Adsorption Properties of Organosulfur Compounds on Zeolite Clusters

Nov 7, 2008 - D.L.: e-mail [email protected], phone +82-31-280-9324. ... The binding energies of tetrahydrothiophene (C4H8S, THT), dimethyl sul...
0 downloads 7 Views 1MB Size
J. Phys. Chem. C 2008, 112, 18955–18962

18955

Adsorption Properties of Organosulfur Compounds on Zeolite Clusters: A Density Functional Theory Calculation Study Doohwan Lee,*,† Jongseob Kim,*,† Hyun Chul Lee,† Kang Hee Lee,† Eun Duck Park,‡ and Hee Chul Woo§ Samsung AdVanced Institute of Technology (SAIT), Mt. 14-1, Nongseo-Dong, Giheung-Gu, Yongin, Republic of Korea 446-712, DiVision of Energy Systems Research, Ajou UniVersity, Wonchun-Dong, Yeongtong-Gu, Suwon, Republic of Korea 443-749, and DiVision of Chemical Engineering, Pukyong National UniVersity, San 100, Yongdang-Dong, Nam-Gu, Pusan, Republic of Korea 608-739 ReceiVed: May 20, 2008; ReVised Manuscript ReceiVed: September 24, 2008

The effects of cations of zeolite for the adsorption of organosulfur compounds were investigated by using density functional theory calculations. The binding energies of tetrahydrothiophene (C4H8S, THT), dimethyl sulfide (C2H6S, DMS), tert-butylmercaptan (C4H10S, TBM), hydrogen sulfide (H2S), and carbonyl sulfide (COS) on the zeolite model clusters [X(HO)3SiOAl(OH)3, X ) H+, Na+ and Ag+] were obtained and compared with those of H2O, CO2, and C1-C3 light hydrocarbons. Compared to the H+ and Na+ cations, the Ag+ cation induces much stronger binding of THT, DMS, TBM, and H2S over H2O suggesting great enhancements in the adsorption selectivity. The order of binding energies of these sulfur compounds is THT > DMS > TBM > H2S > COS, and it does not depend on the cation types. These results agree well with the experimental adsorption uptake and selectivity properties of AgNaY zeolites for organosulfur compounds. 1. Introduction Adsorptive deep desulfurization of hydrocarbon fuels is an alternative method that has been of much interest in recent years for its simplicity and effectiveness over the hydrodesulfurization method.1-5 The adsorption method selectively removes sulfur compounds utilizing adsorbents at ambient temperatures and pressures, in a much simplified way compared to the hydrotreating method that requires complex catalytic conversion to H2S and subsequent chemical fixation on zinc oxides sorbents.6 The adsorption method is particularly favorable for fuel cell applications, where it can enable a great simplification of the system preventing sulfur poisoning of the Pt electrodes. Adsorptive desulfurization has been extensively investigated for hydrocarbon fuels utilizing various metal [Na(I),5,7-10 Ca(II),8,9 Ni(II),2,11,12 Cu(I),1,3,11,13 Cu(II),2,12,14 Zn(II),2,11,12 Pd(II),2 Ag(I),1,4,5,9,12,14,15 and Ce(III)2,14] loaded zeolites (5A,16 ZSM-5,17,18 X,7,19 and Y1-5,7,11,13,14), porous carbons,20,21 and metal oxides22-24 with a particular focus on thiophenic species, refractory compounds difficult to remove by the hydrotreating method. Several studies have explored adsorptive removal of sulfur containing odorants (thiophenes, sulfides, and thiols) from gaseous hydrocarbons utilizing zeolites4,5,8-10,15,25 and activated carbons.26 These studies have shown that sulfur species could be captured reducing the sulfur concentration in the effluents even in parts per billion levels at ambient conditions. In particular, the group of Yang reported that Ag(I) and Cu(I) cation types of Y-zeolites gave superior adsorption capacity and selectivity for thiophenic species via strong π-complexation.1,3,11,13 Satokawa et al. investigated adsorptive removal of dimethyl sulfide (DMS) and tert-butylmercaptan (TBM) odorants from * Corresponding authors. D.L.: e-mail [email protected], phone +82-31-280-9324. J.K.: e-mail [email protected], phone +82-31-280-8168. † SAIT. ‡ Ajou University. § Pukyong National University.

city gas (natural gas) using AgNa-Y zeolites, and reported a significant enhancement in TBM adsorption uptake through silver sulfides formation.4,15 We also reported adsorption properties of tetrahydrothiophene (THT) and TBM on AgNa-Y with detailed site-by-site evaluations of the Ag+, Ag0, H+, and Na+ cations for adsorption contributions.5 Briefly, the Ag+ sites enhanced TBM breakthrough uptake, but its adsorption selectivity was nearly zero in the copresence of THT in the feed stream. In addition, an ion exchange of Na+ with Ag+ did not have an effect on THT adsorption uptake. Our recent experimental studies also showed that adsorption of these sulfur compounds could be inhibited significantly in the presence of H2O and other impurities. These previous studies suggest that adsorption selectivity is a critical factor for desulfurization of hydrocarbon fuels that contain multiple sulfur compounds and antagonizing impurities, but systematic evaluations of various adsorption sites for adsorption strengths and contributions have been rarely presented. In this work, we take a quantum mechanical calculation approach to investigate adsorption properties of common organosulfur compounds (THT, TBM, DMS, H2S, and COS) as well as light C1-C3 hydrocarbons, H2O, and CO2 on model zeolite cluster systems with a particular focus on the effects of H+, Na+, and Ag+ cations. The reason to choose these compounds is that THT, TBM, and DMS (selected from alkylthiophenes, thiols, and sulfides) are widely used sulfur odorants for light hydrocarbon fuels that are a cost-effective H2 source for fuel cells, while CO2, H2O, H2S, and COS are common impurities. Several previous calculation studies on adsorption of individual sulfur species (thiophene,11,27 CH3SH,28,29 C2H5SH,30 H2S31) as well as light hydrocarbons,32-34 H2O,35 and CO236 on zeolites can be found in the literature, but their diverse cluster models and levels in the theories taken for the calculations limit a direct and meaningful comparison of the results. Here we present the binding energies of the compounds on simple zeolite model clusters obtained with highly accurate

10.1021/jp804441q CCC: $40.75  2008 American Chemical Society Published on Web 11/07/2008

18956 J. Phys. Chem. C, Vol. 112, No. 48, 2008

Lee et al.

Figure 1. The optimized structures of the zeolite model clusters: (a) H(HO)3SiOAl(OH)3 [HZ], (b) Na(HO)3SiOAl(OH)3 [NaZ], and (c) Ag(HO)3SiOAl(OH)3 [AgZ]. Mulliken and ESP charges are also shown in bold fonts. ESP charges are in parentheses.

B3LYP and MP2 methods, and discuss the results comparing with experimental sulfur adsorption uptake and selectivity properties of AgNaY zeolites. 2. Calculation Method To study the effects of cations of zeolite for adsorption of the organosulfur compounds, we obtained the binding energies (BEs) between the model zeolite systems and the sulfur compounds through ab initio calculations. As the model systems for zeolite (Figure 1), we used X(HO)3SiOAl(OH)3 [X ) H, Na, Ag], and these were referred to as XZ [X )H, Na, and Ag], respectively. These clusters consisted of a Al-O tetrahedral unit bridged by an oxygen atom with a Si-O tetrahedral unit, and a charge balancing cation (X) resided above the bridging O atom. The dangling O atoms of the clusters were terminated with hydrogen atoms. The structures were fully optimized without any symmetry constraints at the level of density functional theory (DFT), using 6-311+G(2d,p) basis sets. For Ag, Stuttgart RSC 1997 basis sets employing relativistic core potential were used.37 The DFT calculations were conducted with use of Becke’s three-parameter exchange functional together with the correlation functionals of Lee-Yang-Parr (B3LYP).38,39 Vibration frequency analysis was also carried out to confirm whether the optimized geometries were the true minimum energy structures. For more conclusive results, singlepoint Møller-Plesset second-order perturbation (MP2) calculations were further performed on the B3LYP (MP2//B3LYP) predicted geometries. The energies were further modified by using basis set superposition error correction (BSSEC)40 and zero-point energies. Generally, the BEs without BSSEC tend to be overestimated whereas the BEs with full BSSEC, although they are accepted to be more reliable, tend to be underestimated. Therefore, in this study, we denote ∆E as the median value of BSSE corrected and uncorrected BEs plus/minus a half-value

of BSSE, which represents the upper/lower bound of BEs.41 To check the cluster size dependency, we performed two-layer ONIOM (QM/MM) calculations with much larger zeolite systems using B3LYP/6-31+G** as the quantum mechanical part and universal force field as the molecular mechanics part. The details will be discussed later in the Discussion section. All the calculations were performed with Gaussian program suites.42 3. Results 3.1. Zeolite Model Clusters. Figure 1 shows the optimized structures and atomic charges of the zeolite model clusters. Table 1 summarizes the bond lengths and atomic charges of the relevant atoms of the organosulfur compounds and H2O. On the HZ cluster, the charge compensating H+ cation was bonded to the bridging O2 atom with an atomic distance of 0.97 Å. On the NaZ cluster, the Na+ cation resided 2.32 Å above the O2, and 2.31 and 2.19 Å away from the O1 and O3 atoms. On the AgZ cluster, the Ag+ cation resided 2.73 Å above the O2, and 2.30 and 2.22 Å away from the O1 and O3 atoms. These cations showed slight orientations toward the O3, and the order of atomic distance (Å) from the cation to the bridging O2 atom was Ag-O2 (2.73) > Na-O2 (2.32) > H-O2 (0.97). Natural bond order calculations indicated that the H+ cation was closely bonded to the O2 atom, whereas the Na+ and the Ag+ cations were placed in extra-framework of the clusters. The order of Si-O1 and Al-O3 bond lengths (Å) followed the same order as the cation: AgZ (Si-O1 ) 1.70, Al-O3 ) 1.82) > NaZ (Si-O1 ) 1.69, Al-O3 ) 1.80) > HZ (Si-O1 ) 1.63, Al-O3 ) 1.72). The order of atomic charge on the cations was Na (+0.759) > Ag (+0.390) > H (+0.302), and that on the counter O2 atom was NaZ (-1.089) > AgZ (-0.936) > HZ (-0.832). In the case of the charges of sulfur atoms, the sulfur atoms of THT, DMS, TBM, and H2S were negatively charged, whereas

Organosulfur Compounds on Zeolite Clusters

J. Phys. Chem. C, Vol. 112, No. 48, 2008 18957

TABLE 1: Mulliken and ESP Charges and Bond Lengths on the Relevant Atoms of the Sulfur Compounds and H2Oa relevant atom

THT

DMS

TBM

H2S

COS

H2O

O

-0.254 (-0.302) -0.038 (0.061) -

0.074 (-0.238) -0.048 (-0.103) -

-0.330 (-0.397) 0.694 (0.429) 0.076 (0.175) -

-0.216 (-0.283) 0.108 (0.141) -

-0.046 (-0.090) 0.206 (0.322) -0.160 (-0.232)

0.272 (0.378) -0.545 (-0.756)

bond length (Å) S-C, S-H*, or O-H//

1.85

1.82

1.87

1.34*

1.57

0.96**

atomic charge S C H

a

ESP charges are shown in parentheses.

Figure 2. The optimized structures of the zeolite cluster [X(HO)3SiOAl(OH)3 ]-adsorbate complexes.

it was positively charged in COS due to a higher electronegativity of the O atom. 3.2. Adsorption of the Organosulfur Compounds and H2O. Figure 2 shows the optimized structures of THT, DMS, TBM, H2S, and COS on the HZ, NaZ, and AgZ zeolite clusters, respectively. The results clearly indicate that the negatively charged S atom of THT, DMS, TBM, and H2S interacted primarily with the H+, Na+, and Ag+ cation of the clusters, and this gives rise to stabilization of the zeolite cluster-adsorbate complexes to their minimum energy conformations. Adsorption of H2O showed a close resemblance with H2S, while COS stabilized by charge interactions between the cation and the O atom. The spatial orientations of the adsorbates on the clusters differed upon the nature of the cations. On the HZ and NaZ clusters, the sulfur compounds stabilized showing orientations toward the Al-O tetragonal unit with close proximities of the S atom to the H+ and Na+ cations, whereas on the AgZ cluster, the compounds stabilized displaying orientations toward the Si-O tetragonal unit. Table 2 shows B3LYP and MP2//B3LYP predicted negative binding energies (BEs) without and with the zero-point vibrational energy correction denoted by ∆Ee and ∆E0, respectively.

Hereafter, our discussions will be based on the results of ∆E0 (MP2//B3LYP). The results showed that the molar binding energies (-∆E0 in kJ mol-1) of the compounds on the HZ cluster were moderate, and the order was H2O (54.9) > THT (50.3) g DMS (50.1) > TBM (43.0) > H2S (30.3) > COS (19.1). Table 3 shows the structural properties of the corresponding geometry optimized HZ-adsorbate complexes for THT, DMS, TBM, H2S, COS, and H2O. The atomic distance (Å) between the H+ cation and the S atom in the cluster-adsorbate complexes decreased with an order of H2S (2.21) > TBM (2.14) > DMS (2.12) > THT (2.08), and this was exactly the reverse order to their binding energies. The bond distance between the O2 atom and the H+ cation (O2-H+) barely changed, but the O1-H+ and O3-H+ atomic distances increased slightly due to openings of the O2-Al-O3 and O1-Si-O2 bond angles upon binding of the sulfur compounds. The sulfur-carbon (S-C) or sulfur-hydrogen (S-H) bond length (Å) of the adsorbates (THT ) 1.86, DMS ) 1.82, TBM ) 1.88, H2S ) 1.36) practically did not change from those of the free molecules (THT ) 1.85, DMS ) 1.82, TBM ) 1.87, H2S ) 1.34). Adsorption configuration of H2O resembled H2S, but its

18958 J. Phys. Chem. C, Vol. 112, No. 48, 2008

Lee et al.

TABLE 2: Adsorption Energies of Molecules on the Zeolite Clusters (in kJ mol-1)a THT

DMS

TBM

H2S

COS

H2O

∆Ee (B3LYP) ∆E0 (B3LYP) ∆Ee (MP2//B3LYP) ∆E0 (MP2//B3LYP)

-37.7 ( 2.1 -35.3 ( 2.1 -52.7 ( 6.7 -50.3 ( 6.7

HZ [H(HO)3SiOAl(OH)3] -42.4 ( 1.6 -32.4 ( 2.1 -36.6 ( 1.6 -28.8 ( 2.1 -55.9 ( 5.2 -46.5 ( 6.4 -50.1 ( 5.2 -43.0 ( 6.4

-30.4 ( 1.7 -24.1 ( 1.7 -36.6 ( 4.8 -30.3 ( 4.8

-12.8 ( 1.3 -10.9 ( 1.3 -21.0 ( 3.6 -19.1 ( 3.6

-63.2 ( 2.8 -51.7 ( 2.8 -66.4 ( 6.7 -54.9 ( 6.7

∆Ee (B3LYP) ∆E0 (B3LYP) ∆Ee (MP2//B3LYP) ∆E0 (MP2//B3LYP)

-46.7 ( 2.0 -44.6 ( 2.0 -58.2 ( 6.3 -56.1 ( 6.3

NaZ [Na(HO)3SiOAl(OH)3] -39.2 ( 1.6 -39.2 ( 1.9 -36.5 ( 1.6 -36.8 ( 1.9 -48.2 ( 5.2 -47.7 ( 5.8 -45.5 ( 5.2 -45.3 ( 5.8

-27.3 ( 1.6 -22.2 ( 1.6 -31.4 ( 4.5 -26.2 ( 4.5

-22.0 ( 1.9 -21.1 ( 1.9 -20.4 ( 3.5 -19.5 ( 3.5

-60.6 ( 2.6 -52.0 ( 2.6 -61.7 ( 5.6 -53.1 ( 5.6

∆Ee (B3LYP) ∆E0 (B3LYP) ∆Ee (MP2//B3LYP) ∆E0 (MP2//B3LYP)

-100.4 ( 2.3 -95.9 ( 2.3 -99.4 ( 14.2 -95.0 ( 14.2

AgZ [Ag(HO)3SiOAl(OH)3] -98.0 ( 2.0 -89.1 ( 2.2 -92.4 ( 2.0 -85.3 ( 2.2 -94.5 ( 13.1 -90.4 ( 14.2 -88.9 ( 13.1 -86.6 ( 14.2

-80.9 ( 2.3 -68.8 ( 2.3 -67.6 ( 10.8 -55.5 ( 10.8

-14.2 ( 1.2 -16.0 ( 1.2 -19.4 ( 6.6 -21.2 ( 6.6

-74.3 ( 2.9 -58.3 ( 2.9 -65.0 ( 10.3 -49.0 ( 10.3

a See the text for notations. Here, ∆Ee and ∆E0 are negative binding energies without and with zero-point vibrational energy corrections, respectively.

TABLE 3: Relevant Atomic Distances, Bond Lengths, and Bond Angles of the Geometry Optimized HZ-Adsorbate Complexesa THT O1-H+ O2-H+ O3-H+ Si-O1 Si-O2 Al-O2 Al-O3 H+-S or H+-O* S-C, S-H*, C-O**, or O-H† O1-Si-O2 Si-O2-Al O2-Al-O3 a

2.82 1.01 2.87 1.63 1.69 1.91 1.73 2.08 1.86(1.85) 103.6 127.2 100.1

DMS

TBM

atomic distance/bond length (Å) 2.88 2.74 1.00 1.00 3.00 2.77 1.63 1.63 1.72 1.70 1.90 1.91 1.72 1.74 2.12 2.14 1.82(1.82) 1.88 bond angle (deg) 109.7 103.2 117.3 126.9 106.8 98.9

H2S 2.72 0.99 2.76 1.63 1.70 1.92 1.74 2.21 1.36(1.35)* 103.0 126.7 98.6

COS 2.69 0.97 2.66 1.63 1.70 1.93 1.73 1.92* 1.16** 102.7 127.7 96.8

H2O 2.74 1.02 2.64 1.63 1.69 1.90 1.76 1.57* 0.99(0.96)† 103.4 126.8 97.1

See the text and Figure 1 for notations.

TABLE 4: Relevant Atomic Distances, Bond Lengths, and Bond Angles of the Geometry Optimized NaZ-Adsorbate Complexesa THT O1-Na+ O2-Na+ O3-Na+ Si-O1 Si-O2 Al-O2 Al-O3 Na+-S or Na+-O* S-C, S-H*, C-O**, or O-H† O1-Si-O2 Si-O2-Al O2-Al-O3 a

2.36 2.38 2.24 1.68 1.61 1.79 1.79 2.83 1.86(1.86) 100.9 130.9 97.3

DMS

TBM

atomic distance/bond length (Å) 2.34 2.35 2.38 2.37 2.23 2.23 1.68 1.68 1.61 1.61 1.80 1.79 1.79 1.79 2.85 2.87 1.83(1.83) 1.89 bond angle (deg) 100.7 100.8 129.2 130.8 95.6 97.4

H2S 2.31 2.34 2.27 1.69 1.61 1.79 1.81 2.91 1.36(1.35)/ 100.3 130.0 95.4

COS 2.34 2.35 2.21 1.69 1.61 1.79 1.80 2.40* 1.16** 100.4 130.1 95.1

H2O 2.32 2.36 2.36 1.68 1.61 1.79 1.80 2.29* 0.98(0.96)† 100.6 130.7 98.2

See the text and Figure 1 for notations.

adsorption energy was much higher than that of H2S and other sulfur compounds. Table 4 displays the structural details of the geometry optimized NaZ-adsorbate complexes. The binding energy of THT was greater than that of other compounds, and the order was THT (56.1) > H2O (53.1) > DMS (45.5) g TBM (45.3) > H2S (26.2) > COS (19.5). The Na+-S atomic distance (Å) in the complexes decreased following an order of H2S (2.91)

> TBM (2.87) > DMS (2.85) >THT (2.83), which was also exactly the reverse order to their binding energies. The formation of cluster-adsorbate complexes led to slight increases in the atomic distance between the O1, O2, and O3 atoms and the Na+ cation as well as the O2-Al-O3 bond angle, but the extents of these increases were smaller than those of the HZ cluster. The average sulfur-carbon (S-C) or sulfur-hydrogen (S-H) bond length (Å) of the adsorbates (THT ) 1.86, DMS ) 1.83,

Organosulfur Compounds on Zeolite Clusters

J. Phys. Chem. C, Vol. 112, No. 48, 2008 18959

TABLE 5: Relevant Atomic Distances, Bond Lengths, and Bond Angles of the Geometry Optimized AgZ-Adsorbate Complexesa THT O1-Ag+ O2-Ag+ O3-Ag+ Si-O1 Si-O2 Al-O2 Al-O3 Ag+-S or Ag+-O* S-C, S-H*, C-O**, or O-H† O1-Si-O2 Si-O2-Al O2-Al-O3 a

3.38 2.63 2.19 1.66 1.61 1.78 1.81 2.42 1.87(1.87) 106.9 131.1 98.3

DMS

TBM

atomic distance/bond length (Å) 3.48 3.09 2.60 2.58 2.18 2.21 1.67 1.66 1.61 1.62 1.78 1.78 1.81 1.81 2.42 2.43 1.83(1.83) 1.90 bond angle (deg) 105.9 104.6 132.3 130.6 98.1 97.7

H2S 3.58 2.67 2.16 1.68 1.61 1.78 1.81 2.43 1.38(1.35)* 105.7 132.7 99.6

COS 2.45 2.68 2.21 1.69 1.61 1.78 1.81 2.47* 1.17** 102.3 131.5 100.6

H2O 3.45 2.69 2.16 1.69 1.61 1.78 1.82 2.22* 1.00(0.96)† 105.2 133.1 100.7

See the text and Figure 1 for notations.

TBM ) 1.89, H2S ) 1.36) was also practically unchanged from those of the free molecules. Marked increases in the binding energies of the sulfur compounds were observed on the AgZ cluster. The order of binding energy of the compounds was THT (95.0) > DMS (88.9) > TBM (86.6) > H2S (55.5) > H2O (49.0) > COS (21.2). Notice that the binding energies of THT, DMS, TBM, and H2S on the AgZ were roughly two times higher than those on the HZ and NaZ clusters. Conversely, the binding energy of H2O on the AgZ was even lower than that on the HZ and NaZ clusters. Table 5 shows the structural properties of the cluster-adsorbate complexes for THT, DMS, TBM, H2S, COS, and H2O. The atomic distances (Å) between the Ag+ cation and the S atom (Ag+-S) were similar on these compounds (THT ) 2.42, DMS ) 2.42, TBM ) 2.43, H2S ) 2.43). The O2-Ag+ atomic distance decreased upon binding of the sulfur compounds, conversely to the results obtained on the HZ and NaZ clusters. The O1-Ag+ atomic distance increased, while the O3-Ag+ distance remained practically unchanged. In addition, the O1-Si-O2 bond angle increased slightly, whereas the O2-Al-O3 bond angle decreased. The results indicated that the Ag+ cation moved closer to the O2 atom upon binding of the sulfur compounds with preferential openings of the O1-Si-O2 bond angle. The average S-C or S-H bond length (Å) of the adsorbates (THT ) 1.87, DMS ) 1.83, TBM ) 1.90, H2S ) 1.38) was barely changed from that of the free molecules. 3.3. Adsorption of Hydrocarbons and CO2. Figure 3 compares the binding energies of C1-C3 hydrocarbons, CO2, and H2O on the zeolite clusters with those of the sulfur compounds. The structural details of their cluster complexes were not shown here, but they were treated with the identical methods and procedures used for the sulfur compounds. Here we only report their binding energies for comparison. The results showed significant enhancements in the binding energies of hydrocarbons, particularly unsaturated C2H4 (76.0 kJ mol-1) and C3H6 (82.2 kJ mol-1), on the AgZ cluster. The binding energies of the alkenes were much larger than those of the alkanes, and they increased with an increase in the carbon number on all the clusters. The binding energies of CO2 and COS did not show a clear dependency on the cation type and were similar to each other. The results clearly indicated that adsorption of H2O was considerably stronger than that of other compounds (except THT on the NaZ) on the HZ and NaZ clusters. On these clusters, the binding energies of THT, DMS, and TBM were higher than the hydrocarbons ([C3H6]; HZ ) 29.6, NaZ ) 27.0) and CO2 (HZ ) 22.3, NaZ ) 15.3), while the binding energy of H2S

Figure 3. Binding energies of the sulfur compounds, n-hydrocarbon (C1-C3), CO2, and H2O on the X(HO)3SiOAl(OH)3 (X ) H, Na, and Ag) zeolite model clusters.

was comparable to that of C3H6. The results on the AgZ cluster showed great increases in the binding energies of THT, DMS, TBM, H2S, and alkenes over those on the HZ and NaZ clusters. Differently, the binding energies of H2O (49.0 kJ mol-1) and CO2 (20.0 kJ mol-1) remained comparable to those on the HZ and NaZ clusters. The adsorption energies of H2S and COS were lower than those of C2H4 and C3H6. 4. Discussion Adsorption of chemical species on adsorbents is competitive in nature, and the adsorption selectivity is usually described with the selectivity factor (Se) that correlates the difference in adsorption energies of adsorbates by using an Arrhenius type expression such as Se ≈ exp[(E1 - E2)/RT]. The equation predicts a selectivity factor of ∼80 for ∆E of 10 kJ mol-1 at the standard temperature, and this clearly suggests that the relative magnitude of adsorption energies is a critical factor that determines the adsorption property of a material for specific compounds. In the real zeolite systems, the environments that adsorbates experience at the adsorption sites located in the pores may differ from those of our cluster model systems. However, many previous studies have shown that cluster models based ab initio calculations are effective in computational cost with reliable accuracy, and we note that the major attractive interactions for binding of the sulfur compounds, as will be discussed, arise from those between the cations of the clusters and the negatively charged S or O

18960 J. Phys. Chem. C, Vol. 112, No. 48, 2008 atoms of the compounds. Additional hydrogen bonding or van der Waals-type interactions may also exist, but they do not appear to change the order of binding energies among the compounds currently investigated. Later we discuss the cluster size dependency of the binding energy in detail with much larger X3O18Si3Al3 (X ) H, Na, Ag) ring clusters existing inside the supercage of the Y zeolite framework. As shown in Figure 1, the atomic distance between the O2 atom and the cation in the clusters increased with an order of Ag+ > Na+ > H+ reflecting an atomic size dependency. The interactions of the small H+ cation were localized to the O2 atom forming a close bonding to it, while the larger Na+ and Ag+ cations interacted with the adjacent O1, O2, and O3 atoms residing above the cluster framework. This led to slight elongations of the Si-O1 and Al-O3 bond length on the NaZ and AgZ. The minimum energy structures of the cluster-adsorbate complexes (Figure 2) disclosed that interactions between the cation and the electronegative S or O atom were the primary force for adsorption. Clearly, THT, DMS, TBM, and H2S formed the complexes with close proximity of the S atom to the cation, whereas COS binding occurred with neighboring the O atom with the cation of the clusters (the higher electronegativity of the O atom in COS led to the S atom becoming electron deficient, Table 1). It is reasonable to expect that the distance between the atoms of the primary charge interactions would decrease with an increase in the binding energy. The H+-S and Na+-S atomic distances showed exactly the reverse order (H2S > TBM > DMS > THT) to their binding energies (THT > DMS > TBM > H2S) on the HZ and NaZ clusters. Differently, the Ag+-S atomic distances were similar on the AgZ cluster complexes without showing such a correlation, and this mostly appeared due to the dissimilar nature of the soft Ag+ cation compared to the hard H+ and Na+ cations, as could be deduced from the spatial orientations of the adsorbates on the clusters (Figure 2). Considering the high negative charges on the O3 atom, the sulfur compounds were expected to bind giving preferential orientations to the O3 of the Al-O tetrahedral unit through additional interactions with the H atoms of the sulfur compounds. In addition, the primary cation-S interactions were expected to pull the cation from the counter O2 atom loosening the O2-cation atomic distance. These explanations agreed well with the results on the HZ and NaZ clusters. However, unlike the hard H+ and Na+ cations, the soft Ag+ cation on the AgZ cluster was pushed closer to the O2 and O1 atoms with an opening of the O1-Si-O2 bond angle upon accommodations of the sulfur compounds, and this led to the spatial orientations of the compounds to the Si-O tetrahedral unit. The binding energies of the sulfur compounds depend greatly on the nature of the cations. The calculation results compared in Figure 3 indicate that adsorption of the sulfur compounds on the H+ and Na+ cation types of zeolite clusters are limited in the presence of H2O. An intriguing feature of the Ag+ cation was that it markedly enhanced the binding energy of the sulfur compounds retaining the H2O binding energy comparable to those on HZ and NaZ. Notice that the binding energies of THT, DMS, TBM, and H2S on AgZ were about twice as larger as those on HZ and NaZ. This was due to strong pulling interaction between the Ag+ and the S atoms, and this was obvious when the binding energies of H2S and its oxygen analogue, H2O, were compared. The much higher binding energy of THT, DMS, TBM, and H2S (Ag+-S charge interactions) over H2O and COS (Ag+-O charge

Lee et al. interactions) reflected that the Ag+ cation induced much enhanced bonding interactions particularly with the sulfur atoms. The binding energy of unsaturated hydrocarbons, C2H4 and C3H6, was also greatly enhanced by the Ag+ cation indicating the strong binding interactions between the Ag+ cation and the π bonding electrons. The slightly lower binding energy of C3H6 compared to THT, DMS, and TBM, and the increase in the binding energy of alkenes with the carbon number reflected that the binding energy of higher unsaturated hydrocarbons could be larger than that of these sulfur compounds. The above results indicate that the suppression effects of H2O on adsorption of the sulfur compounds can be greatly reduced on the Ag+ cation-type zeolites. This agree well with the experimental work reported by Satokawa et al., where they claimed that Ag-Y zeolites removed sulfur compounds without significant inhibition by H2O presented in the hydrocarbon fuel gas.43 In the previous study, we reported that the order of THT desorption temperature from the H+, Na+, and Ag+ adsorption sites of AgNa-Y zeolites was Ag+ (560 K) > Na+ (470 K) > H+ (450 K).5 The results suggest that THT adsorption energy should follow the same order with these cations, and this agrees well with our calculation results: AgZ (95.0) > NaZ (56.1) > HZ (50.3). One also should notice that the order of binding energy of the sulfur compounds on the clusters was THT > DMS > TBM > H2S > COS, not depending on the cation type. This also nicely agrees with our previous experimental results that THT preferentially adsorbed over TBM with a nearly 100% selectivity on all the H+, Na+, and Ag+ cation types of Y zeolites. These good agreements between the experimental and calculation results indicate that the above sulfur compounds adsorb on the H+, Na+, and Ag+ adsorption sites via nondissociative molecular level interactions. The results disagree in parts with the work of Satokawa et al., where they reported dissociative adsorption of TBM on AgY zeolites via silver sulfides formation.15 The reason for this disagreement is not clear at this point, but our calculation results show that cleavage of the S-C bond of TBM on the AgZ cluster is not likely. The S-C bond length of TBM (1.90 Å) on the AgZ-TBM complex was similar to that of free TBM molecule (1.87 Å), and the Ag+-S atomic distance (2.49 Å) was considerably larger than the Ag-S bond length (2.38 Å) of Ag2S reference. In addition, such dissociative TBM adsorption would lead to a 100% adsorption selectivity of TBM over other molecularly adsorbing sulfur compounds. This contradicts the ∼100% selectivity of THT over TBM on the AgNa-Y zeolites experimentally observed in our previous study (nondissociative adsorption and desorption of THT was shown with the temperature programmed desorption (TPD) experiments).5 In this study we used a simple cluster model to mimic the zeolite structures. To clarify the cluster size effects on the adsorption energy, we further carried out the hybrid quantum mechanical/molecular mechanical (QM/MM) calculations, which are called two-layer ONIOM methods implemented in Gaussian, for a much larger faujasite zeolite structure. Figure 4 shows the entire QM/MM model zeolite system containing 213 atoms. The initial structures were adopted from the experimental X-ray structure of faujasite zeolite, and the H atoms and cations were added. The structures of quantum clusters regions consisting of X3O18Si3Al3 (X ) H, Na, Ag) were optimized by using the B3LYP/6-31+G** method, while the remaining large segments were treated with

Organosulfur Compounds on Zeolite Clusters

J. Phys. Chem. C, Vol. 112, No. 48, 2008 18961 Acknowledgment. H.C.W. acknowledges the support by the Korea Research Foundation Grant funded by the Korean Government (KRF-2005-042-00075). E.D.P. acknowledges the support by the Korea Research Foundation Grant funded by the Korean Government (MEST) (KRF-2007-412-J04001). References and Notes

Figure 4. Binding conformations of H2S and H2O on the Ag+ cation type QM/MM model faujasite zeolite system. Binding energies (BEs) for the H+, Na+, and Ag+ cation form zeolites are also shown.

universal force field [ONIOM(B3LYP/6-31+G**:UFF)]. At the QM/MM optimized geometries, the single-point energy calculations were further performed by using ONIOM(B3LYP/ 6-31+G**:B3LYP/STO-3G). For silver atoms on the quantum region, LanL2DZ basis sets and effective core potentials were used. We only considered the adsorption of two representative molecules, H2O and H2S, on the QM/MM zeolite model due to the computational costs. We found that the optimized QM/ MM structures of the zeolite backbone were very similar to those of our simple zeolite model systems (XZ). However, the positions of cations were quite distorted, because the accessible sites for the cations were limited toward the zeolite’s pore direction. Especially, the bond length between the Ag and O atoms was predicted to be 0.4 Å shorter than those in the AgZ cluster, reflecting that Ag atoms bind more tightly in real zeolites. Because the structures in the QM/ MM models were slightly different form the simple XZ cluster model, the binding energies obtained from both models were also different. However, the trends of relative binding energies which determine the selectivity for sulfur compound remained the same. The binding energies (in kJ mol-1) of H2S/H2O on the H-zeolite were 21.6/62.2 and that on the Na- and Ag-zeolite were 36.2/65.6 and 87.4/63.1, respectively. Clearly, these results showed that adsorption of the sulfur compounds depended on the kinds of cation even in the large cluster models and strongly bound to the Ag+ cation form zeolites. 5. Conclusions The effects of H+, Na+, and Ag+ cations of the zeolite model systems on the adsorption of common organosulfur compounds (THT, DMS, TBM, H2S, and COS), H2O, CO2, and C1-C3 hydrocarbons were investigated with ab initio calculation methods. It is shown that the Ag+ cation gives rise to marked enhancements in the adsorption energies of the sulfur compounds compared to the H+ and Na+ cations via strong Ag+-S binding interactions. Adsorption of the sulfur compounds occur via nondissociative molecular level interactions, and the order of adsorption energies is THT > DMS > TBM > H2S > COS on all the H+-, Na+-, and Ag+-zeolite model clusters, not depending on the cation form. Their binding energies depend on the cluster size, but the order of the energies and relative magnitudes remains the same showing a primary dependency on the cation type. These results show good agreements with the experimental adsorption properties of AgNa-Y zeolites for the sulfur compounds.

(1) Yang, R. T.; Herna´ndez-Maldonado, A. J.; Yang, F. H. Science 2003, 4, 301. (2) Velu, S.; Ma, X.; Song, C. Ind. Eng. Chem. Res. 2003, 42, 5293. (3) Herna´ndez-Maldonado, A. J.; Yang, R. T. J. Am. Chem. Soc. 2004, 126, 992. (4) Satokawa, S.; Kobayashi, Y.; Fujiki, H. Appl. Catal. B 2005, 56, 51. (5) Lee, D.; Ko, E.-Y.; Lee, H. C.; Kim, S.; Park, E. D. Appl. Catal. A 2008, 334, 129. (6) Whitehurst, D. D.; Isoda, T.; Mochida, I. AdV. Catal. 1998, 42, 345. (7) Ng, F. T. T.; Rahman, A.; Ohasi, T.; Jiang, M. Appl. Catal. B 2005, 56, 127. (8) Wakita, H.; Tachibana, Y.; Hosaka, M. Microporous Mesoporous Mater. 2001, 46, 237. (9) Bezverkhyy, I.; Bouguessa, K.; Geantet, C.; Vrinat, M. Appl. Catal. B 2006, 62, 299. (10) Weber, G.; Benoit, F.; Bellat, J.-P.; Paulin, C.; Mougin, P.; Thomas, M. Microporous Mesoporous Mater. 2008, 109, 184. (11) Herna´ndez-Maldonado, A. J.; Yang, F. H.; Qi, G.; Yang, R. T. Appl. Catal. B 2005, 56, 111. (12) Zhang, Z. Y.; Shi, T. B.; Jia, C. Z.; Ji, W. J.; Chen, Y.; He, M. Y. Appl. Catal. B 2008, 82, 1. (13) Herna´ndez-Maldonado, A. J.; Yang, R. T. AIChE J. 2004, 50, 791. (14) Xue, M.; Chitrakar, R.; Sakane, K.; Hirotsu, T.; Ooi, K.; Yoshimura, Y.; Feng, Q.; Sumida, N. J. Colloid Interface Sci. 2005, 285, 487. (15) Shimizu, K.; Kobayashi, N.; Satsuma, A.; Kojima, T.; Satokawa, S. J. Phys. Chem. B 2006, 110, 22570. (16) Salem, A. S. H. Ind. Eng. Chem. Res. 1994, 33, 336. (17) Weitkamp, J.; Schwark, M.; Ernest, S. J. Chem. Soc., Chem. Commun. 1991, 1133. (18) Garcia, C. L.; Lercher, J. A. J. Phys. Chem. 1991, 95, 10729. (19) Salem, A. S. H.; Hamid, H. S. Chem. Eng. Technol. 1997, 20, 342. (20) Haji, S.; Erkey, C. Ind. Eng. Chem. Res. 2003, 42, 6933. (21) Lee, S. H. D.; Kumar, R.; Krumpelt, M. Sep. Purif. Technol. 2002, 26, 247. (22) Larrubia, M. A.; Gutierrez-Alejandre, A.; Ramirez, J.; Busca, G. Appl. Catal. A 2002, 224, 167. (23) Ma, X.; Sun, L.; Song, C. Catal. Today 2002, 77, 107. (24) Jeevanandam, P.; Klabunde, K. J.; Tetzler, S. H. Microporous Mesoporous Mater. 2005, 79, 101. (25) de Wild, P. J.; Nyqvist, R. G.; de Bruijn, F. A.; Stobbe, E. R. J. Power Sources 2006, 159, 995. (26) Futami, H.; Hashizume, Y. Proceedings of the International Gas Research Conference, 1989; Gas Research Institute: Chicago, IL, 1990; p 1592. (27) Sosc´; un, H.; Castellano, O.; Herna´ndez, J.; Hinchliffe, A. Int. J. Quantum Chem. 2002, 87, 240. (28) Soscu´n, H.; Castellano, O.; Herna´ndez, J. J. Phys. Chem. B 2004, 108, 5620. (29) Lu¨, R.; Qiu, G.; Liu, C. J. Nat. Gas Chem. 2006, 15, 134. (30) Soscu´n, H.; Castellano, O.; Herna´ndez, J.; Arrieta, F.; Bermu´dez, Y.; Hinchliffe, A.; Brussin, M. R.; Sanchez, M.; Sierraalta, A.; Ruette, F. J. Mol. Catal. A: Chem. 2007, 278, 165. (31) Soscu´n, H.; O’Malley, P.; Hinchliffe, A. J. Mol. Struct. (THEOCHEM) 1995, 341, 237. (32) Tielens, F.; Geerlings, P. Chem. Phys. Lett. 2002, 354, 474. (33) Limtrakul, J.; Nanok, T.; Jungsuttiwong, S.; Khongpracha, P.; Truong, T. N. Chem. Phys. Lett. 2001, 349, 161. (34) Pantu, P.; Boekfa, B.; Limtrakul, J. J. Mol. Catal. A: Chem. 2007, 277, 171. (35) Limtrakul, J.; Treesukol, P.; Ebner, C.; Sansone, R.; Probst, M. Chem. Phys. 1997, 215, 77. (36) Plant, D. F.; Maurin, G.; Deroche, I.; Gaberova, L.; Llewellyn, P. L. Chem. Phys. Lett. 2006, 426, 387. (37) Dolg, M.; Stoll, H.; Preuss, H.; Pitzer, R. M. J. Phys. Chem. 1993, 97, 5852. (38) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785. (39) Becke, A. D. Phys. ReV. A 1988, 38, 3098. (40) Boys, S. F.; Bernardi, F. Mol. Phys. 1970, 19, 553. (41) (a) Kim, J.; Kim, K. S. J. Chem. Phys. 1998, 109, 5886. (b) Kim, J.; Lee, H. M.; Suh, S. B.; Majumdar, D.; Kim, K. S. J. Chem. Phys. 2000, 113, 5229.

18962 J. Phys. Chem. C, Vol. 112, No. 48, 2008 (42) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; 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.; Bakken, V.; 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,

Lee et al. 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; Gaussian, Inc.; Wallingford, CT, 2004. (43) Satokawa, S.; Kobayashi, Y. U.S. Patent 6,875,410, 2005.

JP804441Q