MT

Da

Mb

c

A

a

ARRAA

KHCSHDO

1

aiom[dcccnid

hAM

i

0h

Applied Catalysis A: General 475 (2014) 134– 139

Contents lists available at ScienceDirect

Applied Catalysis A: General

j ourna l h omepa ge: www.elsev ier .com/ locate /apcata

etal–support interactions revisited by theoretical calculations:he influence of organic ligands for preparing Ni/SiO2 catalysts

iego Valenciaa,b,∗, Laura Penab, Víctor Hugo Ucc, Isidoro García-Cruza,∗

Programa de Ingeniería Molecular, Instituto Mexicano del Petróleo, Eje Central Lázaro Cárdenas 152, Colonia San Bartolo Atepehuacán, Gustavo A.adero, Mexico D. F., 07730, MexicoFacultad de Química, Universidad Nacional Autónoma de México, Ciudad Universitaria, Coyoacán, México D.F., 04510, MexicoÁrea de Física Atómica Molecular Aplicada (FAMA), CBI, Universidad Autónoma Metropolitana-Azcapotzalco, Av. San Pablo 180, Col. Reynosa Tamaulipas,zcapotzalco, Mexico D. F., 02200, Mexico

r t i c l e i n f o

rticle history:eceived 1 October 2013eceived in revised form 6 January 2014ccepted 8 January 2014vailable online 21 January 2014

eywords:eterogeneous catalysts

a b s t r a c t

We performed a theoretical study of the interaction between Ni-containing compounds and amorphousSiO2 support, which have been used for preparing supported catalysts. The aim of this work is to under-stand the effect of the chemical structure of the metal species and its interaction with the support duringthe preparation of heterogeneous catalysts. The metal–support interactions were revisited for a series ofNi-containing compounds with different ligands adsorbed on SiO2. We studied some of most commonlyused ligands for preparing Ni/SiO2 catalysts such as ethylenediamine, citric acid and EDTA. Replacementof the first atom of the ligand bound to Ni by the –OH surface group of the support was modeled using

helating agentsupportydrogen bondFTNIOM

the ONIOM combined quantum mechanics/molecular mechanics (QM/MM) approach. The adsorptiongeometry and the Gibbs energy are clearly related to the ligand structure. We also noted the stabilizationof the adsorbed compounds through cooperative hydrogen bonds with the –OH moiety adjacent surfacegroup present on SiO2. We point out the importance of the ligand structure in the metal–support inter-actions. These results can be useful to understand the metal– and ligand–support interactions during thepreparation of supported catalysts.

. Introduction

Preparation of heterogeneous catalysts has received muchttention because diverse knowledge is needed to understandts main aspects [1]. This interdisciplinary field has been devel-ped with great interest because the catalytic performance of theaterials depends on several ways in the preparation strategies

2,3]. Supported catalysts have been used for industrial processesue to their thermal, mechanical and chemical stability [4,5]. Theomposition of these catalysts typically consists in three mainomponents: active phase, promoter and support. Their chemicalomposition is closely related to the catalytic process and its eco-

omic viability. At the molecular level, the catalytic performance

s related to the composition, size of active particles, morphology,ispersion and metal–support interactions [6–10].

∗ Corresponding authors at: Instituto Mexicano del Petroleo, San Bartolo Atepe-uacan, Gustavo A. Madero 07730, Mexico D.F., Mexico and Universidad Nacionalutonoma de Mexico, Facultad de Quımica, Ciudad Universitaria, Coyoacan 04510,exico D.F, Mexico. Tel.: +52 55 9175 6903; fax: +52 55 9175 6380.

E-mail addresses: dvalencia@comunidad.unam.mx (D. Valencia),garcia@imp.mx (I. García-Cruz).

926-860X/$ – see front matter © 2014 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.apcata.2014.01.018

© 2014 Elsevier B.V. All rights reserved.

The metal precursor used to prepare supported catalysts is animportant factor on their characteristics and performance [11–15].The use of organic ligands during the impregnation or grafting themetal species to the support has provided a good key to modify themetal precursors. Modification of the metal species in the impreg-nation solution is one of the most important goals of the organicligands. Election of each one is based on its stability constant whichis closely related to the stability of the metal complexes in aqueoussolution. Typically, the hydrotreating catalysts have shown a betterperformance when they were prepared using ligands with higherstability constants [16–19]. For that reason, some of the compoundsused are polydentate ligands, called chelating agents. Nowadays,they have also been used for preparing industrial catalysts forFischer–Tropsch [20], oxidation [21], hydrogenation [22,23] andhydrodechlorination [24] processes. In general, the catalytic per-formance of the catalysts prepared with chelating agents is betterthan the one for those prepared with the typical hexaaquo-basedcompounds.

The mechanism by which the chelating agents act during the

preparation of supported catalysts is still discussed. It involvesthe chemical speciation of metal ions in aqueous solutions, pH ofthese ones, isoelectric point of the support, thermal treatment anddecomposition of organic compounds in the catalytic materials as

ysis A: General 475 (2014) 134– 139 135

wtr[(tstud

thcicilbrshtttTwRmets

stpars

2

aW[emhihteciTt

sHttrct

Table 1Bond length of nickel and X(n) atoms of ligands shown in Fig. 1.

Bonda Compound

[Ni(H2O)6]2+ [Ni(en)3]2+ [Ni(cit)2]4− [Ni(EDTA)]2−b

Ni–X(1) 2.07 2.17 2.11 2.07Ni–X(2) 2.07 2.17 2.13 2.10Ni–X(3) 2.07 2.17 2.09 2.15Ni–X(4) 2.07 2.17 2.13 2.10Ni–X(5) 2.07 2.17 2.09 2.15Ni–X(6) 2.07 2.17 2.11 2.07

D. Valencia et al. / Applied Catal

ell as specific metal–support interactions [25–29]. Characteriza-ion of the supported catalysts prepared with chelating agents hasevealed that the metal species were well dispersed on the support30,31]. The size of the catalytic particles, after different treatmentse.g. thermal, reducing, oxidation or sulfidation), is smaller due tohe use of the chelating agents. The thermal treatment of the metalpecies deposited on the oxide supports provided some insight intohe effect of the organic compounds. Those species can also reactnder N2 and H2 atmospheres at different temperatures and formifferent intermediates [32,33].

The effect of the ligand structure on the metal–support interac-ions is still difficult to understand. For this reason, several authorsave proposed that the support can act as another ligand thatan change the chemical equilibrium of the metal species dur-ng the preparation of heterogeneous catalysts [34–36]. This rolean be expected if the substitution of the ligand takes place dur-ng the impregnation or grafting step. The chemical equilibrium ofigand substitution is difficult to study when the ligand is a supportecause it has a lot of surface atoms, typically oxygen, which shouldeplace the atoms’ ligand in the coordination sphere of the metalpecies [27,37]. A comprehensive picture of this process has beenypothetically developed [38]. Moreover, computational studies ofhe support are difficult because its large number of atoms andhe current theoretical methods to study them. On the other hand,he simulation of the support requires its proper representation.here are two main approximations for this purpose: slab modelith periodic boundary conditions [39] and QM/MM models [40].egarding the accuracy of the computational chemistry model, DFTethods allow to study the support effects properly. The new gen-

ration of density functionals could be an useful tool to estimatehe energy changes and specific interactions which occur in theupported catalysts.

Since the preparation of supported catalysts requires the under-tanding of the properties at the metal-support interface, the aim ofhis work is to obtain insight into the ligand structure effect on thereparation of Ni/SiO2 catalysts from the molecular standpoint. Toccomplish this purpose, we performed a theoretical study of theeplacement of an atom of the ligand in the metal coordinationphere by an –OH surface group of the amorphous SiO2 support.

. Computational details

The Ni(II) compounds studied have octahedral symmetrynd their electronic configuration are [Ar]3d8 for all cases.e used four different Ni-containing compounds. Those are

Ni(H2O)6]2+, [Ni(en)3]2+, [Ni(cit)2]4− and [Ni(EDTA)]2−, wheren = ethylenediamine, cit3− = citrate and EDTA4− = ethylenedia-inetetraacetate. These compounds were selected because they

ave been used in several experimental studies when Ni(II) wasncorporated to the support [16–24]. The hexaaquo complexas been the typical Ni(II) precursor. The optimized struc-ures of these compounds were obtained using the hybrid metaxchange–correlation density functional M06 [41]. This newly-reated functional has shown a broad applicability to many areasncluding metal transition chemistry [42]. We utilized the standardZVP basis set. All harmonic vibrational frequencies were real;herefore the structures are minima on the potential energy surface.

The support model consists in a large SiO2. This model was con-tructed from amorphous SiO2 coordinates with a composition of190Si286O577. We performed two-layer ONIOM calculations due to

he large number of atoms. According to the ONIOM approach [43],

he calculation of energies can be simplified by treating the activeegion as a model and it can be calculated with QM. The real systemontains all the atoms, and it is calculated only at the MM level. In awo-layer calculation, the total energy of the whole system can be

a X=O or N.b X(3) and X(5) are nitrogen atoms in EDTA.

expressed within the framework of the ONIOM methodology as:E(ONIOM) = Elow

real + Ehighmodel − Elow

model. We treated the Ni-containingcompounds and two hydroxyl groups from the SiO2 surface at thehigh level with the rest of the support at the low level. The accuracyof the QM/MM method, particularly the ONIOM method, dependssignificantly on the choice of the level of calculations for regionsof high- and low-levels of accuracy. In this study, the high-levelregion was calculated using the hybrid meta exchange–correlationdensity functional M06 and standard TZVP basis set and the low-level with the Universal force field (UFF) calculations. All harmonicvibrational frequencies were real. All calculations were performedwith the Gaussian 09 software [44].

The adsorption energies were calculated as: GMS−(GM + GS),where GMS is the Gibbs energy of the Ni(II) precursor adsorbed onthe support, GM is the Gibbs energy of the Ni(II) precursor and GSis the Gibbs energy of the support.

3. Results and discussion

Optimized geometries for Ni-containing compounds exhibitedan octahedral symmetry for all the ligands studied (Fig. 1). Itis in good agreement with crystal field theory, the d orbitals ofNi(II) interact with the electron donating atoms in the ligand. Thismeans that coordination compounds were formed through O andN electron donating atoms around the Ni(II) atom. The bond dis-tances of Ni-X(n) are displayed in Table 1. The highest symmetrywas observed for [Ni(H2O)6]2+ and [Ni(en)3]2+ compounds, all Ni-X(n) bond lengths are 2.07 A and 2.17 A, respectively. We can seesome differences in those bond distances for the [Ni(cit)2]4− and[Ni(EDTA)]2− compounds. They are also octahedral compounds, butexhibited three different bond distances. These differences can beexplained by a distortion of the octahedral symmetry due to thoseligands are binding the Ni(II) ion. It is important to note that themain differences of these compounds are related to the number ofatoms that binds the Ni(II) ion. Water is the poorest ligand in thisseries because only one atom in one molecule binds the Ni(II) atom.The chelating agents have two (en), three (cit) and six (EDTA) atomsin one molecule which are binding the nickel ion in the coordinationcompounds.

We simulated the chemical reaction which takes place duringthe replacement of the first atom of the chelating agent by one oxy-gen atom at the surface support (Fig. 2). In this theoretical approach,the metal–support interactions between the [Ni(L)n]y compoundsand the SiO2 support depend on the ligand structure. The replace-ment showed in Fig. 2 is important to understand the effect of thesupport as a ligand and the one of the coordination compounds’stability during the preparation of Ni-based supported catalysts.

The optimized geometry for the adsorption of [Ni(H2O)6]2+ is

shown in Fig. 3. The octahedral symmetry around the Ni(II) ion ismaintained after the geometry optimization and the substitutionof one H2O molecule from the coordination sphere of [Ni(H2O)6]2+

is clearly observed. In this compound adsorbed on SiO2, the O(6)

136 D. Valencia et al. / Applied Catalysis A: General 475 (2014) 134– 139

Fig. 1. Optimized geometries for (a) [Ni(H2O)6]2+, (b) [Ni(en)3]2+, (c) [Ni(cit)2]4− and (d) [Ni(EDTA)]2− .

Fig. 2. Chemical reaction for the ligand substitution,which represents the rep

Fig. 3. Optimized geometry for [Ni(H O) ]2+ on SiO . The high-level part is shownb

2 6 2

y a ball-and-stick representation and the low-level part by wireframe model.

lacement of the first atom of the ligand bound to Ni(II) by the support.

atom from the H2O molecule is replaced by the O(1)sup atom fromthe SiO2. This replacement is driven by the affinity of the oxygensurface atom from the support and the Ni(II) ion, where the sup-port is acting as another ligand. It is important to note that anotherinteraction is formed between the adjacent –OH group from thesupport and the leaving H2O molecule. It can be seen a hydrogenbond between the H(1) atom in the water molecule and the O(2)sup

atom from the support with a bond distance 1.69 A. This hydro-gen bond between the substituted H2O molecule and the supportshould be taken into account to deeper understand the processeswhich take place during the preparation of supported catalysts.

The [Ni(en)3]2+ compound was also adsorbed on the SiO2 sup-port. In this compound, the ethylenediamine ligand has two atomsin one molecule which are binding the metal ion. The optimizedgeometry is depicted in Fig. 4. The N(6) atom of the en ligand isalso replaced by the O(1)sup atom of the SiO2. The N substitution ofthe en ligand by the O atom from the support exhibited different

features than the one observed in the [Ni(H2O)6]2+ compound. Theimportance of the hydrogen bonds between the en ligand and thesupport is clearly observed. The structure of en ligand requires the

D. Valencia et al. / Applied Catalysis A: General 475 (2014) 134– 139 137

Fa

stiswgTlmtwtd

stiacas[thd

Fa

ig. 4. Optimized geometry for [Ni(en)3]2+ on SiO2. The high-level part is shown by ball-and-stick representation and the low-level part by wireframe model.

tabilization of the compound because the other N(4) atom fromhe –NH2 group in the coordination sphere is binding the Ni(II)on. In this line, the chelate effect is remarkable to maintain thetable structure. There are two hydrogen bonds in this compoundith the surface atoms of the support at the same –NH2 terminal

roup, those are N(6)–H(1)sup (1.76 A) and H(1)–O(2)sup (1.91 A).he metal–support interactions need the cooperation between theeaving –NH2 group and the –OH moieties at the surface support to

aintain the structure of the coordination compound. The effect ofhe support as a ligand is also remarkable to form hydrogen bondsith the atoms which are replaced. These ligand–support interac-

ions could be a necessary step to stabilize the adsorbed compoundsuring the preparation of the supported catalysts.

The optimized geometry for [Ni(cit)2]4− adsorbed on SiO2 ishown in Fig. 5. The octahedral symmetry was also conserved afterhe interaction with the support. The O(6) atom of the cit ligands replaced by the O(1)sup of the SiO2. This substitution resulted in

hydrogen bond formation distinct than the one observed in thease of [Ni(H2O)6]2+ and [Ni(en)3]2+ because these two ligands have

hydrogen atom (O(6)-H(1) and N(6)-H(1), respectively) in theubstituted groups which is able to interact with the O(2)sup. TheNi(cit)2]4− compound does not have hydrogen atoms in the substi-

uted moiety. Citrate ligand has an oxygen atom, O(6), which formsydrogen bonds with the H(2)sup atom of the support, the bondistance between them is 1.65 A. We also detected the presence

ig. 5. Optimized geometry for [Ni(cit)2]4− on SiO2. The high-level part is shown by ball-and-stick representation and the low-level part by wireframe model.

Fig. 6. Optimized geometry for [Ni(EDTA)]2− on SiO2. The high-level part is shownby a ball-and-stick representation and the low-level part by wireframe model.

of another hydrogen bond in this system in different region of thecitrate molecule, O(3)–H(1)sup with a bond length of 1.91 A. As wementioned above, the presence of hydrogen bonds between theligands and the support are also important for preparing heteroge-neous catalysts using cit as chelating agent.

EDTA is a strong chelating agent. It has six atoms in its struc-ture that can interact with the Ni(II) ion. The chelate effect in[Ni(EDTA)]2− is the largest for all the studied compounds in thiswork. The optimized geometry for this compound adsorbed on SiO2is shown in Fig. 6. The O(6) atom of EDTA in the coordination sphereof the Ni-containing compound was also replaced by the O(1)sup

atom of SiO2. The O(6) atom forms two hydrogen bonds with theH(1)sup and H(2)sup atoms of the support, their bond distances are:1.65 A and 1.91 A, respectively. The optimized geometry for thissystem is similar than the obtained one for the cit because the oxy-gen atom of the ligand can interact with the hydrogen atoms ofthe surface support. Moreover, EDTA is only one ligand that bindsthe nickel atom in the coordination compound, the replacement ofO(6) by the oxygen atom of the support preserved the octahedralsymmetry around the metal ion with a hydrogen bond formationto perform this process.

Table 2 gathers the values of Ni–X(n) bond distances for allthe Ni-containing compounds adsorbed on SiO2. The replacementof the X(6) atom in each ligand modifies all the bond distancesbetween the O or N and Ni atoms. We found some differences withrespect to those bond distances for the pure coordination com-pounds. No matter the coordination compound, we noted that allthe Ni–X(n) bond distances are different in each compound after

they interact with silica. In this line, if the support can be viewedas a ligand, it resulted in symmetry distortion of Ni-containingcompounds because SiO2 can be considered as bulky one. The effect

Table 2Selected bond lengths of the Ni-containing compounds adsorbed on SiO2.

Bonda Compound

[Ni(H2O)6]2+ [Ni(en)3]2+ [Ni(cit)2]4− [Ni(EDTA)]2−b

Ni–X(1) 2.07 2.13 2.09 2.04Ni–X(2) 2.09 2.14 2.09 2.06Ni–X(3) 2.04 2.11 2.14 2.05Ni–X(4) 2.03 2.18 2.06 2.07Ni–X(5) 2.07 2.15 2.09 2.09Ni–O(1)sup 2.09 2.23 2.20 2.25X(6)–O(2)sup 2.68 2.90 2.62 2.85X(6)–H(2)sup 3.18 3.54 1.65 1.91H(1)–O(2)sup 1.69 1.91 – –

a X=O or N.b X(3) and X(5) are nitrogen atoms in EDTA.

138 D. Valencia et al. / Applied Catalysis A

Table 3Standard Gibbs energy for the process depicted in Fig. 2.

Compound �G0 (kcal/mol)

[Ni(H2O)6]2+ −32.1[Ni(en)3]2+ −20.4

oaraw2Ttscittirs

deboooNatiNpaptoetg(a((G[[ff

hfgacutgwlc

[6] A.M. Molenbroek, S. Helveg, H. Topsoe, B.S. Clausen, Top. Catal. 52 (2009)

[Ni(cit)2]4− −46.0[Ni(EDTA)]2− −23.4

f the ligand structure on the symmetry of these compounds islso observed on the Ni–O(1)sup bond distance which is the bondeplaced by the support, in other words, the metal-support inter-ction. [Ni(H2O)6]2+ resulted in the smallest bond length (2.09 A),hereas this bond distance for the other compounds is about

.20 A. These differences are closely related to the chelate effect.he chelating agents: en, cit and EDTA are polydentate ligands,herefore the replacement of one atom of each molecule by theupport needed the stabilization of the respective Ni-containingompound, with a larger Ni–O(1)sup bond length. The differencesn this bond length can also be attributed to steric effect due tohe chelating agent structure and the support. On the other hand,he Ni–Sisup bond distance for all the Ni-based compounds stud-ed is around 3.4 A which is in good agreement with the EXAFSesults reported in an experimental work for a related Ni/SiO2ystem [35].

Finally, we calculated the Gibbs energies for the processepicted in Fig. 2 for all of the Ni-containing compounds. Thesenergetic values are important to understand the interactionetween the coordination compounds and the –OH surface groupsf SiO2. They were not calculated previously due to the complexityf the systems and the computational limitations on the accuracyf the available density functionals. The interaction of thesei-containing compounds with silica is an exergonic process forll the cases (Table 3). The reference compound in this work ishe [Ni(H2O)6]2+ because it is the typical compound which ismpregnated when the common nickel salts are used for preparingi-based catalyst supported on SiO2. Some experimental studiesroposed intuitively the use of chelating agents can result in smallerdsorption energy because the substitution of an atom from aolydentate ligand could be more difficult [27,38]. That decrease inhe adsorption energy could be the reason of the larger dispersionf the nickel species on the support. It is a proper explanation of thexperimental results because the replacement of an atom from a bi,ri or hexadentate ligand (e.g. en, cit or EDTA) could be less ener-etically favorable than the one from the typical hexaaquo complexmonodentate ligand). We can note that this intuitive trend is

true fact for almost all the compounds studied [Ni(H2O)6]2+

−32.1 kcal/mol) > [Ni(EDTA)]2− (−23.4 kcal/mol) > [Ni(en)3]2+

−20.4 kcal/mol). However, [Ni(cit)2]4− exhibited the largestibbs energy (−46.0 kcal/mol) even than the obtained one for the

Ni(H2O)6]2+. This energetic discrepancy is not unexpected as theNi(cit)2]4− compound exhibited two hydrogen bonds betweenour atoms when it is adsorbed on SiO2 and they can be responsibleor this larger energy value.

We noted the cooperation of coordinative covalent bonds withydrogen bonds in the metal– and ligand–support interactions

or the Ni-containing compounds adsorbed on SiO2. This is inood agreement with the proposed molecular recognition mech-nism for preparing supported catalysts using amine-based metalomplexes [35]. It is important to note that the Gibbs energy val-es evinced the stabilization of the adsorbed surface species dueo the ligands. This can be reached through cooperative hydro-en bond between the ligands and the support. Furthermore,

e are able to establish the importance of both the metal– and

igand–support interactions during the preparation of supportedatalysts.

: General 475 (2014) 134– 139

Based on the present theoretical results, we can point out thespecial treatment of the support as a bulky ligand. Due to its largernumber of atoms it is also able to form hydrogen bonds withthe ligands during the replacement of those in the coordinationsphere of the metal species. The instances of precise molecu-lar identification of adsorbed metal complexes are still scarce byexperimental techniques [27]. Proper molecular simulations canprovide insights into the mechanism by which organic ligands actduring the preparation of supported catalysts and their impact onthe metal–support interactions. Our results could sustain the chem-ical species interacting with the support, which were suggestedexperimentally, but not fully confirmed [32,35,37]. The cooperativemetal– and ligand–support interactions produce a supramolecu-lar system which could explain the good dispersion of the metalspecies in the supported catalysts when they were prepared usingpolydentate ligands. It could be a little diffusion of the metal specieswhen they are forming supramolecular structures adsorbed on thesupport and it can avoid their agglomeration. We consider theligand–support interactions are very important and they shouldbe taken into account to understand the mechanisms which occurduring the preparation of supported catalysts. According to ourtheoretical calculations, the ligand–support interactions can beclassified into two groups for preparing supported catalysts: (i)L → Sup and (ii) Sup → L, in the first case a H atom from the organicligand forms H bond with an O atom of the surface support, in thelatter case a H atom from support can interact with an electroneg-ative atom from the organic ligand.

4. Conclusions

The effect of the ligand structure during the preparation of sup-ported catalysts was elucidated by means of ONIOM calculations.The structure of the ligand is relevant to form the metal–supportinteraction. We noted important differences between the typicalnickel precursor [Ni(H2O)6]2+ and the ones prepared with chelatingagents when they interact with the support. The replacement ofeach ligand depends on its stability as it was shown by the calcu-lated Gibbs energy. The substitution of the ligand atom bound to Nirequires the cooperative participation of the –OH groups at the sur-face support. It was clearly demonstrated that a proper treatmentof the support as a bulky ligand needs special attention to formmetal–support and also ligand–support interactions. The hydrogenbond is the main interaction in the latter and it strongly dependson the ligand structure.

Acknowledgments

The authors wish to thank Laboratorio de Visual-ización y Cómputo Paralelo de la Universidad AutónomaMetropolitana—Iztapalapa for CPU time in the Aitzaloa Supercom-puter and LUFAC Computación S.A. de C.V. D.V. also acknowledgesInstituto Mexicano del Petróleo for their kind hospitality, as wellas R. Olivares-Amaya for insightful discussions.

References

[1] B.M. Weckhuysen, Angew. Chem. Int. Ed. 48 (2009) 4910–4943.[2] J.A. Schwarz, C. Contescu, A. Contescu, Chem. Rev. 95 (1995) 477–510.[3] M. Campanati, G. Fornasari, A. Vaccari, Catal. Today 77 (2003) 299–314.[4] M. Di Serino, R. Tesser, L. Pengmei, E. Santacesaria, Energy Fuels 22 (2008)

207–217.[5] R.A. Sheldon, M. Wallau, I.W.C.E. Arends, U. Schuchardt, Acc. Chem. Res. 31

(1998) 485–493.

1303–1311.[7] F. Zaera, J. Phys. Chem. Lett. 1 (2010) 621–627.[8] T.W. Hansen, J.B. Wagner, P.L. Hansen, S. Dahl, H. Topsoe, C.J.H. Jacobsen, Science

294 (2001) 1508–1510.

ysis A

[[[

[

[

[

[

[

[[[[[[[

[[

[[

[[

[

[

[

[[[

[[

[

[

[[[

[

D. Valencia et al. / Applied Catal

[9] S.J. Tauster, S.C. Fung, R.L. Garten, J. Am. Chem. Soc. 100 (1978) 170–175.10] D.W. Goodman, Catal. Lett. 99 (2005) 1–4.11] S. Wang, G.Q. Lu, Appl. Catal., A 169 (1998) 271–280.12] P. Kim, H. Kim, J. Bong Joo, W. Kim, I. Kyu Song, Jongheop Yi, J. Mol. Catal. A:

Chem. 256 (2006) 178–183.13] A. Maroto-Valiente, M. Cerro-Alarcón, A. Guerrero-Ruiz, I. Rodríguez-Ramos,

Appl. Catal., A 283 (2005) 23–32.14] H. Kusama, K. Kitamura Bando, K. Okabe, H. Arakawa, Appl. Catal., A 205 (2001)

285–294.15] G. Wu, C. Zhang, S. Li, Z. Han, T. Wang, X. Ma, J. Gong, ACS Sustainable Chem.

Eng. 1 (2013) 1052–1062.16] K. Hiroshima, T. Mochizuki, T. Honma, T. Shimizu, M. Yamada, Appl. Surf. Sci.

121/122 (1997) 433–436.17] T. Shimizu, K. Hiroshima, T. Honma, T. Mochizuki, M. Yamada, Catal. Today 45

(1998) 271–276.18] R. Cattaneo, T. Shido, R. Prins, J. Catal. 185 (1999) 199–212.19] L. Pena, D. Valencia, T. Klimova, Appl. Catal., B 147 (2014) 879–887.20] T. Mochizuki, T. Hara, N. Koizumi, M. Yamada, Appl. Catal., A 317 (2007) 97–104.21] D. Prieto-Centurión, J. Notestein, J. Catal. 279 (2011) 103–110.22] K.-Q. Sun, E. Marceau, M. Che, Phys. Chem. Chem. Phys. 8 (2006) 1731–1738.23] S. Schimpf, C. Louis, P. Claus, Appl. Catal., A 318 (2007) 45–53.24] A.I. Tsyganok, K. Suzuki, S. Hamakawa, K. Takehira, T. Hayakawa, Catal. Lett. 77

(2001) 75–86.25] J.P. Brunelle, Pure Appl. Chem. 50 (1978) 1211–1229.26] G. Kishan, L. Coulier, V.H.J. de Beer, J.A.R. van Veen, J.W. Niemantsverdriet, J.

Catal. 196 (2000) 180–189.27] J.-F. Lambert, M. Che, J. Mol. Catal. A: Chem. 162 (2000) 5–18.28] J. Escobar, M.C. Barrera, J.A. de los Reyes, J.A. Toledo, V. Santes, J.A. Colín, J. Mol.

Catal. A: Chem. 287 (2008) 33–40.29] D. Valencia, T. Klimova, Catal. Commun. 21 (2012) 77–81.30] A.J. van Dillen, R.J.A.M. Terörde, D.J. Lensveld, J.W. Geus, K.P. de Jong, J. Catal.

216 (2003) 257–264.31] Y.S. Cho, J.C. Park, B. Lee, J. Yi, Catal. Lett. 81 (2002) 89–96.

: General 475 (2014) 134– 139 139

32] F. Négrier, E. Marceau, M. Che, J.-M. Giraudon, L. Gengembre, A. Löfberg, J. Phys.Chem. B 109 (2005) 2836–2845.

33] M.G. O’Brien, S.D.M. Jacques, M. Di Michiel, P. Barnes, B.M. Weckhuysen, A.M.Beale, Chem. Sci. 3 (2012) 509–523.

34] D. Spielbauer, H. Zeilinger, H. Knözinger, Langmuir 9 (1993) 460–466.35] S. Boujday, J.-F. Lambert, M. Che, Chem. Phys. Chem. 5 (2004) 1003–1013.36] J. Lu, P. Serna, C. Aydin, N.D. Browning, B.C. Gates, J. Am. Chem. Soc. 133 (2011)

16186–16195.37] X. Carrier, E. Marceau, M. Che, Pure Appl. Chem. 78 (2006) 1039–1055.38] M. Che, O. Clause, Ch. Marcilly, Impregnation and ion exchange, in: G. Ertl, H.

Knözinger, J. Weitkamp (Eds.), Preparation of Solid Catalysts, Weinheim, 1999,pp. 315–340, cap. 4.1.1.

39] C. Arrouvel, M. Digne, M. Breysse, H. Toulhoat, P. Raybaud, J. Catal. 222 (2004)152–166.

40] G. Li, E.M. Sproviero, R.C. Snoeberger III, N. Iguchi, J.D. Blakemore, R.H. Crabtree,G.W. Brudvig, V.S. Batista, Energy Environ. Sci. 2 (2009) 230–238.

41] Y. Zhao, D.G. Truhlar, Theor. Chem. Acc. 120 (2008) 215–241.42] Y. Zhao, D.G. Truhlar, Acc. Chem. Res. 41 (2008) 157–167.43] S. Dappricha, I. Komáromia, K.S. Byuna, K. Morokuma, M.J. Frisch, J. Mol. Struct.

Teochem. 461 (1999) 1–21.44] M.J. 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, J.E. Peralta Jr., F. Ogliaro, M.Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Nor-mand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N.Rega, J.M. 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, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J.Cioslowski, D.J. Fox, Gaussian 09, Revision D.01, Gaussian, Inc., Wallingford, CT,2009.