CHEMPHYSCHEM ARTICLES DOI: 10.1002/cphc.201300931
Conversion of CO2 and C2H6 to Propanoic Acid over a AuExchanged MCM-22 Zeolite Winyoo Sangthong,[a, b, c] Michael Probst,[e] and Jumras Limtrakul*[a, b, c, d] Finding novel catalysts for the direct conversion of CO2 to fuels and chemicals is a primary goal in energy and environmental research. In this work, density functional theory (DFT) is used to study possible reaction mechanisms for the conversion of CO2 and C2H6 to propanoic acid over a gold-exchanged MCM-22 zeolite catalyst. The reaction begins with the activation of ethane to produce a gold ethyl hydride intermediate. Hydrogen transfers to the framework oxygen leads then to gold ethyl adsorbed on the Brønsted-acid site. The energy barriers for these steps of ethane activation are 9.3 and 16.3 kcal mol1, respectively. Two mechanisms of propanoic acid formation are investigated. In the first one, the insertion of CO2 into the AuH bond of the first intermediate yields gold carboxyl
ethyl as subsequent intermediate. This is then converted to propanoic acid by forming the relevant CC bond. The activation energy of the rate-determining step of this pathway is 48.2 kcal mol1. In the second mechanism, CO2 interacts with gold ethyl adsorbed on the Brønsted-acid site. Propanoic acid is formed via protonation of CO2 by the Brønsted acid and the simultaneous formation of a bond between CO2 and the ethyl group. The activation energy there is 44.2 kcal mol1, favoring this second pathway at least at low temperatures. Gold-exchanged MCM-22 zeolite can therefore, at least in principle, be used as the catalyst for producing propanoic acid from CO2 and ethane.
1. Introduction The search for methods to prevent carbon dioxide (CO2) from entering the atmosphere and turn it into a useful functional group has attracted substantial interest from energy and engineering sciences, and might also be beneficial from an environmental perspective. Indeed, first industrial applications exist already.[1] Concurrent utilization of both CO2 and other unwanted chemicals to generate valuable compounds would be even more ideal.[2]
[a] Dr. W. Sangthong, Prof. Dr. J. Limtrakul Laboratory for Computational and Applied Chemistry Department of Chemistry, Faculty of Science and Center of Nanotechnology Kasetsart University Research and Development Institute Kasetsart University, Bangkok 10900 (Thailand) Fax: (+ 66) 2-562-5555 E-mail: [email protected] [b] Dr. W. Sangthong, Prof. Dr. J. Limtrakul NANOTEC Center for Nanoscale Materials for Green Nanotechnology Kasetsart University, Bangkok 10900 (Thailand) [c] Dr. W. Sangthong, Prof. Dr. J. Limtrakul Center for Advanced Studies in Nanotechnology and its Applications in Chemical, Food, and Agricultural Industries Kasetsart University, Bangkok 10900 (Thailand) [d] Prof. Dr. J. Limtrakul PTT Group Frontier Research Center, PTT Public Company Limited 555 Vibhavadi Rangsit Road, Chatuchak, Bangkok 10900 (Thailand) [e] Prof. Dr. M. Probst Institute of Ion Physics and Applied Physics University of Innsbruck, 6020 Innsbruck (Austria) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201300931.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The system discussed here explores such a type of reaction utilizing a class of nanomaterials that can capture CO2 from an exhaust stream and convert it into a useful and valuable product for industry. Carboxylic acids are ideally produced from the direct reaction of carbon dioxide with alkanes and propanoic acid is such obtained from the reaction between CO2 and ethane. Propanoic acid is widely used as an intermediate in the production of other chemicals, especially polymers. To achieve this task, ethane activation is also an important initial step to overcome thermodynamic limitations. Theoretical and experimental works on the catalytic activation of light alkanes by means of activation of the CH or CC bonds have been reported and the Lewis and Brønsted-acid sites of zeolites were found to serve as potential catalysts for these alkane activations.[3] The conversion of ethane and carbon dioxide investigated in this work consists of a double challenge: the cleavage of two strong bonds with dissociation energies of 101.1 and 192.4 kcal mol1 for CH and C=O, respectively. Metal and metal-supported catalysts such as Pd/carbon, Pt/alumina,[4] Pd/SiO2, Rh/SiO2,[5] CoCu,[6] and CoPd/TiO2[7] are widely used for the direct synthesis of acetic acid from carbon dioxide and methane. Many studies propose a stepwise mechanism for this reaction starting with the activation of methane. Metal-exchanged zeolites are also expected to be good candidates for alkane activation, as their ability to activate CH bonds of alkanes, especially ethane, has been consolidated both experimentally[8] and theoretically.[9] Recently, the utilization of Au-exchanged zeolites for methane activation[10] and to further the reaction of CO2 with such a prior-activated intermediate on a Au-exchanged ZSM-5 zeoChemPhysChem 2014, 15, 514 – 520
514
CHEMPHYSCHEM ARTICLES lite to produce acetic acid[11] has been demonstrated theoretically. However, this process requires high activation energy. Au-exchanged zeolites, including Au/NaY, Au/Na-ZSM-5, Au-ZSM-5, and Au/Na-MOR, have been successfully synthesized.[12] Moreover, it has been proven that Au + is the dominant active site for the adsorption of CO and H2O on zeoliteencapsulated gold catalysts.[12e, 13] MCM-22 zeolites are nanoporous materials with a peculiar structure that was firstly synthesized by Mobil researchers in 1990, and was originally used as a catalyst for hydrocarbon conversion reactions.[14] To the best of our knowledge, its use in the conversion of CO2 and ethane to propanoic acid has not yet been explored. In this study, we examine the feasibility using computational methods and demonstrate a new reaction mechanism. This could offer insights for experimentalists into the design of such a catalyst.
2. Results and Discussion The MCM-22 zeolite was represented by the 14T cluster model as illustrated in Figure 1. The lattice coordinates of the model were obtained from its crystallographic structures.[15] One aluminum atom is substituted for a silicon atom at the T1 position to generate the Brønsted-acid site. In the model of the Au + -exchanged MCM-22 (Au/MCM-22), the Brønsted site of the zeolite is exchanged by a Au cation. This position is considered to be the most stable configuration for this model.[16] Its stoichiometric composition is (SiO)13Al(AuO)H28. In this study, the 14T cluster of the AuI-MCM-22 zeolite acts as a catalyst for the conversion of carbon dioxide and ethane to propanoic acid. This model has been proved to be accurate for the activation of methane over a AuI-exchanged zeolite in one of our previous works.[10] In the optimized structure of the zeolite active site, the AuI cation is bound to two oxygen atoms close to the Al atom with distances of Au···Oa and Au···Ob of 2.30 and 2.31 , respectively. The Au···Al distance is 3.08 , which is consistent with the experimentally reported
www.chemphyschem.org value of 3.20 for the AuAl bond in the Au-exchanged FAU zeolite.[17] The charge of AuI obtained by natural bond orbital (NBO) analysis is + 0.70e. The conversion mechanism can proceed over two possible pathways. In both cases the reaction begins with the activation of ethane to produce a gold ethyl hydride intermediate (INT 1). After its formation, the hydrogen can transfer from the AuI cation to the regional framework oxygen leaving a gold ethyl complex adsorbed on the Brønsted-acid site (INT 2). Then, in the first pathway, carbon dioxide and the gold ethyl hydride intermediate (INT 1) react. This involves only the metal center of the catalyst. In the second pathway a possible bifunctionality of the catalyst is considered. The reaction occurs between carbon dioxide and the gold ethyl adsorbed on the Brønsted-acid site (INT 2). The
Table 1. Optimized geometrical parameters and calculated natural charge for the species involved in the ethane activation to produce the gold ethyl hydride intermediate (INT 1) and the gold ethyl adsorbed on the Brønsted-acid site (INT 2) over the 14T cluster model of AuI-MCM-22 calculated at the M06 L/6-31G(d,p) level of theory. Parameter distances [] AuC1 AuH1 C1H1 AuOa AuOb OaH1
Isolated cluster
ADS
– – 1.09 2.30 2.31 –
2.46 1.88 1.15 2.14 2.58 –
TS 1
2.16 1.58 1.61 2.25 2.35 –
INT 1
2.08 1.55 2.41 2.27 2.27 3.08
TS 2
2.06 1.55 2.55 3.09 2.21 2.12
INT 2
2.04 2.23 3.52 3.10 2.37 1.00
angles [8] aSiaOaAl aSibObAl aOaAlOb
126.0 125.9 95.4
124.9 126.4 95.4
124.8 125.6 93.2
126.4 123.4 93.2
124.1 126.5 97.1
126.7 123.8 94.6
charges [q] Au C1 H1
+ 0.70 -0.70 + 0.23
+ 0.62 0.71 + 0.18
+ 0.63 0.64 + 0.09
+ 0.64 0.55 + 0.03
+ 0.63 0.63 + 0.16
+ 0.19 0.67 + 0.54
details of each mechanism are reported in the following sections. Selected geometrical parameters of the complexes and their partial charges are documented in Tables 1–3. 2.1. CH Activation of Ethane
Figure 1. Structures of a) MCM-22 zeolite, b) 120T extended framework on which single-point calculations were performed, and c) 14T optimized reaction center.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
In a previous work,[11] we have shown that the adsorption of CO2 over AuI-ZSM-5 is less favorable than that of methane in the acetic acid production and that methane activation is needed in the first step of the reaction. Without catalyst, the direct conversion of methane ChemPhysChem 2014, 15, 514 – 520
515
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 2. Optimized geometrical parameters and calculated natural charge for the species involved in the reaction between the carbon dioxide and the gold ethyl hydride intermediate (INT 1) to produce propanoic acid over the 14T cluster model of AuI-MCM-22 calculated at the M06 L/6-31G(d,p) level of theory.
Table 3. Optimized geometrical parameters and calculated natural charge for the species involved in the reaction between the carbon dioxide and the gold ethyl adsorbed on the Brønsted-acid site (INT 2) to produce propanoic acid over the 14T cluster model of AuI-MCM-22 calculated at the M06 L/6-31G(d,p) level of theory.
Parameter
Parameter
distances [] AuC1 AuC2 AuH1 C1C2 C2O1 O1H1 AuOa AuOb
Isolated cluster
COADS 1
2.08 – 1.55 – 1.17 – 2.27 2.27
2.08 4.20 1.54 3.76 1.17 2.89 2.26 2.27
TS 3
2.07 2.67 1.93 3.17 1.24 1.14 2.27 2.50
INT 3
2.10 2.05 2.68 2.75 1.34 0.97 2.27 2.30
TS 4
2.28 2.21 2.78 1.79 1.35 0.97 2.18 2.48
PROD
3.35 3.01 – 1.50 1.32 0.97 2.83 2.14
distances [] AuC1 AuC2 OaH1 C1C2 C2O1 O1H1 AuOa AuOb
Isolated cluster
COADS 2
2.04 – 1.00 – – – 3.10 2.37
2.04 3.30 0.98 4.60 1.17 2.04 3.20 2.41
TS 5
2.22 2.23 1.50 1.94 1.30 1.04 4.13 2.59
PROD
3.35 3.01 – 1.50 1.32 0.97 2.83 2.14
angles [8] aSiaOaAl aSibObAl aOaAlOb
126.4 123.4 93.2
126.7 123.4 93.1
124.0 126.7 95.0
125.9 124.0 92.9
125.6 125.4 94.1
126.2 124.1 95.2
angles [8] aSiaOaAl aSibObAl aOaAlOb
126.7 123.8 94.6
127.2 122.8 96.2
125.6 122.8 97.9
126.2 124.1 95.2
charges [q] Au C1 H1 C2 O1
+ 0.64 0.55 + 0.03 + 1.03 0.51
+ 0.66 0.55 + 0.01 + 1.03 0.50
+ 0.40 0.67 + 0.41 + 1.05 0.61
+ 0.67 0.58 + 0.52 + 0.67 0.68
+ 0.66 0.67 + 0.51 + 0.69 0.69
+ 0.58 0.59 + 0.54 + 0.90 0.63
charges [q] Au C1 H1 C2 O1
+ 0.19 0.67 + 0.54 + 1.03 0.51
+ 0.12 0.67 + 0.58 + 1.06 0.55
+ 0.61 0.71 + 0.57 + 0.78 0.71
+ 0.58 0.59 + 0.54 + 0.90 0.63
and carbon dioxide to carboxylic acid is kinetically very unfavtransition state (TS 1), the ethane C1H1 bond is cleaved by orable with an Ea of 83.8 kcal mol1, as calculated at the M06 L/ the Au cation and hydride is transferred from the C1 atom to 6-31G(d,p) level, or 87.6 kcal mol1 if calculated at the MP2/6the metal center, leading to distances of 1.61 and 1.58 for 31G(d,p) level of theory. Therefore, a catalyst is necessary for C1···H and AuI···H, respectively. The distance between C1 and this conversion. In the reaction under consideration here, ethane takes the place of methane, starting with its physisorption on the active site. The energetic profile and geometries of relevant complexes are given in Figure 2. Selected geometrical parameters of the structures obtained along the reaction pathway and the natural population analysis (NPA) charges are tabulated in Table 1. We found that the h2 configuration (denoted ADS) is the most stable one. The C1 atom of ethane is 2.46 away from the AuI cation and the distance between AuI and the nearest H atom is 1.88 . The charge of Au becomes less positive (+0.62e) as compared to the bare zeolite. The C1H1 bond is lengthened from 1.09 to 1.15 . The adsorption energy is Figure 2. Energy profile of the Au-MCM-22-catalyzed reaction from ethane to the ethyl hydride (INT 1) and the 19.2 kcal mol1. At the first gold ethyl (INT 2) intermediate. Energies in kcal mol1. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 514 – 520
516
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Au decreases from 2.46 to 2.16 . The activation energy of this step is predicted to be 9.3 kcal mol1. This is lower than required for the activation of methane on a Au-MCM-22 catalyst (12 kcal mol1).[10] We note that this order is opposite to the strength of the CH bonds of these alkanes. The CH2CH3 group and the H atom stay coordinated to the AuI cation. The relative energy of INT 1 is 18.1 kcal mol1 with respect to the isolated systems. After that, the hydrogen atom can transfer from AuI to the oxygen atom of the zeolite framework while the CH2CH3 group remains bound to the metal. At the second transition state (TS 2), the Au···H1 and Oa···H1 distances become 1.55 and 2.12 and the AuC1 bond length is 2.06 . This step requires an activation energy of 16.3 kcal mol1. The charge of H1 increases slightly from + 0.03e in the INT 1 to + 0.16e in TS 2. Finally the second intermediate (INT 2) is formed. Gold ethyl, AuCH2CH3, is strongly adsorbed on the Brønsted-acid site with a relative energy of 21.9 kcal mol1 (INT 2). We checked the sufficiency of the size of the 5T region being relaxed by extending it to all atoms of the model except of the terminating atoms. We found that the results for ethane activation to be quite similar to the one obtained with the smaller 5T cluster. The values are included in the Supporting Information. Therefore, we feel confident that the practical relaxation scheme in this work is sufficiently accurate. The Wiberg bond order from the NBO analysis describes the bond formations and the bond breakings of the complexes along the reaction coordinate. In the Supporting Information the bond orders of several bonds are plotted against the reaction coordinate. From ADS to INT 1, the bond order of C1H1 decreases while the AuC1 and AuH1 bond orders increase to about 0.5, indicating the breaking of the C1H1 bond and the formation of the two new bonds to the AuI cation. From INT 1 to INT 2, the bond order of C1H1 decreases while the OaH1 bond orders increase from about zero to 0.57, indicating the generation of the Brønsted-acid site in the zeolite framework. Comparing the two steps of ethane activation over AuI-MCM-22, the apparent activation energies of both steps are below zero, meaning reactions should also occur easily. The stability of the adsorption complex (ADS) and the intermediates (INT 1 and INT 2) are comparable.
ethane over AuI-MCM-22 zeolite requires moderate activation energies and that the intermediates are thermodynamically stable compared to the adsorption complex. In this part, we examine the insertion of CO2 into these intermediates. Two pathways are possible: 1) The reaction between carbon dioxide and the gold ethyl hydride intermediate (INT 1) and 2) the reaction between carbon dioxide and gold ethyl adsorbed on the Brønsted-acid site (INT 2). In the first pathway, AuI acts as a monofunctional catalyst. Optimized geometries and the energetic diagram are shown in Figure 3. Selected geometrical parameters of the structures obtained along the reaction pathway and the NPA charges are given in Table 2. Firstly, the carbon dioxide adsorbs over the gold ethyl hydride intermediate (INT 1). The relative energy of this coadsorption complex, COADS, is 24.7 kcal mol1. In principle, the carbon dioxide molecule can be inserted into either the AuC1 or the AuH1 bonds of this intermediate. The first reaction produces a HAuOOCCH2CH3 unit which is bulkier than the HOOCAuCH2CH3 unit resulting from the second reaction. After the insertion into the AuH1 bond a gold carboxyl ethyl intermediate (INT 3) is formed. At the transition state, TS 3, the C2O1 bond distance of the carbon dioxide is increased from 1.17 to 1.24 , while the AuC2 and the O1H1 bond are forming with bond distances of 2.67 and 1.14 , respectively. The AuH1 bond is lengthened to 1.93 . This step requires an activation energy of 48.2 kcal mol1. The changes in the NPA charges from the coadsorption complex (COADS 1) to the transition state (TS 3) show that the charge of H1 increases from + 0.01 to + 0.41 e while the one of Au decreases from + 0.66 to + 0.40 e, showing a transfer of electron density from H1 to Au. The reaction proceeds with the formation of a gold
2.2. The Formation of Propanoic Acid In the previous section it was shown that the activation of
Figure 3. Energy profile of the Au-MCM-22-catalyzed reaction of carbon dioxide and the ethyl hydride intermediate (INT 1) to propanoic acid. Energies are in kcal mol1.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 514 – 520
517
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
carboxyl ethyl intermediate (INT 3) with a relative energy of 9.9 kcal mol1 with respect to the isolated system. There, both the C(O)OH and the CH2CH3 group bind to Au with an Au C1 distance of 2.10 and an AuC2 distance of 2.05 . The distance between the carbon atoms of the two groups is 2.75 . The Wiberg bond order demonstrates that the C2=O1 double bond of CO2 changes into a single bond. The AuC2 and the O1H1 bond form during this step (see the Supporting Information). The next step is the formation of propanoic acid by the connection of carboxyl C1 and ethyl C2. At the reaction between adsorbed carbon dioxide and gold ethyl (INT 2) to propanoic transition state, TS 4, the C1C2 Figure 4. Energy profile of the acid. Energies are in kcal mol1. distance becomes 1.79 while the bond distances of AuC1 and AuC2 increase from 2.10 and 2.05 to 2.28 and 2.21 , cess of insertion of a carbon dioxide molecule, forward rate respectively. This bond formation needs an activation energy constants, k, have been calculated for both pathways. The of 24.2 kcal mol1. The NPA charges change only slightly from DG ¼6 values from the 14T cluster model were calculated from the intermediate (INT 3) to the transition state (TS 4) except the energy differences corrected with the corresponding frefor the one of the C1 atom, which obtains electron density quency contributions. They were then used to derive the rate from Au and decreases from 0.58 to 0.67 e. The propanoic constants in the temperature range from 298.15 to 798.15 K acid is bound to the Au atom through the lone pair electrons using the Equation (1): of its carbonyl group. The AuO distance is 2.13 . The relative energy of this complex is 24.3 kcal mol1. The release of the kB T DG# ð1Þ kðTÞ ¼ exp ð Þ propanoic acid requires then a desorption energy of 34.9 kcal h RT mol1. The reaction profile of the second possible pathway, the reDG ¼6 is the difference between the free energy of the transiaction between carbon dioxide and gold ethyl on the Brønsttion state and the adsorption or intermediate complex, kB is ed-acid site (INT 2), is shown in Figure 4. Selected geometrical the Boltzmann’s constant, h is the Plank’s constant, T is the parameters of the structures obtained along the reaction pathtemperature in Kelvin, and R is the universal gas constant. The way and the NPA charges are given in Table 3. This pathway values of k for the elementary steps of the carbon dioxide inbegins with the adsorption of carbon dioxide on gold ethyl adsertion into the intermediates over AuI-MCM-22 at different sorbed on the Brønsted acid (COADS 2). This complex stabilizes the system by 5.0 kcal mol1. The reaction takes place via a bifunctional catalysis, including the protonation of the Table 4. Reaction rate constants (k) for the propanoic acid formation carbon dioxide by the Brønsted acid and the concurrent forstep from the monofunctional gold catalysis (kTS 3) and the concerted bifunctional mechanism (kTS 5) mation of the CC bond between the carbon dioxide and the gold ethyl group. At the transition state, TS 5, the distances Temperature [K] Rate constants OaH1 and O1H1 are 1.50 and 1.04 , respectively, while the kTS 5 kTS 5/kTS 3 kTS 3 distance between the carbon dioxide carbon atom (C2) and 23 21 298.15 2.80 10 1.75 10 62.6 the carbon atom of the ethyl group (C1) is 1.94 . The changes 3.84 1017 34.7 348.15 1.11 1018 6.68 1014 21.9 398.15 3.05 1015 in the NPA charges from the coadsorption complex (COADS 2) 2.16 1011 15.2 448.15 1.42 1012 to the transition state (TS 5) shows electron density transfers 2.16 109 11.3 498.15 1.91 1010 from Au (+ 0.12 e! + 0.61 e) to C2 (+ 1.06 e! + 0.78 e). This 8 8 9.25 10 8.89 548.15 1.04 10 step requires an activation energy of 44.2 kcal mol1, which is 2.09 106 7.27 598.15 2.88 107 2.91 105 6.15 648.15 4.73 106 lower than that determined for the first pathway (48.2 kcal 2.77 104 5.36 698.15 5.16 105 mol1). Thus the second pathway appears to be favored. 1.93 103 4.78 748.15 4.06 104 The rate constants for the reaction as a function of tempera798.15 2.44 103 1.06 102 4.34 ture can be examined by transition state theory. For the pro 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 514 – 520
518
CHEMPHYSCHEM ARTICLES temperatures are tabulated in Table 4. One sees that the path over Au acting as a bifunctional catalyst dominates at low temperatures, whereas at high temperatures the rate constants of the two pathways are comparable. The activation enthalpies and the free energies for the step of propanoic acid formation by the monofunctional (TS 3) and by the concerted bifunctional (TS 5) gold catalysis at different temperatures are available in the Supporting Information. One of the key issues for understanding adsorption phenomena and catalytic activities in a zeolitic system is the role of the confining van der Waals interactions in the framework. The extended 120T cluster of the supercage of MCM-22 zeolite was used for single-point calculations to study the effects of the framework. The relative energies of the systems involved in the reaction are summarized in Table 5. It was found that all complexes are stabilized by the zeolite framework. This results in slightly higher adsorption energies for both the reactant and the product. In contrast, the confinement effect lowers the activation energy of each system in the MCM-22 framework, except for that of TS 1. This could be explained by long-range interactions which play a role in the adsorption of ethane and propanoic acid and also influence the ion-pair-like transition state.
3. Conclusions The direct conversion of carbon dioxide and ethane to propanoic acid over gold-exchanged MCM-22 zeolite has been investigated by DFT calculations with the M06 L functional. The reaction begins with the activation of ethane to produce a gold ethyl hydride intermediate and then hydrogen transfers to the framework oxygen to produce gold ethyl adsorbed on the Brønsted-acid site. The energy barriers for these two steps of ethane activation are 9.3 and 16.3 kcal mol1, respectively. Two mechanisms of propanoic acid formation were investigated. In the first one, the insertion of CO2 into the AuH bond of the first intermediate yields a gold carboxyl ethyl intermediate. Subsequent conversion of this intermediate to propanoic acid occurs through the formation of a CC bond. The activation energy of the rate-determining step of this pathway is 48.2 kcal mol1. In the second mechanism investigated, CO2 interacts with gold ethyl adsorbed on the Brønsted-acid site. Propanoic acid is formed by the protonation of CO2 by the Brønsted acid and the concurrent bond formation between CO2 and the ethyl group. The activation energy of this reaction is 44.2 kcal mol1. From these results, the Au-MCM-22 zeolite should exhibit good performance for ethane activation. The reaction of CO2 with the gold ethyl and the Brønsted-acid site is found to be the rate-limiting step for the formation of propanoic acid. The high activation energy means that a sufficiently high temperature is needed in order to enable the catalytic CO2 conversion with a reasonable rate.
Computational Method In zeolite materials, the van der Waal interactions are more pronounced than in conventional heterogeneous catalysts. As a result, 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org dispersion-corrected functionals which overcome these are needed for the investigation of catalytic processes inside the zeolite pores. In the M06 family this kind of interaction is described efficiently.[18] Therefore, the M06-L functional, which is recommended for investigation of the confined interactions of adsorbates with zeolites involving transition metals, was employed in this work. The 631G(d,p) basis set was employed for all light atoms, while Au was described by the LANL2DZ effective core potential and the corresponding basis set. During geometry optimizations, only the 5T active region and the adsorbates were allowed to relax while the rest of the structure was fixed at the crystallographic coordinates. Transition states were located with the “Berny Algorithm”[19] and were checked to ensure that they have only one imaginary frequency in which the displacement vectors of all transition state structures correspond to the bond-forming and bond-breaking processes for the reaction steps. All systems were considered to be in the singlet state, which is the ground state of the Au ion. Single-point calculations were applied on a larger 120T cluster, which includes a supercage of the MCM-22 zeolite in order to cover the confinement effect from the zeolite framework. The NBO analysis was used to obtain charge distributions and bond orders of the systems.[20] All calculations were performed with the Gaussian 09 code.[21]
Table 5. Energies of the components in the conversion of CO2 and C2H6 to propanoic acid process over the Au-exchanged MCM-22 zeolite, given relative to the energy of the reactants.
Systems
ADS TS 1 INT 1 TS 2 INT 2 COADS 1 TS 3 INT 3 TS 4 COADS 2 TS 5 PROD
Relative energy [kcal mol1] 14T M06-L/6-31G(d,p) (optimized)
120T M06-L/6-31G(d,p) (single point)
19.2 9.9 18.1 1.8 21.9 24.7 23.5 9.9 14.3 26.9 17.3 24.3
21.2 11.4 (Ea.1 = 9.8) 18.6 4.9 (Ea.2 = 13.7) 22.1 24.8 20.0 (Ea.3 = 44.8) 13.5 12.0 (Ea.4 = 25.5) 27.0 13.7 (Ea.5 = 40.7) 28.6
(Ea.1 = 9.3) (Ea.2 = 16.3)
(Ea.3 = 48.2) (Ea.4 = 24.2) (Ea.5 = 44.2)
Acknowledgements This work was supported in part by grants from the National Science and Technology Development Agency (NSTDA Chair Professor funded by the Crown Property Bureau under the management of the National Science and Technology Development Agency and NANOTEC Center of Excellence funded by the National Nanotechnology Center), the Kasetsart University Research and Development Institute (KURDI), and the Commission on Higher Education, Ministry of Education (the “National Research University Project of Thailand (NRU)” and the “National Center of Excellence for Petroleum, Petrochemical and Advanced Materials (NCE-PPAM)”). ChemPhysChem 2014, 15, 514 – 520
519
CHEMPHYSCHEM ARTICLES Keywords: carbon dioxide conversion · density functional theory calculations · ethane · propanoic acid · zeolites [1] G. A. Olah, Angew. Chem. 2013, 125, 112 – 116; Angew. Chem. Int. Ed. 2013, 52, 104 – 107. [2] a) B. Eliasson, C.-j. Liu, U. Kogelschatz, Ind. Eng. Chem. Res. 2000, 39, 1221 – 1227; b) V. Havran, M. P. Dudukovic, C. S. Lo, Ind. Eng. Chem. Res. 2011, 50, 7089 – 7100; c) Y. Li, C.-J. Liu, B. Eliasson, Y. Wang, Energy Fuels 2002, 16, 864 – 870; d) C.-J. Liu, B. Xue, B. Eliasson, F. He, Y. Li, G.-H. Xu, Plasma Chem. Plasma Process. 2001, 21, 301 – 310; e) D. Shi, Y. Feng, S. Zhong, Catal. Today 2004, 98, 505 – 509; f) J.-g. Wang, C.-j. Liu, Y.-p. Zhang, B. Eliasson, Chem. Phys. Lett. 2003, 368, 313 – 318. [3] a) D. C. Tranca, N. Hansen, J. A. Swisher, B. Smit, F. J. Keil, J. Phys. Chem. C 2012, 116, 23408 – 23417; b) P. M. Zimmerman, D. C. Tranca, J. Gomes, D. S. Lambrecht, M. Head-Gordon, A. T. Bell, J. Am. Chem. Soc. 2012, 134, 19468 – 19476. [4] E. M. Wilcox, G. W. Roberts, J. J. Spivey, Catal. Today 2003, 88, 83 – 90. [5] Y.-H. Ding, W. Huang, Y.-G. Wang, Fuel Process. Technol. 2007, 88, 319 – 324. [6] W. Huang, K. C. Xie, J. P. Wang, Z. H. Gao, L. H. Yin, Q. M. Zhu, J. Catal. 2001, 201, 100 – 104. [7] W. Huang, W. Z. Sun, F. Li, AIChE J. 2010, 56, 1279 – 1284. [8] a) V. B. Kazansky, I. R. Subbotina, N. Rane, S. R. A. van, E. J. M. Hensen, Phys. Chem. Chem. Phys. 2005, 7, 3088 – 3092; b) G. Caeiro, R. H. Carvalho, X. Wang, M. A. N. D. A. Lemos, F. Lemos, M. Guisnet, F. R. Ribeiro, J. Mol. Catal. A 2006, 255, 131 – 158; c) A. G. Stepanov, V. N. Parmon, D. Freude, Kinet. Catal. 2007, 48, 521 – 534; d) A. Itadani, H. Torigoe, T. Yumura, T. Ohkubo, H. Kobayashi, Y. Kuroda, J. Phys. Chem. C 2012, 116, 10680 – 10691; e) Y. G. Kolyagin, I. I. Ivanova, Y. A. Pirogov, Solid State Nucl. Magn. Reson. 2009, 35, 104 – 112. [9] a) E. A. Pidko, V. B. Kazansky, E. J. M. Hensen, S. R. A. Van, J. Catal. 2006, 240, 73 – 84; b) E. A. Pidko, S. R. A. van, J. Phys. Chem. C 2007, 111, 2643 – 2655; c) H. A. Aleksandrov, G. N. Vayssilov, Catal. Today 2010, 152, 78 – 87; d) O. A. Anunziata, J. Cussa, A. R. Beltramone, Catal. Today 2011, 171, 36 – 42. [10] S. Wannakao, C. Warakulwit, K. Kongpatpanich, M. Probst, J. Limtrakul, ACS Catal. 2012, 2, 986 – 992. [11] W. Panjan, J. Sirijaraensre, C. Warakulwit, P. Pantu, J. Limtrakul, Phys. Chem. Chem. Phys. 2012, 14, 16588 – 16594. [12] a) J. C. Fierro-Gonzalez, B. C. Gates, J. Phys. Chem. B 2004, 108, 16999 – 17002; b) Z.-X. Gao, Q. Sun, H.-Y. Chen, X. Wang, W. M. H. Sachtler, Catal. Lett. 2001, 72, 1 – 5; c) M. M. Mohamed, T. M. Salama, M. Ichikawa, J. Colloid Interface Sci. 2000, 224, 366 – 371; d) M. M. Mohamed, T. M. Salama, R. Ohnishi, M. Ichikawa, Langmuir 2001, 17, 5678 – 5684; e) S. Qiu, R. Ohnishi, M. Ichikawa, J. Phys. Chem. 1994, 98, 2719 – 2721; f) S. Qiu, R. Ohnishi, M. Ichikawa, J. Chem. Soc. Chem. Commun. 1992, 19, 1425 – 1427;
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
[13] [14]
[15] [16] [17] [18]
[19] [20] [21]
g) T. M. Salama, T. Shido, R. Ohnishi, M. Ichikawa, J. Phys. Chem. 1996, 100, 3688 – 3694. M. M. Mohamed, M. Ichikawa, J. Colloid Interface Sci. 2000, 232, 381 – 388. a) M. Bjorgen, S. Akyalcin, U. Olsbye, S. Benard, S. Kolboe, S. Svelle, J. Catal. 2010, 275, 170 – 180; b) B. Onida, F. Geobaldo, F. Testa, F. Crea, E. Garrone, Stud. Surf. Sci. Catal. 2000, 130, 2951 – 2956; c) X. Wang, W. Dai, G. Wu, L. Li, N. Guan, M. Hunger, Microporous Mesoporous Mater. 2012, 151, 99 – 106. M. E. Leonowicz, J. A. Lawton, S. L. Lawton, M. K. Rubin, Science 1994, 264, 1910 – 1913. D. Zhou, Y. Bao, M. Yang, N. He, G. Yang, J. Mol. Catal. A 2006, 244, 11 – 19. J. Lu, C. Aydin, N. D. Browning, B. C. Gates, Angew. Chem. 2012, 124, 5944-5948; Angew. Chem. Int. Ed. 2012, 51, 5842 – 5846. a) B. Boekfa, S. Choomwattana, P. Khongpracha, J. Limtrakul, Langmuir 2009, 25, 12990 – 12999; b) Y. Zhao, D. G. Truhlar, J. Phys. Chem. C 2008, 112, 6860 – 6868; c) Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215 – 241; d) Y. Zhao, D. G. Truhlar, Rev. Mineral. Geochem. 2010, 71, 19 – 37; e) M. Mantina, R. Valero, D. G. Truhlar, J. Chem. Phys. 2009, 131, 064706/1 – 5; f) C. Kumsapaya, K. Bobuatong, P. Khongpracha, Y. Tantirungrotechai, J. Limtrakul, J. Phys. Chem. C 2009, 113, 16128 – 16137; g) J. Sirijaraensre, J. Limtrakul, Phys. Chem. Chem. Phys. 2013, 15, 18093 – 18100; h) T. Maihom, P. Khongpracha, J. Sirijaraensre, J. Limtrakul, ChemPhysChem 2013, 14, 101 – 107. a) C. Gonzalez, H. B. Schlegel, J. Chem. Phys. 1989, 90, 2154 – 2161; b) H. B. Schlegel, J. Comput. Chem. 1982, 3, 214 – 218. a) A. E. Reed, F. Weinhold, J. Chem. Phys. 1983, 78, 4066 – 4073; b) A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899 – 926. Gaussian 09 (Revision C.01), M. J. T. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels,. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian, Inc., Wallingford CT, 2010.
Received: October 11, 2013 Revised: November 25, 2013 Published online on December 20, 2013
ChemPhysChem 2014, 15, 514 – 520
520
Conversion of CO2 and C2H6 to Propanoic Acid over a AuExchanged MCM-22 Zeolite Winyoo Sangthong,[a, b, c] Michael Probst,[e] and Jumras Limtrakul*[a, b, c, d] Finding novel catalysts for the direct conversion of CO2 to fuels and chemicals is a primary goal in energy and environmental research. In this work, density functional theory (DFT) is used to study possible reaction mechanisms for the conversion of CO2 and C2H6 to propanoic acid over a gold-exchanged MCM-22 zeolite catalyst. The reaction begins with the activation of ethane to produce a gold ethyl hydride intermediate. Hydrogen transfers to the framework oxygen leads then to gold ethyl adsorbed on the Brønsted-acid site. The energy barriers for these steps of ethane activation are 9.3 and 16.3 kcal mol1, respectively. Two mechanisms of propanoic acid formation are investigated. In the first one, the insertion of CO2 into the AuH bond of the first intermediate yields gold carboxyl
ethyl as subsequent intermediate. This is then converted to propanoic acid by forming the relevant CC bond. The activation energy of the rate-determining step of this pathway is 48.2 kcal mol1. In the second mechanism, CO2 interacts with gold ethyl adsorbed on the Brønsted-acid site. Propanoic acid is formed via protonation of CO2 by the Brønsted acid and the simultaneous formation of a bond between CO2 and the ethyl group. The activation energy there is 44.2 kcal mol1, favoring this second pathway at least at low temperatures. Gold-exchanged MCM-22 zeolite can therefore, at least in principle, be used as the catalyst for producing propanoic acid from CO2 and ethane.
1. Introduction The search for methods to prevent carbon dioxide (CO2) from entering the atmosphere and turn it into a useful functional group has attracted substantial interest from energy and engineering sciences, and might also be beneficial from an environmental perspective. Indeed, first industrial applications exist already.[1] Concurrent utilization of both CO2 and other unwanted chemicals to generate valuable compounds would be even more ideal.[2]
[a] Dr. W. Sangthong, Prof. Dr. J. Limtrakul Laboratory for Computational and Applied Chemistry Department of Chemistry, Faculty of Science and Center of Nanotechnology Kasetsart University Research and Development Institute Kasetsart University, Bangkok 10900 (Thailand) Fax: (+ 66) 2-562-5555 E-mail: [email protected] [b] Dr. W. Sangthong, Prof. Dr. J. Limtrakul NANOTEC Center for Nanoscale Materials for Green Nanotechnology Kasetsart University, Bangkok 10900 (Thailand) [c] Dr. W. Sangthong, Prof. Dr. J. Limtrakul Center for Advanced Studies in Nanotechnology and its Applications in Chemical, Food, and Agricultural Industries Kasetsart University, Bangkok 10900 (Thailand) [d] Prof. Dr. J. Limtrakul PTT Group Frontier Research Center, PTT Public Company Limited 555 Vibhavadi Rangsit Road, Chatuchak, Bangkok 10900 (Thailand) [e] Prof. Dr. M. Probst Institute of Ion Physics and Applied Physics University of Innsbruck, 6020 Innsbruck (Austria) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201300931.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The system discussed here explores such a type of reaction utilizing a class of nanomaterials that can capture CO2 from an exhaust stream and convert it into a useful and valuable product for industry. Carboxylic acids are ideally produced from the direct reaction of carbon dioxide with alkanes and propanoic acid is such obtained from the reaction between CO2 and ethane. Propanoic acid is widely used as an intermediate in the production of other chemicals, especially polymers. To achieve this task, ethane activation is also an important initial step to overcome thermodynamic limitations. Theoretical and experimental works on the catalytic activation of light alkanes by means of activation of the CH or CC bonds have been reported and the Lewis and Brønsted-acid sites of zeolites were found to serve as potential catalysts for these alkane activations.[3] The conversion of ethane and carbon dioxide investigated in this work consists of a double challenge: the cleavage of two strong bonds with dissociation energies of 101.1 and 192.4 kcal mol1 for CH and C=O, respectively. Metal and metal-supported catalysts such as Pd/carbon, Pt/alumina,[4] Pd/SiO2, Rh/SiO2,[5] CoCu,[6] and CoPd/TiO2[7] are widely used for the direct synthesis of acetic acid from carbon dioxide and methane. Many studies propose a stepwise mechanism for this reaction starting with the activation of methane. Metal-exchanged zeolites are also expected to be good candidates for alkane activation, as their ability to activate CH bonds of alkanes, especially ethane, has been consolidated both experimentally[8] and theoretically.[9] Recently, the utilization of Au-exchanged zeolites for methane activation[10] and to further the reaction of CO2 with such a prior-activated intermediate on a Au-exchanged ZSM-5 zeoChemPhysChem 2014, 15, 514 – 520
514
CHEMPHYSCHEM ARTICLES lite to produce acetic acid[11] has been demonstrated theoretically. However, this process requires high activation energy. Au-exchanged zeolites, including Au/NaY, Au/Na-ZSM-5, Au-ZSM-5, and Au/Na-MOR, have been successfully synthesized.[12] Moreover, it has been proven that Au + is the dominant active site for the adsorption of CO and H2O on zeoliteencapsulated gold catalysts.[12e, 13] MCM-22 zeolites are nanoporous materials with a peculiar structure that was firstly synthesized by Mobil researchers in 1990, and was originally used as a catalyst for hydrocarbon conversion reactions.[14] To the best of our knowledge, its use in the conversion of CO2 and ethane to propanoic acid has not yet been explored. In this study, we examine the feasibility using computational methods and demonstrate a new reaction mechanism. This could offer insights for experimentalists into the design of such a catalyst.
2. Results and Discussion The MCM-22 zeolite was represented by the 14T cluster model as illustrated in Figure 1. The lattice coordinates of the model were obtained from its crystallographic structures.[15] One aluminum atom is substituted for a silicon atom at the T1 position to generate the Brønsted-acid site. In the model of the Au + -exchanged MCM-22 (Au/MCM-22), the Brønsted site of the zeolite is exchanged by a Au cation. This position is considered to be the most stable configuration for this model.[16] Its stoichiometric composition is (SiO)13Al(AuO)H28. In this study, the 14T cluster of the AuI-MCM-22 zeolite acts as a catalyst for the conversion of carbon dioxide and ethane to propanoic acid. This model has been proved to be accurate for the activation of methane over a AuI-exchanged zeolite in one of our previous works.[10] In the optimized structure of the zeolite active site, the AuI cation is bound to two oxygen atoms close to the Al atom with distances of Au···Oa and Au···Ob of 2.30 and 2.31 , respectively. The Au···Al distance is 3.08 , which is consistent with the experimentally reported
www.chemphyschem.org value of 3.20 for the AuAl bond in the Au-exchanged FAU zeolite.[17] The charge of AuI obtained by natural bond orbital (NBO) analysis is + 0.70e. The conversion mechanism can proceed over two possible pathways. In both cases the reaction begins with the activation of ethane to produce a gold ethyl hydride intermediate (INT 1). After its formation, the hydrogen can transfer from the AuI cation to the regional framework oxygen leaving a gold ethyl complex adsorbed on the Brønsted-acid site (INT 2). Then, in the first pathway, carbon dioxide and the gold ethyl hydride intermediate (INT 1) react. This involves only the metal center of the catalyst. In the second pathway a possible bifunctionality of the catalyst is considered. The reaction occurs between carbon dioxide and the gold ethyl adsorbed on the Brønsted-acid site (INT 2). The
Table 1. Optimized geometrical parameters and calculated natural charge for the species involved in the ethane activation to produce the gold ethyl hydride intermediate (INT 1) and the gold ethyl adsorbed on the Brønsted-acid site (INT 2) over the 14T cluster model of AuI-MCM-22 calculated at the M06 L/6-31G(d,p) level of theory. Parameter distances [] AuC1 AuH1 C1H1 AuOa AuOb OaH1
Isolated cluster
ADS
– – 1.09 2.30 2.31 –
2.46 1.88 1.15 2.14 2.58 –
TS 1
2.16 1.58 1.61 2.25 2.35 –
INT 1
2.08 1.55 2.41 2.27 2.27 3.08
TS 2
2.06 1.55 2.55 3.09 2.21 2.12
INT 2
2.04 2.23 3.52 3.10 2.37 1.00
angles [8] aSiaOaAl aSibObAl aOaAlOb
126.0 125.9 95.4
124.9 126.4 95.4
124.8 125.6 93.2
126.4 123.4 93.2
124.1 126.5 97.1
126.7 123.8 94.6
charges [q] Au C1 H1
+ 0.70 -0.70 + 0.23
+ 0.62 0.71 + 0.18
+ 0.63 0.64 + 0.09
+ 0.64 0.55 + 0.03
+ 0.63 0.63 + 0.16
+ 0.19 0.67 + 0.54
details of each mechanism are reported in the following sections. Selected geometrical parameters of the complexes and their partial charges are documented in Tables 1–3. 2.1. CH Activation of Ethane
Figure 1. Structures of a) MCM-22 zeolite, b) 120T extended framework on which single-point calculations were performed, and c) 14T optimized reaction center.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
In a previous work,[11] we have shown that the adsorption of CO2 over AuI-ZSM-5 is less favorable than that of methane in the acetic acid production and that methane activation is needed in the first step of the reaction. Without catalyst, the direct conversion of methane ChemPhysChem 2014, 15, 514 – 520
515
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Table 2. Optimized geometrical parameters and calculated natural charge for the species involved in the reaction between the carbon dioxide and the gold ethyl hydride intermediate (INT 1) to produce propanoic acid over the 14T cluster model of AuI-MCM-22 calculated at the M06 L/6-31G(d,p) level of theory.
Table 3. Optimized geometrical parameters and calculated natural charge for the species involved in the reaction between the carbon dioxide and the gold ethyl adsorbed on the Brønsted-acid site (INT 2) to produce propanoic acid over the 14T cluster model of AuI-MCM-22 calculated at the M06 L/6-31G(d,p) level of theory.
Parameter
Parameter
distances [] AuC1 AuC2 AuH1 C1C2 C2O1 O1H1 AuOa AuOb
Isolated cluster
COADS 1
2.08 – 1.55 – 1.17 – 2.27 2.27
2.08 4.20 1.54 3.76 1.17 2.89 2.26 2.27
TS 3
2.07 2.67 1.93 3.17 1.24 1.14 2.27 2.50
INT 3
2.10 2.05 2.68 2.75 1.34 0.97 2.27 2.30
TS 4
2.28 2.21 2.78 1.79 1.35 0.97 2.18 2.48
PROD
3.35 3.01 – 1.50 1.32 0.97 2.83 2.14
distances [] AuC1 AuC2 OaH1 C1C2 C2O1 O1H1 AuOa AuOb
Isolated cluster
COADS 2
2.04 – 1.00 – – – 3.10 2.37
2.04 3.30 0.98 4.60 1.17 2.04 3.20 2.41
TS 5
2.22 2.23 1.50 1.94 1.30 1.04 4.13 2.59
PROD
3.35 3.01 – 1.50 1.32 0.97 2.83 2.14
angles [8] aSiaOaAl aSibObAl aOaAlOb
126.4 123.4 93.2
126.7 123.4 93.1
124.0 126.7 95.0
125.9 124.0 92.9
125.6 125.4 94.1
126.2 124.1 95.2
angles [8] aSiaOaAl aSibObAl aOaAlOb
126.7 123.8 94.6
127.2 122.8 96.2
125.6 122.8 97.9
126.2 124.1 95.2
charges [q] Au C1 H1 C2 O1
+ 0.64 0.55 + 0.03 + 1.03 0.51
+ 0.66 0.55 + 0.01 + 1.03 0.50
+ 0.40 0.67 + 0.41 + 1.05 0.61
+ 0.67 0.58 + 0.52 + 0.67 0.68
+ 0.66 0.67 + 0.51 + 0.69 0.69
+ 0.58 0.59 + 0.54 + 0.90 0.63
charges [q] Au C1 H1 C2 O1
+ 0.19 0.67 + 0.54 + 1.03 0.51
+ 0.12 0.67 + 0.58 + 1.06 0.55
+ 0.61 0.71 + 0.57 + 0.78 0.71
+ 0.58 0.59 + 0.54 + 0.90 0.63
and carbon dioxide to carboxylic acid is kinetically very unfavtransition state (TS 1), the ethane C1H1 bond is cleaved by orable with an Ea of 83.8 kcal mol1, as calculated at the M06 L/ the Au cation and hydride is transferred from the C1 atom to 6-31G(d,p) level, or 87.6 kcal mol1 if calculated at the MP2/6the metal center, leading to distances of 1.61 and 1.58 for 31G(d,p) level of theory. Therefore, a catalyst is necessary for C1···H and AuI···H, respectively. The distance between C1 and this conversion. In the reaction under consideration here, ethane takes the place of methane, starting with its physisorption on the active site. The energetic profile and geometries of relevant complexes are given in Figure 2. Selected geometrical parameters of the structures obtained along the reaction pathway and the natural population analysis (NPA) charges are tabulated in Table 1. We found that the h2 configuration (denoted ADS) is the most stable one. The C1 atom of ethane is 2.46 away from the AuI cation and the distance between AuI and the nearest H atom is 1.88 . The charge of Au becomes less positive (+0.62e) as compared to the bare zeolite. The C1H1 bond is lengthened from 1.09 to 1.15 . The adsorption energy is Figure 2. Energy profile of the Au-MCM-22-catalyzed reaction from ethane to the ethyl hydride (INT 1) and the 19.2 kcal mol1. At the first gold ethyl (INT 2) intermediate. Energies in kcal mol1. 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 514 – 520
516
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
Au decreases from 2.46 to 2.16 . The activation energy of this step is predicted to be 9.3 kcal mol1. This is lower than required for the activation of methane on a Au-MCM-22 catalyst (12 kcal mol1).[10] We note that this order is opposite to the strength of the CH bonds of these alkanes. The CH2CH3 group and the H atom stay coordinated to the AuI cation. The relative energy of INT 1 is 18.1 kcal mol1 with respect to the isolated systems. After that, the hydrogen atom can transfer from AuI to the oxygen atom of the zeolite framework while the CH2CH3 group remains bound to the metal. At the second transition state (TS 2), the Au···H1 and Oa···H1 distances become 1.55 and 2.12 and the AuC1 bond length is 2.06 . This step requires an activation energy of 16.3 kcal mol1. The charge of H1 increases slightly from + 0.03e in the INT 1 to + 0.16e in TS 2. Finally the second intermediate (INT 2) is formed. Gold ethyl, AuCH2CH3, is strongly adsorbed on the Brønsted-acid site with a relative energy of 21.9 kcal mol1 (INT 2). We checked the sufficiency of the size of the 5T region being relaxed by extending it to all atoms of the model except of the terminating atoms. We found that the results for ethane activation to be quite similar to the one obtained with the smaller 5T cluster. The values are included in the Supporting Information. Therefore, we feel confident that the practical relaxation scheme in this work is sufficiently accurate. The Wiberg bond order from the NBO analysis describes the bond formations and the bond breakings of the complexes along the reaction coordinate. In the Supporting Information the bond orders of several bonds are plotted against the reaction coordinate. From ADS to INT 1, the bond order of C1H1 decreases while the AuC1 and AuH1 bond orders increase to about 0.5, indicating the breaking of the C1H1 bond and the formation of the two new bonds to the AuI cation. From INT 1 to INT 2, the bond order of C1H1 decreases while the OaH1 bond orders increase from about zero to 0.57, indicating the generation of the Brønsted-acid site in the zeolite framework. Comparing the two steps of ethane activation over AuI-MCM-22, the apparent activation energies of both steps are below zero, meaning reactions should also occur easily. The stability of the adsorption complex (ADS) and the intermediates (INT 1 and INT 2) are comparable.
ethane over AuI-MCM-22 zeolite requires moderate activation energies and that the intermediates are thermodynamically stable compared to the adsorption complex. In this part, we examine the insertion of CO2 into these intermediates. Two pathways are possible: 1) The reaction between carbon dioxide and the gold ethyl hydride intermediate (INT 1) and 2) the reaction between carbon dioxide and gold ethyl adsorbed on the Brønsted-acid site (INT 2). In the first pathway, AuI acts as a monofunctional catalyst. Optimized geometries and the energetic diagram are shown in Figure 3. Selected geometrical parameters of the structures obtained along the reaction pathway and the NPA charges are given in Table 2. Firstly, the carbon dioxide adsorbs over the gold ethyl hydride intermediate (INT 1). The relative energy of this coadsorption complex, COADS, is 24.7 kcal mol1. In principle, the carbon dioxide molecule can be inserted into either the AuC1 or the AuH1 bonds of this intermediate. The first reaction produces a HAuOOCCH2CH3 unit which is bulkier than the HOOCAuCH2CH3 unit resulting from the second reaction. After the insertion into the AuH1 bond a gold carboxyl ethyl intermediate (INT 3) is formed. At the transition state, TS 3, the C2O1 bond distance of the carbon dioxide is increased from 1.17 to 1.24 , while the AuC2 and the O1H1 bond are forming with bond distances of 2.67 and 1.14 , respectively. The AuH1 bond is lengthened to 1.93 . This step requires an activation energy of 48.2 kcal mol1. The changes in the NPA charges from the coadsorption complex (COADS 1) to the transition state (TS 3) show that the charge of H1 increases from + 0.01 to + 0.41 e while the one of Au decreases from + 0.66 to + 0.40 e, showing a transfer of electron density from H1 to Au. The reaction proceeds with the formation of a gold
2.2. The Formation of Propanoic Acid In the previous section it was shown that the activation of
Figure 3. Energy profile of the Au-MCM-22-catalyzed reaction of carbon dioxide and the ethyl hydride intermediate (INT 1) to propanoic acid. Energies are in kcal mol1.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 514 – 520
517
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
carboxyl ethyl intermediate (INT 3) with a relative energy of 9.9 kcal mol1 with respect to the isolated system. There, both the C(O)OH and the CH2CH3 group bind to Au with an Au C1 distance of 2.10 and an AuC2 distance of 2.05 . The distance between the carbon atoms of the two groups is 2.75 . The Wiberg bond order demonstrates that the C2=O1 double bond of CO2 changes into a single bond. The AuC2 and the O1H1 bond form during this step (see the Supporting Information). The next step is the formation of propanoic acid by the connection of carboxyl C1 and ethyl C2. At the reaction between adsorbed carbon dioxide and gold ethyl (INT 2) to propanoic transition state, TS 4, the C1C2 Figure 4. Energy profile of the acid. Energies are in kcal mol1. distance becomes 1.79 while the bond distances of AuC1 and AuC2 increase from 2.10 and 2.05 to 2.28 and 2.21 , cess of insertion of a carbon dioxide molecule, forward rate respectively. This bond formation needs an activation energy constants, k, have been calculated for both pathways. The of 24.2 kcal mol1. The NPA charges change only slightly from DG ¼6 values from the 14T cluster model were calculated from the intermediate (INT 3) to the transition state (TS 4) except the energy differences corrected with the corresponding frefor the one of the C1 atom, which obtains electron density quency contributions. They were then used to derive the rate from Au and decreases from 0.58 to 0.67 e. The propanoic constants in the temperature range from 298.15 to 798.15 K acid is bound to the Au atom through the lone pair electrons using the Equation (1): of its carbonyl group. The AuO distance is 2.13 . The relative energy of this complex is 24.3 kcal mol1. The release of the kB T DG# ð1Þ kðTÞ ¼ exp ð Þ propanoic acid requires then a desorption energy of 34.9 kcal h RT mol1. The reaction profile of the second possible pathway, the reDG ¼6 is the difference between the free energy of the transiaction between carbon dioxide and gold ethyl on the Brønsttion state and the adsorption or intermediate complex, kB is ed-acid site (INT 2), is shown in Figure 4. Selected geometrical the Boltzmann’s constant, h is the Plank’s constant, T is the parameters of the structures obtained along the reaction pathtemperature in Kelvin, and R is the universal gas constant. The way and the NPA charges are given in Table 3. This pathway values of k for the elementary steps of the carbon dioxide inbegins with the adsorption of carbon dioxide on gold ethyl adsertion into the intermediates over AuI-MCM-22 at different sorbed on the Brønsted acid (COADS 2). This complex stabilizes the system by 5.0 kcal mol1. The reaction takes place via a bifunctional catalysis, including the protonation of the Table 4. Reaction rate constants (k) for the propanoic acid formation carbon dioxide by the Brønsted acid and the concurrent forstep from the monofunctional gold catalysis (kTS 3) and the concerted bifunctional mechanism (kTS 5) mation of the CC bond between the carbon dioxide and the gold ethyl group. At the transition state, TS 5, the distances Temperature [K] Rate constants OaH1 and O1H1 are 1.50 and 1.04 , respectively, while the kTS 5 kTS 5/kTS 3 kTS 3 distance between the carbon dioxide carbon atom (C2) and 23 21 298.15 2.80 10 1.75 10 62.6 the carbon atom of the ethyl group (C1) is 1.94 . The changes 3.84 1017 34.7 348.15 1.11 1018 6.68 1014 21.9 398.15 3.05 1015 in the NPA charges from the coadsorption complex (COADS 2) 2.16 1011 15.2 448.15 1.42 1012 to the transition state (TS 5) shows electron density transfers 2.16 109 11.3 498.15 1.91 1010 from Au (+ 0.12 e! + 0.61 e) to C2 (+ 1.06 e! + 0.78 e). This 8 8 9.25 10 8.89 548.15 1.04 10 step requires an activation energy of 44.2 kcal mol1, which is 2.09 106 7.27 598.15 2.88 107 2.91 105 6.15 648.15 4.73 106 lower than that determined for the first pathway (48.2 kcal 2.77 104 5.36 698.15 5.16 105 mol1). Thus the second pathway appears to be favored. 1.93 103 4.78 748.15 4.06 104 The rate constants for the reaction as a function of tempera798.15 2.44 103 1.06 102 4.34 ture can be examined by transition state theory. For the pro 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 514 – 520
518
CHEMPHYSCHEM ARTICLES temperatures are tabulated in Table 4. One sees that the path over Au acting as a bifunctional catalyst dominates at low temperatures, whereas at high temperatures the rate constants of the two pathways are comparable. The activation enthalpies and the free energies for the step of propanoic acid formation by the monofunctional (TS 3) and by the concerted bifunctional (TS 5) gold catalysis at different temperatures are available in the Supporting Information. One of the key issues for understanding adsorption phenomena and catalytic activities in a zeolitic system is the role of the confining van der Waals interactions in the framework. The extended 120T cluster of the supercage of MCM-22 zeolite was used for single-point calculations to study the effects of the framework. The relative energies of the systems involved in the reaction are summarized in Table 5. It was found that all complexes are stabilized by the zeolite framework. This results in slightly higher adsorption energies for both the reactant and the product. In contrast, the confinement effect lowers the activation energy of each system in the MCM-22 framework, except for that of TS 1. This could be explained by long-range interactions which play a role in the adsorption of ethane and propanoic acid and also influence the ion-pair-like transition state.
3. Conclusions The direct conversion of carbon dioxide and ethane to propanoic acid over gold-exchanged MCM-22 zeolite has been investigated by DFT calculations with the M06 L functional. The reaction begins with the activation of ethane to produce a gold ethyl hydride intermediate and then hydrogen transfers to the framework oxygen to produce gold ethyl adsorbed on the Brønsted-acid site. The energy barriers for these two steps of ethane activation are 9.3 and 16.3 kcal mol1, respectively. Two mechanisms of propanoic acid formation were investigated. In the first one, the insertion of CO2 into the AuH bond of the first intermediate yields a gold carboxyl ethyl intermediate. Subsequent conversion of this intermediate to propanoic acid occurs through the formation of a CC bond. The activation energy of the rate-determining step of this pathway is 48.2 kcal mol1. In the second mechanism investigated, CO2 interacts with gold ethyl adsorbed on the Brønsted-acid site. Propanoic acid is formed by the protonation of CO2 by the Brønsted acid and the concurrent bond formation between CO2 and the ethyl group. The activation energy of this reaction is 44.2 kcal mol1. From these results, the Au-MCM-22 zeolite should exhibit good performance for ethane activation. The reaction of CO2 with the gold ethyl and the Brønsted-acid site is found to be the rate-limiting step for the formation of propanoic acid. The high activation energy means that a sufficiently high temperature is needed in order to enable the catalytic CO2 conversion with a reasonable rate.
Computational Method In zeolite materials, the van der Waal interactions are more pronounced than in conventional heterogeneous catalysts. As a result, 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org dispersion-corrected functionals which overcome these are needed for the investigation of catalytic processes inside the zeolite pores. In the M06 family this kind of interaction is described efficiently.[18] Therefore, the M06-L functional, which is recommended for investigation of the confined interactions of adsorbates with zeolites involving transition metals, was employed in this work. The 631G(d,p) basis set was employed for all light atoms, while Au was described by the LANL2DZ effective core potential and the corresponding basis set. During geometry optimizations, only the 5T active region and the adsorbates were allowed to relax while the rest of the structure was fixed at the crystallographic coordinates. Transition states were located with the “Berny Algorithm”[19] and were checked to ensure that they have only one imaginary frequency in which the displacement vectors of all transition state structures correspond to the bond-forming and bond-breaking processes for the reaction steps. All systems were considered to be in the singlet state, which is the ground state of the Au ion. Single-point calculations were applied on a larger 120T cluster, which includes a supercage of the MCM-22 zeolite in order to cover the confinement effect from the zeolite framework. The NBO analysis was used to obtain charge distributions and bond orders of the systems.[20] All calculations were performed with the Gaussian 09 code.[21]
Table 5. Energies of the components in the conversion of CO2 and C2H6 to propanoic acid process over the Au-exchanged MCM-22 zeolite, given relative to the energy of the reactants.
Systems
ADS TS 1 INT 1 TS 2 INT 2 COADS 1 TS 3 INT 3 TS 4 COADS 2 TS 5 PROD
Relative energy [kcal mol1] 14T M06-L/6-31G(d,p) (optimized)
120T M06-L/6-31G(d,p) (single point)
19.2 9.9 18.1 1.8 21.9 24.7 23.5 9.9 14.3 26.9 17.3 24.3
21.2 11.4 (Ea.1 = 9.8) 18.6 4.9 (Ea.2 = 13.7) 22.1 24.8 20.0 (Ea.3 = 44.8) 13.5 12.0 (Ea.4 = 25.5) 27.0 13.7 (Ea.5 = 40.7) 28.6
(Ea.1 = 9.3) (Ea.2 = 16.3)
(Ea.3 = 48.2) (Ea.4 = 24.2) (Ea.5 = 44.2)
Acknowledgements This work was supported in part by grants from the National Science and Technology Development Agency (NSTDA Chair Professor funded by the Crown Property Bureau under the management of the National Science and Technology Development Agency and NANOTEC Center of Excellence funded by the National Nanotechnology Center), the Kasetsart University Research and Development Institute (KURDI), and the Commission on Higher Education, Ministry of Education (the “National Research University Project of Thailand (NRU)” and the “National Center of Excellence for Petroleum, Petrochemical and Advanced Materials (NCE-PPAM)”). ChemPhysChem 2014, 15, 514 – 520
519
CHEMPHYSCHEM ARTICLES Keywords: carbon dioxide conversion · density functional theory calculations · ethane · propanoic acid · zeolites [1] G. A. Olah, Angew. Chem. 2013, 125, 112 – 116; Angew. Chem. Int. Ed. 2013, 52, 104 – 107. [2] a) B. Eliasson, C.-j. Liu, U. Kogelschatz, Ind. Eng. Chem. Res. 2000, 39, 1221 – 1227; b) V. Havran, M. P. Dudukovic, C. S. Lo, Ind. Eng. Chem. Res. 2011, 50, 7089 – 7100; c) Y. Li, C.-J. Liu, B. Eliasson, Y. Wang, Energy Fuels 2002, 16, 864 – 870; d) C.-J. Liu, B. Xue, B. Eliasson, F. He, Y. Li, G.-H. Xu, Plasma Chem. Plasma Process. 2001, 21, 301 – 310; e) D. Shi, Y. Feng, S. Zhong, Catal. Today 2004, 98, 505 – 509; f) J.-g. Wang, C.-j. Liu, Y.-p. Zhang, B. Eliasson, Chem. Phys. Lett. 2003, 368, 313 – 318. [3] a) D. C. Tranca, N. Hansen, J. A. Swisher, B. Smit, F. J. Keil, J. Phys. Chem. C 2012, 116, 23408 – 23417; b) P. M. Zimmerman, D. C. Tranca, J. Gomes, D. S. Lambrecht, M. Head-Gordon, A. T. Bell, J. Am. Chem. Soc. 2012, 134, 19468 – 19476. [4] E. M. Wilcox, G. W. Roberts, J. J. Spivey, Catal. Today 2003, 88, 83 – 90. [5] Y.-H. Ding, W. Huang, Y.-G. Wang, Fuel Process. Technol. 2007, 88, 319 – 324. [6] W. Huang, K. C. Xie, J. P. Wang, Z. H. Gao, L. H. Yin, Q. M. Zhu, J. Catal. 2001, 201, 100 – 104. [7] W. Huang, W. Z. Sun, F. Li, AIChE J. 2010, 56, 1279 – 1284. [8] a) V. B. Kazansky, I. R. Subbotina, N. Rane, S. R. A. van, E. J. M. Hensen, Phys. Chem. Chem. Phys. 2005, 7, 3088 – 3092; b) G. Caeiro, R. H. Carvalho, X. Wang, M. A. N. D. A. Lemos, F. Lemos, M. Guisnet, F. R. Ribeiro, J. Mol. Catal. A 2006, 255, 131 – 158; c) A. G. Stepanov, V. N. Parmon, D. Freude, Kinet. Catal. 2007, 48, 521 – 534; d) A. Itadani, H. Torigoe, T. Yumura, T. Ohkubo, H. Kobayashi, Y. Kuroda, J. Phys. Chem. C 2012, 116, 10680 – 10691; e) Y. G. Kolyagin, I. I. Ivanova, Y. A. Pirogov, Solid State Nucl. Magn. Reson. 2009, 35, 104 – 112. [9] a) E. A. Pidko, V. B. Kazansky, E. J. M. Hensen, S. R. A. Van, J. Catal. 2006, 240, 73 – 84; b) E. A. Pidko, S. R. A. van, J. Phys. Chem. C 2007, 111, 2643 – 2655; c) H. A. Aleksandrov, G. N. Vayssilov, Catal. Today 2010, 152, 78 – 87; d) O. A. Anunziata, J. Cussa, A. R. Beltramone, Catal. Today 2011, 171, 36 – 42. [10] S. Wannakao, C. Warakulwit, K. Kongpatpanich, M. Probst, J. Limtrakul, ACS Catal. 2012, 2, 986 – 992. [11] W. Panjan, J. Sirijaraensre, C. Warakulwit, P. Pantu, J. Limtrakul, Phys. Chem. Chem. Phys. 2012, 14, 16588 – 16594. [12] a) J. C. Fierro-Gonzalez, B. C. Gates, J. Phys. Chem. B 2004, 108, 16999 – 17002; b) Z.-X. Gao, Q. Sun, H.-Y. Chen, X. Wang, W. M. H. Sachtler, Catal. Lett. 2001, 72, 1 – 5; c) M. M. Mohamed, T. M. Salama, M. Ichikawa, J. Colloid Interface Sci. 2000, 224, 366 – 371; d) M. M. Mohamed, T. M. Salama, R. Ohnishi, M. Ichikawa, Langmuir 2001, 17, 5678 – 5684; e) S. Qiu, R. Ohnishi, M. Ichikawa, J. Phys. Chem. 1994, 98, 2719 – 2721; f) S. Qiu, R. Ohnishi, M. Ichikawa, J. Chem. Soc. Chem. Commun. 1992, 19, 1425 – 1427;
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
[13] [14]
[15] [16] [17] [18]
[19] [20] [21]
g) T. M. Salama, T. Shido, R. Ohnishi, M. Ichikawa, J. Phys. Chem. 1996, 100, 3688 – 3694. M. M. Mohamed, M. Ichikawa, J. Colloid Interface Sci. 2000, 232, 381 – 388. a) M. Bjorgen, S. Akyalcin, U. Olsbye, S. Benard, S. Kolboe, S. Svelle, J. Catal. 2010, 275, 170 – 180; b) B. Onida, F. Geobaldo, F. Testa, F. Crea, E. Garrone, Stud. Surf. Sci. Catal. 2000, 130, 2951 – 2956; c) X. Wang, W. Dai, G. Wu, L. Li, N. Guan, M. Hunger, Microporous Mesoporous Mater. 2012, 151, 99 – 106. M. E. Leonowicz, J. A. Lawton, S. L. Lawton, M. K. Rubin, Science 1994, 264, 1910 – 1913. D. Zhou, Y. Bao, M. Yang, N. He, G. Yang, J. Mol. Catal. A 2006, 244, 11 – 19. J. Lu, C. Aydin, N. D. Browning, B. C. Gates, Angew. Chem. 2012, 124, 5944-5948; Angew. Chem. Int. Ed. 2012, 51, 5842 – 5846. a) B. Boekfa, S. Choomwattana, P. Khongpracha, J. Limtrakul, Langmuir 2009, 25, 12990 – 12999; b) Y. Zhao, D. G. Truhlar, J. Phys. Chem. C 2008, 112, 6860 – 6868; c) Y. Zhao, D. G. Truhlar, Theor. Chem. Acc. 2008, 120, 215 – 241; d) Y. Zhao, D. G. Truhlar, Rev. Mineral. Geochem. 2010, 71, 19 – 37; e) M. Mantina, R. Valero, D. G. Truhlar, J. Chem. Phys. 2009, 131, 064706/1 – 5; f) C. Kumsapaya, K. Bobuatong, P. Khongpracha, Y. Tantirungrotechai, J. Limtrakul, J. Phys. Chem. C 2009, 113, 16128 – 16137; g) J. Sirijaraensre, J. Limtrakul, Phys. Chem. Chem. Phys. 2013, 15, 18093 – 18100; h) T. Maihom, P. Khongpracha, J. Sirijaraensre, J. Limtrakul, ChemPhysChem 2013, 14, 101 – 107. a) C. Gonzalez, H. B. Schlegel, J. Chem. Phys. 1989, 90, 2154 – 2161; b) H. B. Schlegel, J. Comput. Chem. 1982, 3, 214 – 218. a) A. E. Reed, F. Weinhold, J. Chem. Phys. 1983, 78, 4066 – 4073; b) A. E. Reed, L. A. Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899 – 926. Gaussian 09 (Revision C.01), M. J. T. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, N. J. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels,. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox, Gaussian, Inc., Wallingford CT, 2010.
Received: October 11, 2013 Revised: November 25, 2013 Published online on December 20, 2013
ChemPhysChem 2014, 15, 514 – 520
520
