Catalytic Mechanism of MCM-41 Supported Phosphoric Acid Catalyst

The desulfurization of fluid catalytic cracking (FCC) gasoline by alkylation over solid acid catalysts is considered to be a viable and less costly pa...
0 downloads 0 Views 1017KB Size
ARTICLE pubs.acs.org/EF

Catalytic Mechanism of MCM-41 Supported Phosphoric Acid Catalyst for FCC Gasoline Desulfurization by Alkylation: Experimental and Theoretical Investigation Rong Wang,† Yonghong Li,*,†,‡ Benshuai Guo,† and Hongwei Sun§ †

Key Laboratory for Green Chemical Technology of State Education Ministry, Tianjin University, Tianjin 300072, P. R. China National Engineering Research Center for Distillation Technology, Tianjin 300072, P. R. China § Department of Chemistry, Nankai University, Tianjin 300071, P. R. China ‡

ABSTRACT: The desulfurization of fluid catalytic cracking (FCC) gasoline by alkylation over solid acid catalysts is considered to be a viable and less costly path to meet environmental regulations of sulfur emissions. However, side reactions in the process lead to significant levels of coke, which will greatly reduce the lifetime of the catalyst. In this paper, the catalytic mechanism of MCM-41 supported phosphoric acid catalyst for gasoline desulfurization by alkylation has been investigated by using experimental methods and quantum chemical calculations to study the catalytic behavior for the adsorption and reaction of different reactants, which can help optimize the reaction conditions and preparation methods of the catalyst for a more efficient alkylation process. The results showed that both the typical main and side reactions in the alkylation process started from a stable alkoxide intermediate that was formed by protonation of olefin adsorbed on the catalyst. Thiophenic compounds were more inclined to be adsorbed on the alkoxide intermediate than olefins for further reaction, and the activation energy for the alkylation of thiophenic sulfurs with alkenes was obviously lower than that for alkene oligomerization. Moreover, the thiophene alkylation was exothermic while the olefin oligomerization was endothermic. On the basis of these findings obtained by experimental and theoretical investigation, two methods that might be useful to further inhibit the occurrence of side reactions and improve the catalyst performance in the alkylation process were proposed.

1. INTRODUCTION At present, fluid catalytic cracking (FCC) gasoline with a high concentration of thiophenic compounds is, by far, the most important sulfur emission contributor.1 Considering that the deep desulfurization of gasoline is receiving increased attention in the research community as a result of the increasingly stringent environmental requirements on fuel specifications, many approaches for deep desulfurization have been put into use at home and abroad.2,3 The olefinic alkylation of thiophenic sulfur (OATS) process is one of these approaches that can be handled under relatively mild conditions with a minimal loss of octane number and without any hydrogen consumption. It can be seen as a good alternative to the conventional catalytic hydro-desulfurization process.2,3 However, the published literature has proved that competitive reactions such as aromatic alkylation and alkene oligomerization in the OATS process can produce undesired side products, which can cause the deactivation of solid acid catalysts by adsorbing on the active sites and plugging the catalyst pores.48 To realize the large-scale industrialization of FCC gasoline desulfurization by alkylation technology, the catalytic stability of solid acid catalysts must be improved. How to prolong the life of the catalyst by improving the catalytic selectivity and without affecting its highly catalytic activity for thiophenic sulfur compound alkylation is the first and foremost problem. As is known to us, the reaction mechanisms are closely related to the deactivation, selectivity, and activity of the catalyst, and so, the study on reaction mechanisms in the OATS r 2011 American Chemical Society

process is very beneficial to solve the problem of catalyst deactivation and can be used as the theoretical guidance to develop solid acid catalysts with superior catalytic performances, to optimize reaction conditions to reduce side reactions, and to prolong the lifetime of the catalyst. However, there have been very few theoretical studies on the reactions in the OATS process catalyzed by solid acid catalysts, except the intrinsic kinetics studies on the alkylation mechanism of 3-methylthiophene (3MT) with 2-methyl-2-butene (2M2B) over USY zeolite and Grace Davison silica supported phosphoric acids catalyst.9,10 The mechanism study on the catalysis of solid acid catalysts during the OATS process by means of quantum chemistry calculation appears to be novel. Furthermore, previous work undertaken by our group has proved that purely siliceous MCM-41 zeolite supported phosphoric acids (denoted as SPAM) compared with HY zeolite (Si/Al = 8) and kieselguhr supported phosphoric acids possessed a relatively excellent catalytic performance for the OATS process in real FCC gasoline, not only for its preferable catalytic activity for the alkylation of thiophenic compounds but also for its superior selectivity, which could obviously suppress alkene oligomerization.11 Therefore, in this work, the effect of temperature on the conversion versus reaction time of thiophenic compounds, representative alkenes, and aromatic hydrocarbons during the Received: May 12, 2011 Revised: August 1, 2011 Published: August 12, 2011 3940

dx.doi.org/10.1021/ef200705b | Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Table 1. Properties and Fraction Ranges of FCC Gasoline in the Experiments properties

datum 1

concentration of sulfurs (mg 3 L ) constitutors of hydrocarbons (wt %) aromatics

288.99

olefins

37.4

saturated hydrocarbons

47.3

15.3

result of Engler distillation test (°C) initial bp

34

50 vol %

84

90 vol %

155

final bp

186

OATS process over the SPAM in real FCC gasoline was investigated by designing a series of experiments. On the basis of a large number of reliable experimental data, the typical main and side reactions in the OATS process over the SPAM were studied from a microscopic point of view by the method of density functional theory (DFT) in quantum chemical calculations. The theory research could reveal the adsorption behavior of different reactants on the SPAM surface, as well as the catalytic role of the acidic proton on active sites of the catalyst. The reliability of the theoretical calculations could be verified by the existing experimental results. The combination of experimental and theoretical results in this paper was advantageous to the further application of SPAM in the OATS process.

2. EXPERIMENTAL SECTION 2.1. Catalysts and Materials. The solid phosphoric acid catalyst used in experiments (SPAM) and its purely siliceous carrier MCM-41 zeolite were both synthesized in our laboratory in accordance with the optimal preparation methods.12,13 The properties and fraction ranges of the FCC gasoline material are both listed in Table 1. 2.2. Equipment and Measurements for Catalytic Experiments. The tests for the effect of temperature on the conversion curves of different reactants catalyzed by the SPAM during the OATS process in FCC gasoline were carried out in a 100 mL stirred batch reactor with autogenous pressure. In each run, the autoclave was charged with 4.5 g of catalyst and 90 mL of fresh gasoline with a stirring rate of 600 rpm, which was sufficient to avoid the external mass transfer resistance. When the set reaction temperature was reached, samples were periodically taken out at the times 0.10, 0.50, 1.00, 1.50, and 2.00 h. During the product sampling, the initial 0.05 mL liquid was not reserved to eliminate the impact of any previous sample leftover in the sampling needle (volume less than 0.05 mL) on the product analysis. The volume of each sample was less than 0.05 mL, so the influence on weight ratio of catalyst to gasoline can be ignored. Taking into account that at least eight samplings were necessary to describe the conversion curves, the experiment was repeated under the same conditions and samples were taken out at the reaction times 0.25, 0.75, 1.25, and 1.75 h. Finally, the samples taken out in the two experiments were all analyzed to obtain the conversion curves of thiophenic compounds, various alkenes, and aromatics at different reaction temperatures over the catalyst. Analyses were performed on a FULI 9790 gas chromatograph (GC) (Zhejiang, China) that was equipped with two detectors (FID, OV-101 column 30 m  0.25 mm  0.50 μm; FPD, OV-101 column 60 m  0.25 mm  0.50 μm). Meanwhile, the different species of sulfur compounds and hydrocarbons in the samples were identified by the analysis of a GCSCD (sulfur compound detector) with a HP-1 column (30 m  0.32

mm  0.25 μm) and a GC-MS (mass spectrometer) with a HP5-MS column (30 m  0.25 mm  0.25 μm), respectively. It should be noted that the FPD analysis for sulfur components in FCC gasoline shows that the major sulfur compounds are thiophene and its derivatives. The experimental results revealed that most of thiophenic compounds were transformed into much heavier molecules during the OATS process catalyzed by the SPAM, except 2,3,5trimethylthiophene (C3-T).12 Considering that the boiling point (bp) of the C3-T (about 170 °C) is very close to the final boiling point of the feed used in experiments (186 °C, as shown in Table 1) and the boiling points of more than 90 wt % of thiophenic sulfurs in the feed are lower than 170 °C, the catalytic activity of SPAM for the alkylation of sulfur compounds can be measured by the total conversion of thiophenic compounds with boiling points lower than that of the C3-T (170 °C). Moreover, because the content of thiophenic sulfurs in the feed (Table 1) is tiny, the alkenes consumed in the thiophenic compound alkylation can be neglected, and its conversion is considered as the consumption in side reactions, such as alkene oligomerization and aromatic alkylation. Therefore, the catalytic selectivity of SPAM for the OATS process can be determined by the conversions of different alkene and aromatic hydrocarbons. 2.3. Computational Details for DFT Study. All calculations were done by the Gaussian 03 programs package.14 The B3LYP hybrid density functional combined with 6-31+G(d,p) basis sets were employed for all geometry optimizations. No geometric constraints were used in the optimizations.15 The ultrafine integration grid was used to ensure convergence. The single point energies of optimized geometries were calculated by the MP2/6-311+G(d,p) level to obtain more realistic values.16 Vibrational frequencies were calculated at the same level of optimization to identify the nature of the stationary points and obtain the zero-point-energy (ZPE) correction. The reaction trajectories were determined by the intrinsic reaction coordinate (IRC) method to verify the correctness of obtained theoretical structures of transition states.

3. RESULTS AND DISCUSSION 3.1. Effect of Temperature on the Conversion Curves of Different Reactants in FCC Gasoline during the OATS Process over the SPAM. According to the experimental procedures

in section 2.2, the relationships between conversions of different reactants and reaction time during the OATS process over the SPAM were investigated under various reaction temperatures from 130160 °C. The experimental results are shown in Figure 1 and Table 2. As shown in Figure 1a, the conversion curves of the total thiophenic compounds with boiling points lower than 170 °C revealed that increasing the temperature was beneficial to increase the catalyst activity for sulfur compound alkylation in the suitable temperature range 130150 °C. However, it seemed that, if the temperature was too high, it was not conducive to the alkylation of thiophenic sulfurs, as a decreasing tendency in the conversion was observed when the temperature was higher than 150 °C. Moreover, the initial conversion rate of thiophenic sulfurs was very fast under every investigated temperature, and there was a tiny increase in the conversion rates when the reaction time was prolonged past 1 h. On the basis of the FID analysis for hydrocarbon components in FCC gasoline, there are four kinds of alkenes, C4-, C5-, C6-, and C7-olefins, in the feed and their contents in total olefins are 9.7, 51.8, 27.3, and 11.2 wt %, respectively. Given that the concentration of C4-olefins is relatively low and the impact of C4-olefin oligomers on the catalyst stability is much smaller than 3941

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Figure 1. Conversion curves of total thiophenic compounds with boiling points lower than 170 °C and various alkenes at different reaction temperatures: (a) total thiophenic compounds; (b) C5-olefins; (c) C6-olefins; and (d) C7-olefins.

Table 2. Conversion of Aromatic Hydrocarbons with Time at Different Temperatures conversions of aromatics (%) at different times (h) temp (°C) 0

0.25

0.5

0.75

1

1.25

1.5

1.75

2

130

0

0

0

0

0

0

0

0

0

140

0

0

0

0

0

0

0

0

0

150

0

0

0

0

0

0

0

0.6

0.8

160

0

0

0

0.9

1.5

2.3

2.9

3.1

3.8

that of the other three alkene oligomers, the conversion curves of C4-olefins under various temperatures were not studied in this work. Figure 1bd displays the conversion curves of C5∼C7-olefins under different reaction temperatures, respectively, indicating that higher temperatures were favorable for the alkene conversions, especially for the C7-olefin conversion. The conversion of C7-olefins was distinctly higher at 160 °C than at the other three temperatures. The conversion rates for C5 and C6-olefins in the initial stage (before 0.5 h) of the OATS process catalyzed by the

SPAM were relatively slow, but these conversion rates increased obviously after 0.5 h. This may be caused by the competition of thiophenic compound alkylation at the initial reaction stage. Although the conversion rate of C7-olefins was a little different and it was fast in the initial reaction stage compared with in the middle and later stages, the conversion of C7-olefins was lower than that of C5- and C6-olefins at every reaction time and temperature, especially after 0.5 h. Moreover, the main components of alkenes in the gasoline feed were C5- and C6-olefins, whose high conversions were adverse to maintaining the stability of the SPAM during the OATS process. The total conversions of the aromatic hydrocarbons at different times in the experiments under various temperatures are listed in Table 2, revealing that the high temperature was beneficial to the aromatic alkylation, as almost no alkyl-aromatics were found in the products over the temperature range 130150 °C and some aromatics in the feed were converted at the higher temperature 160 °C. The effect of the alkylation products of aromatics on the catalyst stability was tiny compared with that of the oligomers of alkenes when the temperature was lower than 150 °C, as few of the aromatics were converted and their content in the feed was only 15.3 wt %. So, the aromatic alkylation could be ignored, and the main side 3942

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

Figure 2. Optimized geometries of the cluster model of SPAM for the theoretical calculation.

reaction was alkene oligomerization during the OATS process catalyzed by the SPAM when the reaction temperature was lower than 150 °C. Therefore, the experimental results in Figure 1 and Table 2 indicated that the reaction temperature and time had a great effect on the catalytic activity and selectivity of the SPAM during the OATS process of FCC gasoline. The high temperature was very advantageous to the occurrence of side reactions such as alkene oligomerization and aromatic alkylation. Although the catalyst activity for alkylation of thiophenic compounds could be improved by increasing temperature over the relatively low temperature range 130150 °C and more of the sulfur compounds in the feed were transformed into much heavier molecules, the catalyst selectivity would be reduced significantly at the high temperature 160 °C. Furthermore, the conversion rate of the thiophenic sulfurs at the initial reaction stage was obviously faster than those of the aromatics and various alkenes under every investigated temperature, which proved that the alkylation rate of thiophenic compounds with alkenes was faster than the reaction rate of alkene oligomerization and aromatic alkylation. However, the conversion rate of alkenes increased visibly when the sulfur content was greatly decreased in the reaction system, especially after 1 h. So, the optimal temperature for the OATS process of FCC gasoline catalyzed by the SPAM was 150 °C, and the reaction time should not be more than 1 h. These reaction conditions were conducive to obtain excellent catalyst performance. The above experimental results could be used to prove the reliability of the following theoretical calculation outcomes. 3.2. DFT Studies on the Catalytic Mechanism of the SPAM in the OATS Process. An improved understanding of how reactant molecules interact with acid sites of the OATS catalysts is necessary to develop a rational approach to the application of these catalysts. Therefore, the DFT method was employed to explore the possible mechanism of adsorption and reaction over the SPAM during the OATS process at the atomic level. Previous literature has reported that the activity of solid phosphoric acid (SPA) is due to the acidity of the liquid glass solution of phosphoric acids supported on the siliceous support.17,18 The PO3H2 or PO2H groups are formed on the surface of the catalyst by the interaction between the POH bonds of phosphoric acids and the surface silanols of the siliceous carrier. So, we used the computationally manageable cluster containing the acidic active group PO2H (Figure 2) to represent the SPAM for cost-effective calculation. Although the cluster model cannot accurately represent the catalyst environment, it is particularly suited to describe local phenomena such

ARTICLE

as the interaction of organic molecules with active sites and the forming and breaking of bonds by high-quality theoretical methods. Calculations with such clusters have proven them to be adequate for qualitative descriptions of chemical rearrangements that occur locally on the active sites.18,19 Furthermore, our purpose was to study the adsorption property of SPAM for different reactants as well as the reaction pathway of the typical main and side reactions in the OATS process over the SPAM. All the investigated energies, for example adsorption and activation energies, are relative, and the cluster model can make the reasonable estimates for them. On the basis of the experimental results, the olefin oligomerization was the major side reaction in the OATS process catalyzed by the SPAM and the aromatic compounds had little effect on the alkylation of thiophenic sulfurs. Moreover, the most representative thiophene and olefin in the feed are thiophene (T) and 2-methyl-2-butene (2M2B), respectively. Therefore, T and 2M2B were selected as the model reactants for the quantum chemical calculations. 3.2.1. Theoretical Investigation on Reactants Adsorption on the SPAM in the OATS Process. Considering that the adsorption of the reactants on the catalyst surface is the first step in reactions catalyzed by solid catalysts and the adsorption may be the controlled step of reaction, the adsorption of representative reactants, such as T and 2M2B, on the SPAM surface is investigated by the method of theoretical calculation, and the calculation results are shown in Figure 3. The geometric structures shown in Figure 3a and b, as well as the related geometric parameters listed in Table 3, revealed that both 2M2B and T could be physisorbed on the acid site of the catalyst cluster by forming π-complexes, for the interaction between adsorbates and the catalyst did not significantly affect their structures. As shown by the physisorption of 2M2B on the SPAM cluster in Figure 3a, the double bond C2dC3 was stretched from 1.344 to 1.351 Å and the length of acid bond O1—H1 was increased from 0.968 to 0.985 Å, indicating that the adsorption slightly weakened the double bond and the acid bond, which may lead to the protonation of 2M2B. The proton H1 was inclined to approach the C3 atom on the double bond, which was conformed to the general chemical theory. When an alkene is protonated, the proton attaches to the least substituted carbon. By comparison, the physisorption had relatively little impact on the geometric structure of T (Figure 3b), as the extensions of the bonds O1H1 and C6C7 on the π-complex of T were shorter than those on the π-complex of 2M2B, 0.007 and 0.002 Å, respectively. Moreover, the other geometric structures of T were almost unchanged. This may be caused by the aromatic character of the thiophene ring, which led thiophene to be physisorbed on the acid site of catalyst easily and was not beneficial to the thiophene protonation. The catalyst proton H1 tended to be close to the C6 atom on the α site of T because of the electrondonating effect of the S atom. On the basis of the above analysis, the weakly adsorbed 2M2B can be further protonated by the acidic proton H1 on the catalyst. However, it was not stable in a carbonium ion-like form and quickly transformed into a stable alkoxide intermediate (σ-complex of 2M2B on the SPAM, Figure 3c) by forming the covalent bond C2—O2 to one of the bridging oxygen atoms on the catalyst. The calculation showed that it needed to form a transition state (Figure 3d) during the transformation process from π- to σ-complex of 2M2B on the SPAM cluster. There was 3943

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Figure 3. Optimized geometries for the adsorption of different reactants on the SPAM cluster: (a) π-complex of 2M2B on the SPAM cluster, (b) π-complex of T on the SPAM cluster, (c) σ-complex of 2M2B on the SPAM cluster, and (d) transition state (TS) of 2M2B from π- to σ-complex on the SPAM cluster and its vibrational motion.

Table 3. Most Important Geometric Characteristics of Different Reactants and Transition State (TS) during Adsorption on the SPAM Clustera 2M2B + SPAM

TS of 2M2B from π- to

σ-complex of 2M2B

on SPAM

σ-complex on SPAM

on SPAM

π-complex of T + SPAM

T on SPAM

O1H1

0.968

0.985

1.520

2.538

O1H1

0.968

0.978

O1P

1.623

1.617

1.530

1.483

O1P

1.623

1.612

O2P O1PO2

1.473 114.6

1.476 116.0

1.508 116.8

1.586 118.6

O2P O1PO2

1.473 114.6

1.476 115.5

O3PO4

103.4

103.2

102.5

101.5

O3PO4

103.4

103.2

C2C3

1.344

1.351

1.428

1.540

C6H3

1.081

1.081

C3H2

1.091

1.092

1.097

1.095

C6S

1.735

1.735

O2C2

3.567

2.405

1.490

C6C7

1.370

C3H1

2.193

1.204

1.098

C6H1

C2H1 C1C2C3 C2C3C4 a

π-complex of 2M2B

2.354 120.8 128.6

120.9 128.5

2.033

C7H1

119.2 120.9

111.0 116.0

SC6H3 H3C6C7

1.375 2.380 2.414

120.2 128.1

120.2 128.1

Distances in angstroms; angles in degrees.

only one imaginary frequency (129 cm1) at the transition state corresponding to the following movements: the acidic proton H1 on the catalyst was moving toward the C3 atom of 2M2B, while the double bond C2dC3 was elongating from

1.351 to 1.428 Å and the other carbon atom C2 was moving toward the adjacent O2 atom on the SPAM cluster to form a covalent bond C2—O2. The formation of the alkoxide complex was accompanied by structural changes in the active site of 3944

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Figure 4. Energy changes of different reactants and the reaction coordinate for the transition state (TS) formation during the adsorption process of 2M2B on the SPAM cluster.

SPAM (Table 3) for the fracture of acid bond O1—H1 and the formation of covalent bond C2—O2. The bond O1—P and bond angle O3—P—O4 were decreased by 0.134 Å and 1.7°, respectively, while the bond O2—P and bond angle O1—P— O2 were increased by 0.110 Å and 2.6°, respectively. Figure 4 displays the calculated energy profiles of the adsorption of the reactants and the transformation from the π- to the σ-complex of 2M2B on the SPAM cluster. The adsorption energy ΔEads for different adsorption statuses of reactants on the SPAM was calculated by eq 1, where Eadsorbate/SPAM represents the total energy of the reactant adsorbed on the SPAM cluster. ESPAM and Eadsorbate represent the energies of the separated SPAM cluster and the adsorbate structure, respectively. ΔEads ¼ Eadsorbate=SPAM  ESPAM  Eadsorbate

ð1Þ

As shown in Figure 4, all the ΔEads values are negative, which indicates that the lower the value is, the easier the adsorption is. It seems that T may compete with 2M2B to adsorb on the SPAM at the initial stage of adsorption, as the calculated adsorption energy of the π-complex of 2M2B was a little lower than that of T, 9 kJ/ mol. Although the chemisorption of 2M2B was more stable than the physisorption of T (the adsorption energy of the former was lower than that of the later, 42 kJ/mol), the transformation from π- to σ-complex of 2M2B needed time and energy. The calculated activation energy Eact for the transformation was 99 kJ/mol. However, the content of thiophenic compounds was tiny (about 0.034 wt %) compared with the content of alkenes in real FCC gasoline feed (Table 1). It can be considered that almost all of the acidic active sites on the catalyst surface were occupied by 2M2B in the form of π-complexes at the initial stage of adsorption; then the physisorption of 2M2B could be transformed into a stable alkoxide intermediate on the SPAM cluster (σ-complex of 2M2B on the SPAM) by adsorbing enough energy. Therefore, most of active sites on the SPAM were occupied by the stable alkoxide intermediate of 2M2B at the end of the adsorption of the reactants. 3.2.2. Theoretical Investigation on Reactions in the OATS Process over SPAM. Under the premise of explaining the catalytic mechanism for the main and side reactions in the OATS process concisely and effectively, we just studied the reaction mechanisms for the monoalkylation of T with 2M2B and the self-dimerization of 2M2B (reaction formulas as

shown in formulas 2 and 3) over the SPAM by the method of theoretical calculation.

Previous studies have proposed two kinds of mechanisms for the alkylation reaction, such as the alkylation of benzene with short chain alkenes and the dimerization of linear alkenes.16,19 The difference in the two kinds of mechanisms is only whether or not a stabilized alkoxide intermediate is formed during the course of the reaction. Moreover, our calculation for the alkylation mechanism of T with 2M2B over the SPAM shows that the alkylthiophene was formed only after 2M2B interacted with acid sites on the SPAM, because the opposite scenario, T physically adsorbed on acid sites and 2M2B coadsorbed on the π-complex of T, was also evaluated. While this configuration cannot be ruled out on the basis of adsorption energies, a physically meaningful transition state for the formation of alkylthiophene could not be obtained in this case, and the calculation for adsorption indicated that almost all the active sites on the catalyst surface were occupied by 2M2B in the form of σ-complexes after the adsorption of the reactants in real gasoline. Therefore, the reaction mechanism of alkylation and dimerization were both investigated from a stable alkoxide intermediate of 2M2B on the SPAM (denoted as alkylR) at the beginning of the reaction stage. The calculated energy profile for the reaction between the alkoxide complex of 2M2B and T to produce alkylthiophene is shown in Figure 5a. The optimized structures of the involved stationary points are displayed in Figure 6a, c, and e. Correspondingly, the major structural parameters of the geometric structures are tabulated in Table 4. These revealed that further progress of alkylation reaction required the physisorption of a second reactant (T) to the alkoxide group of 2M2B on the SPAM cluster (adsT on alkylR, as shown in Figure 6a), as the adsorption process reduced the total energy of the system about 26 kJ/mol. During the adsorption, T was diffused into the vicinity of the alkoxide group and interacted weakly with alkylR. The covalent 3945

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Figure 5. Calculated energy profiles for the alkylation and dimerization over the SPAM cluster: (a) alkylation of T with the σ-complex of 2M2B and (b) dimerization of 2M2B on the SPAM cluster.

bond O2C2 increased by 0.006 Å, and the bond angle PO2C reduced by 0.9° on the alkylR, while the bond O2P on the active site of the cluster increased by 0.004 Å. After the second reactant T was adsorbed on alkylR, the alkylation between the adsorbed T and alkylR started by adsorbing enough energy. The thiophene alkylation involved the forming of a covalent bond C2C6 between the carbon atoms of 2M2B and T and the proton giving back to the adjacent oxygen atom O1 on the catalyst by breaking the proton H3 on the electron-rich site of T (α site). The whole process was clearly displayed by the vibrational motion corresponding to the single imaginary frequency (186 cm1) at the transition state for alkylation (TS1) in Figure 6c. During the transformation, the covalent bond C2O2 of the alkoxide complex was broken by stretching to form a carbonium ion-like transition state that was not stable and continually attacked the adsorbed T until the bond C2C6 between isopentyl carbocation and thiophene began to form and the proton H3 on thiophene left toward the SPAM cluster. The calculated activation energy (Eact) for the process was 170 kJ/mol. Subsequently, the alkylthiophene (Figure 6e) was desorbed from the SPAM cluster endothermically, which required the desorption energy (ΔEdes) of at least 13 kJ/mol. The energy profile is displayed in Figure 5b combined with the corresponding structures and geometric parameters in Figure 6b, d, and f and Table 4, revealing that the reaction path for the

self-dimerization of 2M2B was very similar to that for the alkylation of T. The dimerization also started from a stable alkoxide intermediate of 2M2B on the SPAM cluster (alkylR); then, a second reactant molecule 2M2B was physically adsorbed on the alkylR (Figure 6b). An isopentyl carbocation was also found in the transition state (TS2) for the self-dimerization of 2M2B (Figure 6d). The vibrational motion corresponding to an imaginary frequency (269 cm1) described the formation of a new covalent bond C2C9 between the carbocation and the least substituted carbon atom C9 on the double bond of 2M2B. The similarity of the reaction paths led us to decide that the olefin oligomerization was an unavoidable side reaction during the OATS process catalyzed by the SPAM. However, there were still some obvious differences between the thiophenic compound alkylation and alkene oligomerization. First, as shown in Figure 5, the energy change after the adsorption of 2M2B on alkylR was lower than that of T, 15 kJ/mol, indicating that T was more inclined to be adsorbed on the alkoxide complex than 2M2B. Moreover, the activation energy for the alkylation of T was obviously less than that for the 2M2B dimerization, 47 kJ/mol, which demonstrated that the alkylation rate of thiophenic compounds with olefins was faster than the self-dimerization rate of alkenes. So, the adsorption rates of the second reactant and the different reaction rates led us to determine that the alkylation of thiophenic sulfurs happened more easily, as compared 3946

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Figure 6. Optimized geometric structures of different stationary states: (a) adsorption of T on the σ-complex of 2M2B on the SPAM cluster (adsT on alkylR), (b) adsorption of 2M2B on the σ-complex of 2M2B on the SPAM cluster (ads2M2B on alkylR), (c) transition state for the alkylation of T and 2M2B on the SPAM cluster (TS1) and its vibrational motion, (d) transition state for dimerization of 2M2B on the SPAM cluster (TS2) and its vibrational motion, (e) alkyl-T on the SPAM cluster, and (f) dimer-2M2B on the SPAM cluster.

to the dimerization of alkenes. This theoretical result was conformed to the experimental outcome in section 3.1, as the conversion rate of thiophenic sulfur compounds at the initial stage of reaction was obviously faster than those of aromatics and various alkenes under every investigated temperature (Figure 1 and Table 2). The consistency of the results obtained by theoretical calculation and experimentation could be used to prove the correctness of the theoretical calculation results. Second, the calculated results in Figure 5 also showed that the alkylation of T was an exothermic process whereas the

dimerization of 2M2B was an endothermic reaction, which was consistent with the thermodynamic calculation results obtained by Belliere et al.9 as well as the experiment results about the effect of temperature on the catalyst performance in section 3.1. The alkylation reaction was favored at low temperatures, and rising temperature could improve the catalyst activity over the range of relatively low temperatures, for the high temperature could enhance the molecules movement and increase the adsorption rate of reactants on the active sites of the catalyst. Moreover, T was more inclined to be adsorbed on the alkoxide 3947

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels

ARTICLE

Table 4. Part Geometric Parameters of Adsorption Complexes of Reactants, Transition States (TS), and Products during Alkylation and Dimerization on the SPAM Clustera alkyl-T T + alkylR

adsT on alkylR

TS1

O1P O2P

1.483 1.586

1.481 1.590

1.521 1.505

O2C2

1.490

1.496

3.895

C2C3

1.540

1.540

1.493

1.559

C3H1

1.095

1.094

1.104

C3H2

1.098

1.098

1.105

ads2M2B

on SPAM 1.611 1.476

2M2B + alkylR

on alkylR

dimer- 2M2B TS2

O1P O2P

1.483 1.586

1.483 1.586

1.516 1.505

O2C2

1.490

1.495

3.936

C2C3

1.540

1.540

1.500

1.562

1.097

C3H1

1.095

1.095

1.104

1.092

1.099

C3H2

1.098

1.098

1.108

1.610 1.475

1.098

O1PO2

118.6

118.1

119.2

115.7

O1PO2

118.6

118.4

120.0

116.2

O3PO4

102.5

103.0

100.5

103.2

O3PO4

102.5

102.7

100.1

103.2

PO2C2 C6C7

126.7 1.370

125.8 1.368

1.399

PO2C2 C8C9

126.7 1.344

126.1 1.345

1.427

1.532

1.081

1.081

1.099

C9H5

1.091

1.090

1.087

1.096

C10H6

1.098

1.093

1.102

2.166

1.511

1.512

1.475

1.347

C6H3

1.376

C8H6 C8C10

2.667

O1H3

2.261

1.955

0.980

O1H6

2.718

2.162

0.982

C2C6

5.720

2.329

1.527

C2C9

5.548

2.455

1.578

6.435 179.8

2.950 152.2

C2C8 C10C8C9H5

6.073 0.1

2.866 3.0

8.0

C2C7 SC6C7H3 a

on SPAM

180.0

0

Distances in angstroms; angles in degrees.

complex than 2M2B during the adsorption of the second reactant for the further reaction, and the alkylation rate of T was faster than that of 2M2B dimerization. Therefore, increasing temperatures could improve the catalytic activity for thiophenic compound alkylation over a range of relatively low temperatures. However, if the temperature was too high, the catalytic activity for the alkylation would be suppressed, as the reaction was an exothermic process. On the contrary, the oligomerization of alkenes as a side reaction was promoted at very high temperatures, which could cause the reduction of the catalyst selectivity. This experimental result further proved the correctness of the theoretical calculations for the reaction mechanisms of typical main and side reactions in the OATS process over the SPAM. On the basis of the above theoretical investigation on the catalytic mechanism of SPAM in the OATS process, two methods that might improve the catalyst performance were proposed. First, reduce the redundant active sites on the catalyst appropriately because there were only strong acids on the catalyst12 and both main and side reactions started from a stable alkoxide intermediate, which was formed by the protonation of olefin adsorbed on the active sites of SPAM. In the further reaction, thiophenic sulfurs were more inclined to be adsorbed on the alkoxide complex than alkenes. Moreover, the activation energy with relatively low value made the thiophenic sulfur alkylation happen easily. However, the content of the sulfur compounds was tiny in the feed, and the demand for stable alkoxide intermediates for the alkylation of thiophenic compounds was also tiny. If there were excessive acid sites on the catalyst for the formation of alkoxide complexes of 2M2B, the redundant alkoxide complexes would take part in the side reactions of alkene oligomerization, which was adverse to the improvement of the catalyst selectivity. Therefore, the decrease of redundant active sites on the catalyst may be an effective way to reduce the side reactions. Second, control the reaction

temperature strictly; because the alkylation of thiophenic sulfurs as the main reaction was exothermic while the oligomerization of alkenes as the side reaction was endothermic, the temperature had a great effect on the catalyst activity and selectivity. The experimental results showed that the optimal temperature range for the alkylation of thiophenic compounds catalyzed by the SPAM in FCC gasoline was 140150 °C to achieve a better catalytic selectivity.

4. CONCLUSIONS Both the experimental and theoretical investigations were carried out to study the catalytic mechanism of SPAM during the FCC gasoline desulfurization by alkylation. The DFT calculation for the reaction paths showed that both main and side reactions in the OATS process over the SPAM started from a stable alkoxide intermediate, which was formed by protonation of olefin adsorbed on the catalyst. Thiophenic compounds were more inclined to be adsorbed on the alkoxide complex than olefins for the further reaction, and the activation energy for alkylation of thiophenic sulfurs with alkenes was obviously lower than that for alkene oligomerization. The alkylation rate of thiophenic sulfur compounds was faster than the reaction rate of alkene oligomerization over the catalyst. Moreover, the thiophene alkylation was exothermic while the olefin oligomerization was endothermic. The theoretical calculation results were very consistent with the experimental outcomes about the effect of temperature on the conversion curves of different reactants during the OATS process of FCC gasoline over the SPAM, demonstrating that the theoretical calculation for the reaction mechanisms was reliable. On the basis of these experimental and theoretical findings, two methods that might inhibit the occurrence of side reactions and be advantageous to the further application of SPAM in the alkylation desulfurization of FCC gasoline 3948

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949

Energy & Fuels were proposed as follows, and they will be proved in further research. (1) Reduce the redundant active sites on the catalyst appropriately. (2) Control the reaction temperature strictly, and the optimal temperature range for the alkylation of thiophenic compounds catalyzed by the SPAM in FCC gasoline was 140150 °C.

ARTICLE

(17) Shon, J. K.; Yuan, X. D.; Ko, C. H.; Lee, H. I.; Thakur, S. S.; Kang, M.; Kang, M. S.; Li, D. H.; Kim, J. N.; Kim, J. M. J. Ind. Eng. Chem. 2007, 13, 1201–1207. (18) Kovalchuk, T. V.; Sfihi, H.; Korchev, A. S.; Kovalenko, A. S.; Il'in, V. G.; Zaitsev, V. N.; Fraissard, J. J. Phys. Chem. B 2005, 109, 13948–13956. (19) Hansen, N.; Br€uggemann, T.; Bell, A. T.; Keil, F. J. J. Phys. Chem. C 2008, 112, 15402–15411.

’ AUTHOR INFORMATION Corresponding Author

*Telephone: 86-22-27406795. Fax: 27403389. E-mail: yhli@ tju.edu.cn.

’ ACKNOWLEDGMENT The authors gratefully acknowledge the financial support provided by the National Natural Science Foundation (No. 20976129) and the Program of Introducing Talents of Discipline to Universities of China (No. 06006). ’ REFERENCES (1) Brunet, S.; Mey, D.; Perot, G.; Bouchy, C.; Diehl, F. Appl. Catal., A 2005, 278, 143–172. (2) Song, C. S. Catal. Today 2003, 86, 211–263. (3) Babich, I. V.; Moulijn, J. A. Fuel 2003, 82, 607–631. (4) Jaimes, L.; Tonetto, G.; Lujan, M.; de Lasa, H. Int. J. Chem. React. Eng. 2008, 6, 1–66. (5) Ito, E.; van Veen, J. A. R. Catal. Today 2006, 116, 446–460. (6) Zhang, Z. K.; Guo, X. Y.; Liu, S. L.; Zhu, X. X.; Xu, L. Y. Fuel Process. Technol. 2008, 89, 1135–1141. (7) Jaimes, L.; Ferreira, M. L.; de Lasa, H. Chem. Eng. Sci. 2009, 64, 2539–2561. (8) Zheng, X. D.; Dong, H. J.; Wang, X.; Shi, L. Catal. Lett. 2009, 127, 70–74. (9) Belliere, V.; Geantet, C.; Vrinat, M.; Ben-Ta^arit, Y.; Yoshimura, Y. Energy Fuels 2004, 18, 1806–1813. (10) Belliere, V.; Lorentz, C.; Geantet, C.; Yoshimura, Y.; Laurenti, D.; Vrinat, M. Appl. Catal., B 2006, 64, 254–261. (11) Shi, R. H.; Li, Y. H.; Wang, R.; Guo, B. S. Catal. Lett. 2010, 139, 114–122. (12) Wang, R.; Li, Y. H. Catal. Commun. 2010, 11, 705–709. (13) Cai, Q.; Lin, W. Y.; Xiao, F. S.; Pang, W. Q.; Chen, X. H.; Zou, B. S. Microporous Mesoporous Mater. 1999, 32, 1–15. (14) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03, Revision C.01; Gaussian Inc.: Wallingford, CT, 2004. (15) Sousa, S. F.; Fernandes, P. A.; Ramos, M.J. J. J. Phys. Chem. A 2007, 111, 10439–10452. (16) Svelle, S.; Kolboe, S.; Swang, O. J. Phys. Chem. B 2004, 108, 2953–2962. 3949

dx.doi.org/10.1021/ef200705b |Energy Fuels 2011, 25, 3940–3949