Direct Ch4-methanol Over Fe-zsm-5 Zeolite

734

J. Phys. Chem. B 2000, 104, 734-740

Direct Methane-Methanol and Benzene-Phenol Conversions on Fe-ZSM-5 Zeolite: Theoretical Predictions on the Reaction Pathways and Energetics Kazunari Yoshizawa,* Yoshihito Shiota, Takashi Yumura, and Tokio Yamabe Department of Molecular Engineering, Kyoto UniVersity, Sakyo-ku, Kyoto 606-8501, Japan, and Institute for Fundamental Chemistry, 34-4 Takano-Nishihiraki-cho, Sakyo-ku, Kyoto 606-8103, Japan ReceiVed: June 7, 1999; In Final Form: NoVember 20, 1999

The reaction pathways and the energetics for the direct methane-methanol and benzene-phenol conversions that occur on the surface of Fe-ZSM-5 zeolite are analyzed from B3LYP DFT computations. We propose a reasonable model for “R-oxygen”, a surface oxygen species responsible for the catalytic reactivities of Fe-ZSM-5 zeolite. Our model involves an iron-oxo species on the AlO4 surface site of the zeolite as a catalytic active center and as a source of oxygen. The essential features of the reaction pathways for the methane-methanol and benzene-phenol conversions are identical, especially in bonding characters. In the initial stages of each reaction, methane or benzene comes into contact with the active iron site of the “Roxygen” model, leading to the reactant (methane or benzene) complex. After the initial complex is formed, each reaction takes place in a two-step concerted manner, via neither radical species nor ionic intermediates. The concerted reaction pathway for the methane (benzene) hydroxylation involves an H atom abstraction and a methyl (phenyl) migration at the iron active center. From computed energetics for the reaction pathways, we predict that the benzene hydroxylation should be energetically more favorable than the methane hydroxylation.

Introduction ZSM-5 zeolite exhibits an extremely high catalytic selectivity for the oxidation of benzene to phenol. Panov and collaborators1 have suggested that the high reactivity of the zeolite should be ascribed to impurity iron, which is added mainly along with the starting ingredients at the step of zeolite synthesis. Samples of ZSM-5 containing 0.07-0.72 wt % of iron calculated for Fe2O3 demonstrated excellent results; the selectivity for phenol was reported to be 100% at 20-25% conversion levels of benzene over Fe-ZSM-5 zeolite at 300-400 °C. A surface oxygen species generated on Fe-ZSM-5 zeolite under N2O decomposition, as indicated by (O) in reaction 1, has been

N2O + ( ) f (O) + N2

(1)

(O) + C6H6 f C6H5OH

(2)

proposed to be responsible for the formation of phenol from benzene (reaction 2); such an oxygen species is called “Roxygen” by Panov et al.1 According to ref 2, Solutia recently developed a one-step technology that produces phenol directly from benzene and N2O. Since this process provides a very high yield and can use waste N2O from the production of adipic acid,2 it is considered to be a hopeful technology in the coming next century. The conversion of methane to methanol has been reported to also take place over Fe-ZSM-5 zeolite using N2O as a source of oxygen. Anderson and Tsai3 reported from high-temperature experiments that the conversion rate of methane is very low and that the products include CO2, CO, and HCHO as byproducts in high concentrations. On the other hand, Panov et * To whom all correspondence should be addressed at Kyoto University. E-mail: [email protected].

al.4 showed that methanol is produced at room temperature from a single-turnover run in high yield (80 %), suggesting that the product methanol that is strongly bound to the surface of zeolite can be extracted by using a mixture of water and acetonitrile. As indicated in reaction 3, “R-oxygen” has been also proposed

(O) + CH4 f CH3OH

(3)

to be responsible for this intriguing catalytic reactivity of the zeolite toward methane. It seems to us that the methanemethanol process is difficult compared to the benzene-phenol process, judging from the limited number of papers published concerning the methane hydroxylation. Once methanol is produced, the further oxidation of methanol into formaldehyde, formic acid, or carbon oxides can take place more easily than the oxidation from methane to methanol. Thus, it may be difficult to control the further oxidation of methanol over FeZSM-5 zeolite. Schro¨der and Schwarz5-7 showed that the bare FeO+ complex reacts with methane and benzene in the gas phase, yielding methanol and phenol, respectively, in high yield. Interestingly, the FeO+ species is also generated in the gas phase from the reaction of Fe+ with nitrous oxide. Thus, an analysis of the interesting gas-phase reactions is the key to understanding the catalytic hydroxylations of methane and benzene over FeZSM-5 zeolite. Shaik et al.8 have proposed that more than one state should be involved in the reactions catalyzed by organometallic systems in remarkable contrast to organic reactions in which one state plays a dominant role; they call such a phenomenon “two-state reactivity”. From density-functional-theory (DFT) computations, we have proposed that the direct methane-methanol conversion by FeO+ should take place in the following manner: FeO+ + CH4 f OFe+(CH4) f [TS1] f HO-Fe+-CH3 f [TS2] f Fe+(CH3-

10.1021/jp991844b CCC: $19.00 © 2000 American Chemical Society Published on Web 01/05/2000

Methane-Methanol and Benzene-Phenol Conversions OH) f Fe+ + CH3OH, in which TS stands for transition state.9 The proposed reaction pathway involves the reactant complex, the hydroxy intermediate, the product complex, and the two transition states (TS1 and TS2). We call this mechanism the “two-step concerted mechanism”. The potential energy surface of the quartet state lies below that of the sextet state in the middle of the reaction pathway, and therefore the quartet potential energy surface plays an essential role in this direct methane hydroxylation. In a recent paper,10 we theoretically demonstrated that the gas-phase benzene-phenol conversion by FeO+ should take place in an identical manner with respect to essential bonding characters: FeO+ + C6H6 f OFe+(C6H6) f [TS1] f HO-Fe+-C6H5 f [TS2] f Fe+(C6H5OH) f Fe+ + C6H5OH. Taking cognizance of the experimental results mentioned above and our previous theoretical analyses,9,10 we set up a working hypothesis that “R-oxygen” should have relevance to an iron-oxo species supported on the surface of Fe-ZSM-5 zeolite and we proposed how such a reactive oxygen species is generated upon decomposition of N2O.11 Our hypothesis is reasonable if a coordinatively unsaturated iron-oxo species is responsible for the C-H bond activation of methane, benzene, and other hydrocarbons. In this article, we address how the hydroxylations of methane and of benzene take place on the surface of Fe-ZSM-5 zeolite using a possible cluster that can model “R-oxygen”. The reaction species and the energetics along the reaction pathways are analyzed in detail from DFT computations. It is useful to investigate the interesting catalytic functions of the zeolite from the viewpoint of quantum chemical calculations because of possible implications for the rational design of high-performance catalysts. Method of Calculation We optimized local minima and saddle points on potential energy surfaces using the hybrid Hartree-Fock/density-functionaltheory (HF/DFT) B3LYP method.12,13 This method consists of the Slater exchange, the Hartree-Fock exchange, the exchange functional of Becke,12 the correlation functional of Lee, Yang, and Parr (LYP),13 and the correlation functional of Vosco, Wilk, and Nusair.14 The hybrid B3LYP method has been reported to provide excellent descriptions of various reaction profiles, particularly in geometries, heats of reaction, barrier heights, and vibrational analyses.15 As indicated in a previous paper,10 the B3LYP method correctly reproduces an experimental dissociation energy of FeO+ (81.4 ( 1.4 kcal/mol).16 The method of choice is therefore appropriate for the present study. For the Fe atom we used the (14s9p5d) primitive set of Wachters,17 and for the H, C, O, Al, and Si atoms we used the 6-31G** basis set,18 a standard double-ζ basis set with polarization functions. The sextet and quartet spin states were considered for the methane hydroxylation and only the quartet state was considered for the benzene hydroxylation. The spin-unrestricted method was applied to such open-shell systems. Computed 〈S2〉 values confirmed that the spin contamination included in calculations was very small (within 0.3% after annihilation of spin contamination). Vibrational analyses were systematically performed for all transition states to confirm that they have only one imaginary mode of vibration and for all reaction intermediates along the methane-methanol conversion pathway. The Gaussian 94 program package19 was used in this study. A Possible Model of “r-Oxygen” Before discussing the reaction pathways for the hydroxylations of methane and benzene on Fe-ZSM-5 zeolite, let us first

J. Phys. Chem. B, Vol. 104, No. 4, 2000 735 CHART 1

set up a reasonable model of “R-oxygen”. To the best of our knowledge, there is no detailed structural information on transition metals supported on zeolite, particularly concerning whether such metal active centers are isolated or coupled. Kucherov and Slinkin20 proposed from ESR measurements that supported Cu, Cr, and Fe cations should be located in nonlattice (surface) positions of ZSM-5 zeolite. Analyses on the active center of Cu-ZSM-5 zeolite,21,22 which exhibits an important catalytic function of NOx decomposition,23 have been carried out using reasonable zeolite models. From semiempirical quantum-chemical calculations using dinuclear model complexes, Filatov et al.24 showed that N2O decomposition does not occur at an Al atomic center but occurs at an Fe atomic center of zeolite, suggesting that “R-oxygen” should derive from an impurity iron supported on ZSM-5 zeolite. Taking the general features of zeolite into account,25 we set up our model complex (Chart 1), in which an iron-oxo species is located on the AlO4 surface site as a catalytic active center and as a source of oxygen. In a previous paper,11 we presented from B3LYP calculations how such a reactive surface species is generated upon decomposition of nitrous oxide (N2O) on the surface of Fe-ZSM-5 zeolite and proposed the possible form of “R-oxygen”. The activation energy for the decomposition of N2O was computed to be 2.4 kcal/mol and the entire reaction 63.5 kcal/mol exothermic; we thus expect the decomposition of N2O to readily occur on the active site of Fe-ZSM-5 zeolite. The iron active center is three-coordinate in this model, but four- or even fivecoordinate iron is also possible in that the metal active center is coordinatively unsaturated. From the viewpoint of known catalytic chemistry, it is in general reasonable to assume that catalytically active metal complexes might be coordinatively unsaturated. If such an iron-oxo species is generated on the surface of the zeolite, it should become a reaction center that affords a coordination site for methane, benzene, and other hydrocarbon molecules. The oxo species is, of course, a source of oxygen in the oxidation of these hydrocarbons. Since the Fe atom is +3 in formal charge, the oxo species -2, the Al atom +3, the OSiH3 group -1, and the OH group -1, the total charge of the model complex is counted as neutral. We considered the sextet and quartet spin states for our computational analyses because of the detailed analyses of the gas-phase reactions by the bare FeO+ complex9,10 discussed above. A computed structure of our model of “R-oxygen” in the quartet state is shown in Chart 1, in which bond distances indicated are in angstroms and bond angles (italic) are in degrees. Atomic charges and spin densities in the sextet and quartet states are listed in Table 1. Computed spin densities reside mainly on the iron-oxo moiety, as expected. In the following section, we will look in detail at the reactions of methane and benzene on “Roxygen” of Fe-ZSM-5 zeolite using this simple, but rather reasonable model.

736 J. Phys. Chem. B, Vol. 104, No. 4, 2000

Yoshizawa et al.

TABLE 1: Computed Atomic Charges and Spin Densities on a Model of “r-oxygen” of Fe-ZSM-5 Zeolite quartet

sextet

atom

atomic charge

spin density

atomic charge

spin density

Fe O(oxo) O Si Al

+1.30 -0.63 -0.85 +0.40 +0.91

+3.15 -0.22 +0.03 +0.01 -0.01

+1.38 -0.67 -0.86 +0.39 +0.91

+3.97 +0.93 +0.05 +0.01 -0.01

Methane Hydroxylation on Fe-ZSM-5 Zeolite Structures of the Reaction Species. Having set up our model of “R-oxygen”, let us next consider the methane hydroxylation on it. In this section we look at computed geometries of the reaction species in the quartet spin state for the conversion of methane to methanol on the model complex. The reaction pathway that we address here is essentially identical to that of the gas-phase methane-methanol conversion by FeO+.9 Since the low-spin quartet potential energy surface provides a lowcost energy pathway, as described later, we take a look at the reaction species on the quartet reaction pathway. Optimized structures of the reaction species in the sextet state are similar to those in the quartet state. The first step of this reaction is C-H bond activation of methane. As shown in Figure 1, a methane (or reactant) complex is formed in the initial stages of the reaction. The methane molecule is weakly bound to the three-coordinate iron of the model complex; computed binding energies of 1.6 and 3.9 kcal/mol in the quartet and sextet states, respectively, are small, but we think that the formation of this complex is very much indicative of methane activation. We predict from DFT computations that the Fe ion is closer to one pair of H atoms of the bound methane; this initially formed species is an η2-CH4 complex. Since methane is bound to the neutral complex, the driving force for this binding is unlikely to be direct electrostatic interactions between the methane molecule and the complex. Methane complexes with an η2- or η3-CH4 binding mode have been proposed to be involved in its C-H bond activation; extensive indirect data have been obtained to support the intermediary of small-alkane complexes.26 One such example was obtained by Billups et al.26i The neutral Co(CH4) complex is formed during photolysis of the CH3CoH complex in an argon matrix, which was confirmed from the observation that the triply degenerate T2 vibrational mode of methane at 1305 cm-1 was observed from FTIR measurements to split into two peaks in the Co(CH4) complex. After this methane complex is formed, one of the H atoms of the methane molecule is abstracted in a concerted fashion via a four-centered transition state (TS1), leading to a reaction intermediate that involves the resultant OH and CH3 groups as ligands. We call this species a hydroxy intermediate, which plays a central role in the methane-methanol conversion. This process can be viewed as a [1,3] migration of the H atom; the dissociating C-H bond and the forming O-H bond were predicted to be 1.309 and 1.357 Å, respectively, the values being appropriate for this concerted electronic process. This fourcentered structure is quite similar to the corresponding transitionstate structure in the gas-phase process by the bare FeO+ complex. This mechanism for the C-H dissociation of hydrocarbon substrates is strikingly different from the conventional radical mechanism that is considered to take place through a transition state with a linear C-H-O array. The second half of the reaction starts from the hydroxy intermediate. The hydroxy intermediate thus formed is then converted into a methanol (or product) complex in a concerted

manner via a three-centered transition state (TS2) in which an Fe-C bond breaking and a C-O bond formation occur simultaneously, as indicated in Figure 1. This transition state correctly connects the hydroxy intermediate and the product complex, being responsible for an intramolecular [1,2] methyl migration. Optimized Fe-C and O-C distances are 2.543 and 1.852 Å, respectively. The three-centered structure of TS2 is thus appropriate for the transition state of the concerted methyl migration from the Fe ion to the OH group. Let us finally look at the product complex that involves methanol as a ligand. The basic structure of the methanol moiety in the product complex is not significantly changed from that of free methanol. The Fe-O distance of 1.989 Å is much longer than 1.608 Å in the reactant complex and 1.765 Å in the hydroxy intermediate. The Fe-O bond of the product complex is thus viewed as a typical coordinate bond in contrast to those of the reactant complex and the hydroxy intermediate. The product complex and the final complex from which methanol is released involve Fe(I) that needs to be reoxidized back to Fe(III) upon decomposition of N2O. In a previous paper,11 we studied from DFT computations the energetics for the conversion of [Fe]+ f [FeO]+ by N2O that regenerates the active center for further reaction. Fe(I) is in general a rather unstable oxidation state for iron, but according to a textbook,27 the existence of Fe(I) cannot be ruled out. Panov et al.4b proposed a diiron model that reasonably allows two-electron oxidation-reduction between Fe(III)/Fe(III) and Fe(II)/Fe(II). Our mechanism is, of course, extended to diiron complexes, as can be seen in our previous papers on the mechanism of methane monooxygenase.28 We have described the essential electronic processes for the methane hydroxylation on a model of “R-oxygen”. The most important point of our proposal for the reaction pathway is that methane is activated on a coordinatively unsaturated iron-oxo species leading to a methane complex and that two-step concerted migrations of hydrogen and methyl successfully convert methane to methanol on the complex. Finally, we should indicate that the structure of the four-sided figure of FeO2Al remains almost unchanged throughout the reaction and that the catalytically active center can be reproduced by the introduction of N2O gas. Energetics for the Hydroxylation of Methane. We present in Figure 2 computed energy diagrams in the sextet and quartet spin states for the methane hydroxylation on the possible model of “R-oxygen” of Fe-ZSM-5 zeolite. The overall reaction is 23.9 kcal/mol endothermic on the quartet surface and 26.0 kcal/ mol endothermic on the sextet surface. The essential features of the energy diagrams are similar to those of the gas-phase process by FeO+.9 We see crossing between the two potential energy diagrams to occur twice in the vicinity of the reactant and the product complexes. Spin inversion can take place in the vicinity of these crossing points. From an inspection of the two energy diagrams, we expect that the quartet potential surface should afford an energetically low-cost reaction pathway whereas the sextet state is the ground state in the entrance and the exit channels. Although the crossing occurs twice near the reactant and the product complexes, the spin inversion has little effect on the energetics of the reaction pathway because the sextet and the quartet states in the entrance and the exit channels are close-lying in energy. We therefore conclude that the quartet potential energy surface should play a dominant role in this reaction, which is somewhat different from that of the reaction pathway for the gas-phase methane-methanol conversion by FeO+,9 where such spin inversions play a very important role in determining the energetics of the reaction pathway.8a

Methane-Methanol and Benzene-Phenol Conversions

J. Phys. Chem. B, Vol. 104, No. 4, 2000 737

Figure 1. Ball-and-stick structures of the reaction intermediates and the transition states for the direct methane hydroxylation on a model of “R-oxygen” of Fe-ZSM-5 zeolite in the quartet state. Bond lengths are in angstroms.

Figure 2. Potential energy diagrams for the direct methane hydroxylation on a model of “R-oxygen” of Fe-ZSM-5 zeolite in the sextet and the quartet states. Note that the reaction proceeds in a two-step manner, via TS1 and TS2. Relative energies, which include zero-point energy corrections, are in kcal/mol. SI means spin inversion.

Let us look in detail at the energetics of the catalytic methane hydroxylation on the model complex. As mentioned above, we can draw general conclusions on the energetics in view of the quartet potential energy surface. The methane molecule is bound to the three-coordinate iron of the model complex with a small binding energy of 1.6 kcal/mol. The activation barrier for the C-H bond cleavage via TS1 was predicted to be 11.4 kcal/ mol if measured from the reactant complex and 9.8 kcal/mol

from the dissociation limit on the quartet potential energy surface. These values are small compared to the C-H dissociation energy of methane (104 kcal/mol). It is interesting to compare the activation energy for the catalytic H atom abstraction on the “R-oxygen” model with those in the gas-phase process by the bare FeO+ complex.9 In the gas-phase process, the corresponding barrier height for TS1 was computed to be 15.7 kcal/mol from the reactant complex and 6.5 kcal/mol from the dissociation limit on the quartet potential energy surface.9b Thus, the energetics for the C-H bond cleavage of methane is comparable in both processes because of the similar fourcentered structures of TS1. The resultant hydroxy intermediate is extremely stable in energy in both spin multiplicities. Therefore, this intermediate might be detected at low temperature using modern spectroscopic techniques. The barrier height for TS2 was predicted to be 41.6 kcal/ mol on the quartet potential energy surface if measured from the hydroxy intermediate, which is extremely stable in energy. Therefore, once the hydroxy intermediate is formed, it seems to us that this activation energy would require very high temperatures for the reaction to be facile. This view is correct if we confine our discussions to the chemical reactions on the computed adiabatic potential. However, we think that either kinetic energy that the reaction system possesses or energy transfer from the external field can overcome this activation barrier. In fact, if we measure the activation barrier for TS2 from the dissociation limit, it becomes 13.2 kcal/mol. Considering this potential energy profile of the reaction system, this electronic process can be thermally accessible. If the reaction system preferentially moves on the quartet potential energy

738 J. Phys. Chem. B, Vol. 104, No. 4, 2000

Yoshizawa et al.

Figure 3. Ball-and-stick structures of the reaction intermediates and the transition states for the direct benzene hydroxylation on a model of “R-oxygen” of Fe-ZSM-5 zeolite. Bond lengths are in angstroms.

surface, 26.6 kcal/mol is required for the release of the product methanol. Even if we consider the spin inversion to take place in the exit channel, this value would not change because the sextet and the quartet states of the product complex, indicated at right in Figure 2, are close-lying in energy. Thus, the quartet potential energy surface plays a dominant role in this reaction. Benzene Hydroxylation on Fe-ZSM-5 Zeolite Structures of the Reaction Species. Economical, direct conversion of benzene to phenol over Fe-ZSM-5 zeolite has attracted the attention of researchers in applied chemistry,2 but little is known of the reaction mechanism by which benzene is oxidized to phenol. We believe that our theoretical study on the benzene hydroxylation on Fe-ZSM-5 zeolite will be of significant use in the interpretation of the catalytic processes and in the rational design of high-performance zeolite catalysts. In a previous paper,10 we addressed the reaction pathway and the energetics for the gas-phase benzene-phenol conversion by the bare FeO+ complex. We can apply our previous results to the catalytic processes on Fe-ZSM-5 zeolite. Figure 3 presents ball-and-stick structures for the reaction intermediates and the transition states along the reaction pathway for the benzene hydroxylation on the “R-oxygen” model. The reaction mechanism is identical to that for the methane hydroxylation discussed earlier in essential bonding characters. Our discussions can be reasonably restricted to the reaction pathway in the quartet state because the quartet potential energy surface is low-lying compared to the sextet state and thus the

reaction system should move preferentially on the quartet potential energy surface judging from the previous study on the corresponding gas-phase process.10 The initially formed complex exhibits an η2-C6H6 mode in which the Fe ion is closer to one pair of carbon atoms of the benzene ring. The benzene ring remains nearly planar, so the complex is a kind of π complex. The binding energy for the reactant complex was predicted to be 10.2 kcal/mol. After this complex is formed, an H atom abstraction occurs via TS1 that connects the reactant complex and the hydroxy intermediate. The C-H bond and the O-H bond in the four-centered structure were computed to be 1.298 and 1.364 Å, respectively. These values are reasonable for a transition-state structure responsible for C-H bond breaking and O-H bond forming. The C-H bond is significantly deviated from the benzene plane, and the four-centered structure is nearly orthogonal to the benzene plane. Thus, the first half of the reaction ends with the formation of the intermediate complex that involves the resultant OH and C6H5 groups as ligands. We next take a look at the hydroxy intermediate. This intermediate is likely to play a central role in the reaction pathway for the benzene-phenol conversion, by analogy with the reaction pathway for the methane-methanol conversion. The FeOH moiety is nearly coplanar with the benzene ring. TS2 is a transition state in which Fe-C bond breaking and C-O bond formation occur in a concerted manner to connect the hydroxy intermediate and the product complex that involves phenol as a ligand. This three-centered transition state is responsible for

Methane-Methanol and Benzene-Phenol Conversions

J. Phys. Chem. B, Vol. 104, No. 4, 2000 739 computed to be 20.2 kcal/mol, which is smaller than 26.6 kcal/ mol in the methanol desorption; see Figure 2. Thus, the product methanol is more strongly bound to the active iron site of FeZSM-5 zeolite than the product phenol. This computational result is in good agreement with the statement of Panov et al.4 that the product methanol is strongly bound to the surface of zeolite and can be extracted by using a mixture of water and acetonitrile. Conclusions

Figure 4. Potential energy diagram for the direct benzene hydroxylation on a model of “R-oxygen” of Fe-ZSM-5 zeolite in the quartet state. Relative energies are in kcal/mol.

the process for a phenyl migration from the iron active center to the oxygen atom. The breaking Fe-C bond and the forming C-O bond were computed reasonably well to be 1.999 and 1.730 Å, respectively. The three-centered (Fe,O,C) structure responsible for the phenyl migration is again nearly orthogonal with the benzene ring. Let us finally look at the product complex. Note that the basic structure of the phenol moiety in the product complex is not significantly changed from that of free phenol. We thus conclude that the direct conversion of benzene to phenol should proceed in a concerted manner, via neither radical species nor ionic intermediates. Our computational results are, generally, in good agreement with the proposals of Panov et al.1 Burch and Howitt29 proposed that the mechanism for the selective oxidation of benzene involves a radical cation as an intermediate, but our proposal is different from theirs. Energetics for the Hydroxylation of Benzene. We show in Figure 4 a computed energy diagram for the benzene hydroxylation on the model of “R-oxygen”. Because the overall reaction is 8.6 kcal/mol endothermic on the quartet surface, we predict that the benzene-phenol conversion should be energetically more favorable than the methane-methanol conversion (23.9 kcal/mol endothermic). This is likely in good agreement with the experimental results of Panov et al.1,4 In the initial stages of the reaction, benzene comes into contact with the iron active site of the “R-oxygen” model to form the reactant complex. The binding energy for the reactant complex was predicted to be 10.2 kcal/mol, which is clearly larger than the value in the methane case (1.6 kcal/mol), due to the higher reactivity of the frontier π orbitals of benzene. After the reactant complex is formed, one of the H atoms of the benzene molecule is abstracted in a concerted manner via TS1. The barrier height for this electronic process was computed to be 15.4 kcal/mol from the reactant complex and 5.2 kcal/mol from the dissociation limit on the quartet surface. We think that this transition state should be thermally accessible. The hydroxy intermediate thus formed is also stable in energy. Therefore, the barrier height for TS2 was predicted to be very large (31.1 kcal/mol) if measured from the hydroxy intermediate. However, from a close inspection of the energy diagram, the height of TS2 is found to be comparable to the dissociation limit. If this reaction is kinetically controlled, the energy difference between the transition state and the dissociation limit is a good measure in considering whether the reaction takes place or not. We therefore think that this transition state is also thermally accessible to lead to the product complex. The energy required for the desorption of phenol from the complex was

The reaction pathways and the energetics for the direct hydroxylation of methane and of benzene that occur on the surface of Fe-ZSM-5 zeolite have been computed and analyzed using the B3LYP DFT method. We proposed a reasonable model for “R-oxygen”, a surface oxygen species that is responsible for the catalytic reactivities of Fe-ZSM-5 zeolite; our model involves an iron-oxo species on the AlO4 surface site as an active center. The salient features of the two reaction pathways proposed are identical in essential bonding characters. In the initial stages of each reaction, methane or benzene comes into contact with the active iron site of the “R-oxygen” model leading to the reactant complex. H atom abstraction from methane or benzene occurs via a four-centered transition state (TS1), resulting in the formation of the hydroxy intermediate. The hydroxy intermediate is next transformed into the product complex via a three-centered transition state (TS2). These electronic processes take place in a two-step concerted manner with a stable reaction intermediate. Our proposals for the zeolitecatalyzed reactions are essentially identical to those for the gasphase methane-methanol and benzene-phenol conversions by the bare iron-oxo complex.9,10 The overall reaction for the methane hydroxylation is 23.9 kcal/mol endothermic on the quartet surface and 26.0 kcal/mol endothermic on the sextet surface, and that for the benzene hydroxylation is 8.6 kcal/mol endothermic on the quartet surface. Thus, we predict that the benzene-phenol conversion should be energetically more favorable than the methane-methanol conversion. We think that the “two-step concerted mechanism” should be widely applicable in interpreting how hydrocarbon hydroxylations are catalyzed by transition-metal oxides. The reaction pathways proposed in this article may have relevance to certain enzymatic processes concerning hydrocarbon hydroxylations, in which various iron-oxo complexes play a central role.30,31 Our mechanistic proposals are completely consistent in the gasphase,9,10 enzymatic,28 and zeolite-catalyzed reactions. Acknowledgment. We are grateful to the “Research for the Future” Program of the Japan Society for the Promotion of Science (JSPS-RFTF96P00206) and to a Grant-in-Aid for Scientific Research on the Priority Area “Molecular Biometallics” from the Ministry of Education, Science, Sports, and Culture of Japan for their support of this work. Y.S. is grateful to the JSPS for a graduate fellowship. Computations were partly carried out at the Supercomputer Laboratory of Kyoto University and at the Computer Center of the Institute for Molecular Science. References and Notes (1) (a) Panov, G. I.; Sobolev, V. I.; Kharitonov, A. S. J. Mol. Catal. 1990, 61, 85. (b) Panov, G. I.; Sheveleva, G. A.; Kharitonov, A. S.; Romannikov, V. N.; Vostrikova, L. A. Appl. Catal. A: General 1992, 82, 31. (c) Sobolev, V. I.; Kharitonov, A. S.; Paukshtis, Y. A.; Panov, G. I. J. Mol. Catal. 1993, 84, 117. (d) Sobolev, V. I.; Panov, G. I.; Kharitonov, A. S.; Romannikov, V. N.; Volodin, A. M.; Ione, K. G. J. Catal. 1993, 139, 435. (e) Panov, G. I.; Kharitonov, A. S.; Sobolev, V. I. Appl. Catal. A:

740 J. Phys. Chem. B, Vol. 104, No. 4, 2000 General 1993, 98, 1. (f) Volodin, A. M.; Bolshov, V. A.; Panov, G. I. J. Phys. Chem. 1994, 98, 7548. (2) Chem. Eng. News 1998, April 6, 21. (3) Anderson, J. R.; Tsai, P. J. Chem. Soc., Chem. Commun. 1987, 1435. (4) (a) Sobolev, V. I.; Dubkov, K. A.; Panna, O. V.; Panov, G. I. Catal. Today 1995, 24, 251. (b) Dubkov, K. A.; Sobolev, V. I.; Talsi, E. P.; Rodkin, M. A.; Watkins, N. H.; Shteinman, A. A.; Panov, G. I. J. Mol. Catal. A.: Chemical 1997, 123, 155. (5) (a) Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1990, 29, 1433. (b) Schro¨der, D.; Fiedler, A.; Hrusa´k, J.; Schwarz, H. J. Am. Chem. Soc. 1992, 114, 1215. (c) Schro¨der, D.; Schwarz, H.; Clemmer, D. E.; Chen, Y.-M.; Armentrout, P. B.; Baranov, V. I.; Bo¨hme, D. K. Int. J. Mass Spectrom. Ion Processes 1997, 161, 175. (6) (a) Schro¨der, D.; Schwarz, H. HelV. Chim. Acta 1992, 75, 1281. (b) Becker, H.; Schro¨der, D.; Zummack, W.; Schwarz, H. J. Am. Chem. Soc. 1994, 116, 1096. (c) Ryan, M. F.; Sto¨ckigt, D.; Schwarz, H. J. Am. Chem. Soc. 1994, 116, 9565. (7) A review on catalytic reactivities of the first-row transition-metal oxide ions (MO+): Schro¨der, D.; Schwarz, H. Angew. Chem., Int. Ed. Engl. 1995, 34, 1973. (8) (a) Shaik, S.; Danovich, D.; Fiedler, A.; Schro¨der, D.; Schwarz, H. HelV. Chim. Acta 1995, 78, 1393. (b) Shaik, S.; Filatov, M.; Schro¨der, D.; Schwarz, H. Chem. Eur. J. 1998, 4, 193. (9) (a) Yoshizawa, K.; Shiota, Y.; Yamabe, T. Chem. Eur. J. 1997, 3, 1160. (b) Yoshizawa, K.; Shiota, Y.; Yamabe, T. J. Am. Chem. Soc. 1998, 120, 564. (c) Yoshizawa, K.; Shiota, Y.; Yamabe, T. Organometallics 1998, 17, 2825. (10) Yoshizawa, K.; Shiota, Y.; Yamabe, T. J. Am. Chem. Soc. 1999, 121, 147. (11) Yoshizawa, K.; Yumura, T.; Shiota, Y.; Yamabe, T. Bull. Chem. Soc. Jpn 2000, 73, January issue. (12) (a) Becke, A. D. Phys. ReV. 1988, A38, 3098. (b) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (13) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. 1988, B37, 785. (14) Vosco, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200. (15) Baker, J.; Muir, M.; Andzelm, J.; Scheiner, A. In Chemical Applications of Density-Functional Theory; Laird, B. B., Ross, R. B., Ziegler, T., Eds.; ACS Symposium Series 629; American Chemical Society: Washington, DC, 1996. (16) Loh, S. K.; Fisher, E. R.; Lian, L.; Schultz, R. H.; Armentrout, P. B. J. Phys. Chem. 1989, 93, 2601. (17) Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033. (18) (a) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972, 56, 2257. (b) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28, 213. (19) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T. A.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. J. P.; Head-

Yoshizawa et al. Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94; Gaussian Inc.: Pittsburgh, PA, 1995. (20) Kucherov, A. V.; Slinkin, A. A. J. Phys. Chem. 1989, 93, 864. (21) (a) Gru¨nerrt, W.; Hayes, N. W.; Joyner, R.; Shpiro, E. S.; Siddiqui, M. R. H.; Baeva, G. N. J. Phys. Chem. 1994, 98, 10832. (b) Yamashita, H.; Matsuoka, M.; Tsuji, K.; Shioya, Y.; Anpo, M.; Che, M. J. Phys. Chem. 1996, 100, 397. (22) (a) Trout, B. L.; Chakraborty, A. K.; Bell, A. T. J. Phys. Chem. 1996, 100, 4173, 17582. (b) Schneider, W. F.; Haas, K. C.; Ramprasad, R.; Adams, J. B. J. Phys. Chem. 1996, 100, 6032; Haas, K. C.; Schneider, W. F. J. Phys. Chem. 1996, 100, 9292. (c) Rodriguez-Santiago, L.; Sierka, M.; Branchadell, V.; Sodupe, M.; Sauer, J. J. Am. Chem. Soc. 1998, 120, 1545. (23) (a) Iwamoto, M.; Yashiro, H. Catal. Today 1994, 22, 5. (b) Lei, G. D.; Adelman, B. J.; Sarkany, J.; Sachtler, W. M. H. Appl. Catal. B 1995, 5, 245. (c) Spoto, G.; Zecchina, A.; Bordiga, S.; Ricchiardi, G.; Martra, G.; Leofanti, G.; Petrini, G. Appl. Catal. B: EnVironmental 1994, 3, 151. (d) A review: Shelef, M. Chem. ReV. 1995, 95, 209. (24) Filatov, M. J.; Pelmenschikov, A. G.; Zhidomirov, G. M. J. Mol. Catal. 1993, 80, 243. (25) See for example: (a) Sauer, J. Chem. ReV. 1989, 89, 199. (b) van Santen, R. A.; Kramer, G. J. Chem. ReV. 1995, 95, 637. (26) (a) Trevor, D. J.; Cox, D. M.; Kaldor, A. J. Am. Chem. Soc. 1990, 112, 3742. (b) Irikura, K. K.; Beauchamp, J. L. J. Phys. Chem. 1991, 95, 8344. (c) Ranasinghe, Y. A.; MacMahon, T. J.; Freiser, B. S. J. Phys. Chem. 1991, 95, 7721. (d) Van Zee, R. J.; Li, S.; Weltner, W., Jr. J. Am. Chem. Soc. 1993, 115, 2976. (e) Carrol, J. J.; Weisshaar, J. C. J. Am. Chem. Soc. 1993, 115, 800. (f) Burger, P.; Bergman, R. G. J. Am. Chem. Soc. 1993, 115, 10462. (g) Schaller, C. P.; Bonanno, J. B.; Wolczanski, P. T. J. Am. Chem. Soc. 1994, 116, 4133. (h) Zhang, X.-X.; Wayland, B. B. J. Am. Chem. Soc. 1994, 116, 7897. (i) Billups, W. E.; Chang, S.-C.; Hauge, R. H.; Margrave, J. L. J. Am. Chem. Soc. 1995, 117, 1387. (j) van Koppen, P. A. M.; Kemper, P. R.; Bushnell, J. E.; Bowers, M. T. J. Am. Chem. Soc. 1995, 117, 2098. (k) Campbell, M. L. J. Am. Chem. Soc. 1997, 119, 5984. (27) Cotton, F. A.; Wikinson, G. AdVanced Inorganic Chemistry, 4th ed.; Wiley: New York, 1987. (28) (a) Yoshizawa, K. J. Biol. Inorg. Chem. 1998, 3, 318. (b) Yoshizawa, K.; Yamabe, T.; Hoffmann, R. New J. Chem. 1997, 21, 151. (c) Yoshizawa, K.; Ohta, T.; Yamabe, T.; Hoffmann, R. J. Am. Chem. Soc. 1997, 119, 12311. (d) Yoshizawa, K.; Ohta, T.; Yamabe, T. Bull. Chem. Soc. Jpn. 1998, 71, 1899. (29) (a) Burch, R.; Howitt, C. Appl. Catal. A: General 1992, 86, 139. (b) Burch, R.; Howitt, C. Appl. Catal. A: General 1993, 103, 135. (c) Burch, R.; Howitt, C. Appl. Catal. A: General 1993, 106, 167. (30) For cytochrome P450, see: (a) Cytochrome P450: Structure, Mechanism, and Biochemistry, 2nd ed.; Ortiz de Montellano, P. R., Ed.; Plenum: New York, 1995. (b) Sono, M.; Roach, M. P.; Coulter, E. D.; Dawson, J. H. Chem. ReV. 1996, 96, 2841. (31) Reviews on methane monooxygenase: (a) Kurtz, D. M., Jr. J. Biol. Inorg. Chem. 1997, 2, 159. (b) Wallar, B. J.; Lipscomb, J. D. Chem. ReV. 1996, 96, 2625. (c) Que, L., Jr.; Dong, Y. Acc. Chem. Res. 1996, 29, 190. (d) Feig, A. L.; Lippard, S. J. Chem. ReV. 1994, 94, 759. (e) Lipscomb, J. D. Annu. ReV. Microbiol. 1994, 48, 371.