. Thermodynamic Brønsted basicity of clean MgO surfaces determined by their deprotonation ability: Role of Mg2þ-O2- pairs. Catal. Today 2006, 116, 196–205. (42) Petit, C.; Mauge, F.; Lavalley, J.-C. In Hydrotreatment and Hydrocracking of Oil Fractions; Froment, G., Delmon, B., Grange, P., Eds.: Elsevier: The Netherlands, 1997. (43) Rana, M. S.; Srinivas, B. N.; Maity, S. K.; Murali Dhar, G.; Prasada Rao, T. S. R. Origin of cracking functionality of sulfide (Ni) CoMo/SiO2-ZrO2 catalysts. J. Catal. 2000, 195, 31–37. (44) Mochida, I.; Yoneda, Y. Linear free energy relationships in heterogeneous catalysis. I. Dealkylation of alkylbenzenez on cracking catalysts. J. Catal. 1967, 7, 386–392. (45) Mochida, I.; Yoneda, Y. Linear free energy relationships in heterogeneous catalysis. III. Temperature effects in dealkylation of alkylbenzenes on the cracking catalysts. J. Catal. 1967, 8, 223–230.
2725
dx.doi.org/10.1021/ie1008037 |Ind. Eng. Chem. Res. 2011, 50, 2715–2725
