DOI: 10.1002/chem.201403022
Full Paper
& Reaction Mechanisms
Pd-Catalysed Mono- and Dicarbonylation of Aryl Iodides: Insights into the Mechanism and the Selectivity Victor M. Fernndez-Alvarez ,[a, b] Vernica de la Fuente,[a] Cyril Godard,[a] Sergio Castilln,[c] Carmen Claver,[a] Feliu Maseras ,*[b, d] and Jorge J. Carb*[a]
Abstract: The mechanism of the experimentally reported phosphine-free palladium-catalysed carbonylation of aryl iodides with amines in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as base was investigated at the DFT level. Paths were identified for both di- and monocarbonylation, and the calculated selectivity for three different substrates was in agreement with experiment. In dicarbonylation yielding a-ketoamides, formation of the second carbon–carbon bond occurs through reductive elimination in the Pd acyl amide intermediate after DBU-assisted nucleophilic attack of an amine at a terminal CO ligand. This path yields the major product with iodobenzene and the almost
exclusive product with p-methoxyiodobenzene. Two different possible pathways yield the monocarbonylated amide product. In one of them, which affords the minor product for iodobenzene, base-assisted nucleophilic attack of the amine takes place on a Pd-bound acyl ligand. For substrates with electron-withdrawing substituents, such as p-cyanoiodobenzene, aryl migration to the CO ligand is disfavoured, and this allows base-assisted amine attack at a terminal CO ligand early in the catalytic cycle. From the resulting Pd amide aryl complex, the subsequent reductive elimination occurs easily, and monocarbonylation becomes favoured.
Introduction The carbonylation of organic substrates mediated by transition metal complexes is a convenient method for forming carbon– carbon bonds and incorporating carbonyl functional groups.[1] A fascinating synthetic possibility is the sequential incorporation of two molecules of carbon monoxide, which can lead to the synthesis of valuable products.[2] In particular, the Pd-catalysed dicarbonylation of aryl halides provides a direct method for the production of a-ketoamides (Scheme 1, left),[2, 3] which [a] V. M. Fernndez-Alvarez , V. de la Fuente, Dr. C. Godard, Prof. Dr. C. Claver, Dr. J. J. Carb Departament de Qumica Fsica i Inorgnica Universitat Rovira i Virgili C/Marcel.li Domingo s/n, 43007 Tarragona (Spain) Fax: (+ 34) 977-559-563 E-mail: [email protected] [b] V. M. Fernndez-Alvarez , Prof. Dr. F. Maseras Institute of Chemical Research of Catalonia (ICIQ) Av Paı¨sos Catalans 16, 43007 Tarragona, Catalonia (Spain) Fax: (+ 34) 977-920-231 E-mail: [email protected] [c] Prof. Dr. S. Castilln Departament de Qumica Analtica i Orgnica Universitat Rovira i Virgili C/Marcel.li Domingo s/n, 43007 Tarragona (Spain) [d] Prof. Dr. F. Maseras Departament de Qumica Universitat Autnoma de Barcelona 08193 Bellaterra, Catalonia (Spain) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.201403022. Chem. Eur. J. 2014, 20, 10982 – 10989
Scheme 1. Competing reactions involved in the carbonylation process.
are precursors of biologically active compounds and several families of heterocycles.[4] Most of the reported Pd catalysts are modified with phosphine ligands,[5, 6, 7] which limit their application due to inherent problems with oxidation. One of the most efficient phosphine-based methods for dicarbonylation was discovered by Kondo and co-workers,[7] who used a Pd/PtBu3 catalyst in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as base. This reaction produced a-ketoamides selectively without the need for high CO pressures. Recently, some of us reported the first efficient phosphine-free Pd catalyst system for double carbonylation of aryl iodides.[8] In the presence of DBU, the reaction produced excellent conversions and selectivities for a wide range of aryl iodides and amine nucleophiles under atmospheric CO pressure. The phosphine-free protocol was then extended to supported catalysis by Papp and Skoda-Fçldes.[9] Another alternative to phosphines is the use of N-heterocyclic carbenes, which catalysed the dicarbonylation of aryl iodides under mild conditions.[10] The reaction mechanism of Pd/phosphine-catalysed monoand dicarbonylation has been experimentally studied.[11] In the dicarbonylation process the initial steps of the reaction are common to other Pd-mediated carbonylation reactions involv-
10982
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper ing an oxidative addition of the aryl iodide to Pd0 species followed by CO insertion into the Pd–aryl bond. The process proceeds by coordination of CO, nucleophilic attack of the amine at the terminal CO ligand to form a Pd acyl carbamoyl species and reductive elimination to give the a-ketoamide product. The mechanism involving insertion of CO into the Pd acyl species has also been considered.[10, 11a] Dicarbonylation takes place under conditions very similar to those used for monocarbonylation, and the two processes can compete (Scheme 1). Indeed, this procedure cannot be used for aryl iodides with electron-withdrawing substituents, because the selectivity shifts towards the monocarbonylation products.[5d, 8] It seems also clear from previous reports that the nature of the base is crucial in these processes, and that a large excess of base is always required for dicarbonylation.[5c,d, 8] The mechanism of monocarbonylation is controversial, and a different mechanism has been proposed when alcohols are used as nucleophiles instead of amines.[11, 12] A better understanding of the molecular basis of the chemoselectivity appears to be the key to the development of more efficient and selective processes. As a part of our ongoing projects to understand theoretically the role of Pd complexes in CC bond-forming reactions and other aspects of Pd chemistry,[13, 14] we report herein a detailed mechanistic study of the di- and monocarbonylation of different aryl iodides with amines catalysed by the Pd/DBU system. In the phosphine-free palladium-catalysed dicarbonylation process, DBU is fundamental, since it plays the role of ligand, base and eventually nucleophile.[8] Our goals were to characterise the catalytic cycle in which DBU is the ligand, to determine the role of the base and to identify the factors that favour the incorporation of two CO molecules (dicarbonylation) instead of one CO molecule (monocarbonylation) into an organic substrate. Herein, we focus on the reaction between PdL2 (L = DBU, CO) complexes, NH3 and iodobenzene derivatives. We studied three different substrates: iodobenzene, pmethoxyiodobenzene and p-cyanoiodobenzene. This set of substrates is particularly intriguing, because their dicarbonylation:monocarbonylation ratios are 76:24, 98:2 and 1:99, respectively.[8]
Results and Discussion Initially, we analysed the dicarbonylation mechanism shown in Scheme 2, which was previously proposed and based on experimental findings.[8] The mechanism can be divided into several steps: 1) oxidative addition of iodobenzene to 1 a to yield 2 a, 2) migration of the phenyl group to the terminal CO ligand to form Pd acyl complex 3 a, 3) substitution of a DBU ligand by a second equivalent of CO to produce 4 a, 4) nucleophilic attack of ammonia on coordinated CO assisted by a base to yield 5 a and 5) reductive elimination to form the a-ketoamide and restore the catalyst 1 a. Figures 1 and 2 show the structures of the key species, and Figure 3 depicts the overall freeenergy profile of the reaction. Then, we determined the mechanism for monocarbonylation in order to identify the factors governing the selectivity of the process. Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
Scheme 2. Studied mechanism for the dicarbonylation of aryl iodide by the Pd/DBU system with ammonia as nucleophile and DBU as base.
The first carbon–carbon bond-formation step Formation of the spectroscopically characterised Pd acyl complex [PdI(DBU)2{C(O)aryl}] (3 a) involves oxidative addition of aryl iodide to Pd0 and Ph migration to the terminal CO ligand. Both elementary reactions have been studied by computational methods.[15, 16] Therefore, these steps were not analysed in detail, but the eventual consequences of using DBU ligands are discussed. Because the reaction medium contains a small excess of both CO and DBU, several Pd0 species could coexist: [Pd(DBU)2] (1 a’), [Pd(DBU)(CO)] (1 a) and [Pd(CO)2] (1 a’’). Their relative free energies in toluene are + 25.3, 0.0 and + 5.4 kcal mol1, respectively. Thus, we selected the most stable Pd0 species 1 a as the starting point for the catalytic reaction. First, the active species 1 a reacts with iodobenzene to form the cis oxidative-addition product [Pd(DBU)(CO)(Ph)I] (2 a). As previously observed, the CI cleavage leads to a cis PdII complex.[17] The calculated transition state for oxidative addition of aryl iodide lies 21.6 kcal mol1 above the reactants, and the process is slightly endothermic by 0.9 kcal mol1. For the oxidative addition of PhI to [Pd(PH3)2], the calculated free energy barrier in CH2Cl2 solvent at a similar level of theory (B3LYP/PCM) is somewhat lower (17.0 kcal mol1),[15b] that is, introduction of p-acidic ligands such as CO somewhat disfavours the CI bond activation. Migration of the phenyl group to the carbonyl ligand and coordination of an additional DBU ligand forms cis-[Pd(DBU)2{C(O)Ph}I] (3 a), which has remarkable thermodynamic stability (14.4 kcal mol1 below the reactants). The corresponding trans isomer is almost isoenergetic (0.7 kcal mol1 above), because the two large DBU ligands in cis positions adopt an alternate conformation to avoid mutual steric hindrance (see Figure 1). This result is in agreement with the equilibrium between cis and trans species observed by NMR spectroscopy.[8] For phenyl migration we considered a three-centre mechanism[16b] and located transition state TS2 a-3 a with an energy
10983
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Scheme 3. Possible CO ligand exchange in complex 3 a. Free energies in kcal mol1.
Figure 1. Selected molecular structures and main geometric parameters of the first part of the cycle: the transition states for Ph migration to terminal CO (TS2 a-3 a) and for replacement of DBU by free CO (TS3 a-4 a) and the intermediate cis-[Pd(DBU)2{C(O)Ph}I] 3 a. Hydrogen atoms omitted for clarity. Distances in ngstrçm.
barrier of 11.1 kcal mol1 and the phenyl group migrating in the plane of the complex (see Figure 1). Both the energy and the geometry closely resemble those recently calculated for trans-[PdPh(CO)(OEt)(PMe3)] by Lei et al.[16a] The intermediate 3 a is not only thermodynamically stable (DGTol = 14.4 kcal mol1), but its reverse energy barrier is also quite high (25.4 kcal mol1). Thus, the CO migratory insertion step can be classified as an irreversible step in the catalytic cycle. In summary, calculations show that the formation of Pd acyl species by oxidative addition and subsequent CO migratory insertion is not energetically demanding and leads to a thermodynamically stable complex. This result is in full agreement with previous experimental findings.[8] The second carbon–carbon bond-formation step: dicarbonylation Using the spectroscopically characterised Pd acyl complex as a reliable stepping stone, we analysed the mechanism of a-ketoamide formation from complex 3 a. This intermediate can react with a second CO molecule to form a cationic complex by substitution of the iodide ligand (4 a’) or a neutral complex by replacement of a DBU ligand (4 a) (Scheme 3). The substitution of DBU is thermodynamically favoured over that of the iodide ligand by 10.4 kcal mol1, and the calculated energy barrier is also lower ( 2 kcal mol1). Thus, we assumed that the reaction proceeds through the neutral acyl carbonyl palladium complex 4 a, and that acyl and carbonyl ligands are in cis arrangement to allow the second carbonylation of the aryl substrate. In any case, cis-to-trans isomerisation in four-coordinate PdII complexes is known to be an easy process.[16c, 17] Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
Scheme 4. Possible mechanism proceeding from a Pd acyl carbonyl species. i) Second CO insertion into a Pd ketoacyl species and ii) nucleophilic attack of NH3. Free energies in kcal mol1.
Once the acyl carbonyl Pd species is formed, we considered two possibilities (Scheme 4): 1) another CO insertion to form the Pd ketoacyl species 5 a’ and 2) nucleophilic attack of NH3 at the terminal CO ligand to form a Pd acyl amide species. The migratory insertion of CO into the Pd–acyl bond of 4 a has a significant free-energy barrier of 22.5 kcal mol1 (19.2 kcal mol1 for 4 a’). Alternatively, the free ammonia can attack the carbonyl ligand to form Pd amide complex 5 a. Since ammonia and primary amines are not strong nucleophiles, we considered that the attack is assisted by free DBU, which deprotonates the nucleophile to yield complex 5 a and DBUH + I (Scheme 4). We found a transition state TS4 a-5 a involved in the process. Note that dicarbonylation only occurs in the presence of a base.[8] In addition, the protonated DBU takes the iodide as counterion, yielding the neutral acyl amide complex 5 a (Scheme 4 and Figure 2). A similar process involving the deprotonation of an alkyne by an external pyrrolidine base has been proposed for the Pd-catalysed Sonogashira reaction.[15b] The estimated energy barrier of the concerted process (16.0 kcal mol1, from 4 a to TS4 a-5 a) is considerably lower than that of the insertion of CO into the Pd–acyl bond (by 6.5 kcal mol1). Thus, the base-assisted nucleophilic attack of the amine is likely to occur prior to formation of the second carbon–carbon bond. These facts are in agreement with the experimentally proposed mechanism for Pd/phosphine catalyst systems.[11] Finally, the Pd acyl amide intermediate 5 a undergoes reductive elimination via the transition state TS5 a-6 a (10.2 kcal mol1) to give the final a-ketoamide product and regenerate
10984
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper none of the calculated energy barriers, that is, for ammonia attack or reductive elimination, is high. This is also in agreement with experiments, since no intermediate was detected by NMR spectroscopy when the reaction was carried out in the presence of amine due to the rapidity of the reaction on the NMR timescale.[8] Thus, we can conclude that the mechanism proposed in Scheme 2 is a plausible reaction pathway fully consistent with experimental results. The competitive monocarbonylation reaction
Figure 2. Molecular structures and main geometric parameters of selected species involved in the dicarbonylation mechanism: the TS for the base-assisted nucleophilic attack of ammonia on the terminal CO ligand (TS4 a-5 a), the intermediate [Pd(DBU){C(O)Ph}{C(O)NH2}] (5 a), and the TS for the reductive elimination step (TS5 a-6 a). Hydrogen atoms omitted for clarity. Distances in ngstrçm.
the Pd0 species. In the overall mechanism for dicarbonylation (see Figure 3, solid line), the most stable intermediate is Pd acyl complex 3 a. This is consistent with the previous experimental finding that, in the absence of amine, only the intermediate 3 a was detected.[8] Moreover, once 3 a is formed,
To understand the origin of selectivity in this phosphine-free Pd system, it is mandatory to also examine the mechanism of the monocarbonylation reaction. For example, in the case of iodobenzene as substrate, the incomplete selectivity for the dicarbonylation product (76 %)[8] indicated that a competitive monocarbonylation pathway is operative and is close in energy. Taking the calculated catalytic cycle as reference, the monocarbonylation route may separate at different points depending on how and where the nucleophilic attack occurs (Figures 3–5). Thus we analysed three possible routes involving base-assisted attack of the amine: 1) at Pd in 3 a, 2) at the acyl ligand in 4 a and 3) at the terminal CO ligand in 2 a. The first route is an intramolecular stepwise mechanism involving formation of amido complex [Pd(C(O)Ph)(DBU)2(NH2)] (7 a) and subsequent reductive elimination of the NH2 and acyl ligands. This mechanism resembles that proposed for the methanolysis of palladium–acyl bonds.[18] However, intermediate 7 a lies 28.1 kcal mol1 above 3 a and is significantly higher in energy than the specific transition states for dicarbonylation. The ni-
Figure 3. Gibbs free-energy profile for the di- (continuous line) and monocarbonylation (dashed line, mechanism b) of iodobenzene (R = H) with ammonia as nucleophile. Values for the other aryl iodides in parentheses (R = p-OMe) and in brackets (R = p-CN). Energies in kcal mol1. Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
10985
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 4. Molecular structures and main geometric parameters of selected species involved in possible monocarbonylation mechanisms: the amido complex 7 a, the TS for nucleophilic attack on the acyl ligand (TS4 b-5 b), for p-cyanoiodobenzene the TS for nucleophilic attack on the terminal CO ligand (TS2 c-3 c) and Pd aryl amide complex 3 c. Hydrogen atoms omitted for clarity. Distances in ngstrçm.
trogen atom in 7 a shows a hybridised environment (see Figure 4) due to the repulsive interaction between the filled palladium dp orbital and lone-pair pp orbital of the N atom, in contrast to the strong metal–amido bonding interaction found for early transition metals with vacant d orbitals.[19] This would indicate that monocarbonylation is not competitive, which is not true for iodobenzene as substrate; consequently, we can rule out Pd amide formation. Another possibility often proposed for this type of reactions involves direct intermolecular attack of the amine at the acyl ligand (mechanism b, dashed lines in Figure 3). This attack leads to dissociation of the resulting amide species from the Pd0 centre. Similarly to the process from 4 a to 5 a, we located a transition state (TS4 b-5 b in Figure 4) in which free DBU assists nucleophilic attack of the amine by deprotonation and takes the iodide in cis position to form DBUH + I and regenerate the [Pd(DBU)(CO)] catalyst.[20] The TS4 b-5 b structure lies 22.5 kcal mol1 above the most stable intermediate, namely, Pd acyl complex 3 a, that is, 5.4 kcal mol1 higher than that for NH3 attack at the terminal CO ligand (TS4 a-5 a) and only 2.7 kcal mol1 higher than the overall barrier for the second carbon–carbon bond-formation step (TS5 a-6 a). In this case, it is difficult to reproduce the product ratio quantitatively because the transition states are quite different and therefore have little error cancelation. Nevertheless, the results are in good agreement with the experimentally observed high selectivity for dicarbonylation, albeit with non-negligible formation of the monocarbonylation product. The selectivity is very sensitive to the aryl substituent. In particular, p-methoxyiodobenzene (98 % for the a-ketoamide) and Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
p-cyanoiodobenzene (99 % for the amide) showed extreme selectivity.[8] Figure 3 shows the Gibbs free energies for these two selected substrates and compares them with iodobenzene. For the key selectivity-determining process, the amine attack, we observed that the energy barrier for the dicarbonylation route (3 a!TS4 a-5 a) remains very similar upon substitution: 17.1, 17.1 and 16.7 kcal mol1 for p-cyano-substituted, unsubstituted and p-methoxy-substituted aryls. On the other hand, the barrier for the monocarbonylation route (3 a!TS4 b-5 b) varies significantly: 21.1, 22.5 and 30.1 kcal mol1, respectively. As expected, the presence of a strong electron-donor group, such as methoxyl, deactivates the acyl carbon atom for the nucleophilic attack and results in a high energy barrier for monocarbonylation (30.1 kcal mol1). This result explains the excellent selectivity for the dicarbonylation product observed for p-methoxyiodobenzene. For p-cyanoiodobenzene, the attack at the terminal CO ligand and the dicarbonylation route are still preferred with respect to attack to at the acyl ligand and monocarbonylation route by about 5 kcal mol1, in disagreement with experiments. Therefore, we examined whether the selectivity can be also decided early in the catalytic cycle, and for this purpose, we analysed the route for monocarbonylation that starts from compound 2 a. This route (Figure 5, mechanism c, dashed lines) includes nucleophilic attack of the amine at the terminal CO ligand of the Pd phenyl carbonyl complex, followed by reductive elimination of the amide and aryl ligands. For iodobenzene, formation of Pd acyl complex 3 a is faster than DBU-assisted NH3 attack; the energy difference between the transition states is 4.7 kcal mol1. Comparison with the other two aryl substrates shows that the most substrate-sensitive step is the CO insertion (2 a!TS2 a-3 a) with energy barriers of 21.8, 10.2 and 7.9 kcal mol1 for p-cyano, no and p-methoxy substituents. This can be easily understood if the reaction is regarded as a migration of the aryl group to a terminal electrophilic CO ligand, which is therefore disfavoured by the presence of electron-withdrawing substituents. Thus, p-cyanoiodobenzene has the largest energy barrier, and the monocarbonylation route becomes clearly favoured (DG° = 21.8 versus 15.2 kcal mol1). This mechanism explains the experimental observations for the Pd/DBU system, and is in accordance with the proposed intermediates for monocarbonylation by Pd/phosphine catalysts.[11] To complete the description of mechanism c, we note that the aryl amide intermediate 3 c is neither thermodynamically nor kinetically stable. However, its direct barrier (3 c!TS3 c-5 b) to yield the amide product is lower ( 6 kcal mol1) than its reverse barrier ( 9 kcal mol1), and overall mechanism c is less energetically demanding than the dicarbonylation mechanism. In summary, the full study of monocarbonylation reaction suggests that for iodobenzene the observed by-product is formed by mechanism b, amine attack on the acyl group, whereas for p-cyanoiodobenzene mechanism c, that is, reductive elimination from the aryl amide complex, operates.
10986
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 5. Gibbs free-energy profile for the di- (continuous line) and monocarbonylation (dashed line, mechanism c) of p-cyanoiodobenzene (R = p-CN in brackets) with ammonia as nucleophile. Values of the other aryl iodides (R = H) and in parentheses (R = p-OMe). Energies in kcal mol1.
Conclusion The mechanism of the dicarbonylation of aryl iodides with the Pd/DBU catalyst system was characterised by DFT methods. The process can be described as follows: 1) oxidative addition of aryl iodide to the Pd0 catalyst 1 a’ to form the PdII species 2 a, 2) migration of the aryl group to the terminal electrophilic CO ligand to produce Pd acyl intermediate 3 a 3) replacement of a DBU ligand by CO to yield 4 a, 4) DBU-assisted nucleophilic attack of amine at a terminal CO to form the Pd acyl amide species 5 a and 5) reductive elimination of the amide and acyl ligands to yield the a-ketoamide and restore the initial catalyst. The mechanism proposed in this work for dicarbonylation shows neither deep wells nor high energy barriers, that is, it is a plausible pathway for the observed reaction. The lowestenergy intermediates correspond to Pd acyl species, which were spectroscopically characterised. The base DBU enhances the nucleophilicity of the amine through its deprotonation, which facilitates its attack at a terminal carbonyl ligand. The fact that toluene is the best solvent for the process is in agreement with a reaction mechanism in which all intermediates are neutral. To understand the selectivity, we need to consider two different scenarios. For aryl iodide substrates with electron-donor substituents (i.e., p-methoxy), the first carbon–carbon bondChem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
formation step by aryl migration to the CO ligand is fast and results in preferred formation of dicarbonylation products. The monocarbonylation reaction can take place if the acyl ligand is not deactivated enough toward the nucleophilic base-assisted attack of the amine (mechanism b). In such cases (i.e., iodobenzene), the CO and acyl ligands compete for the amine and end the reaction before the second carbon–carbon bond is formed. In contrast, for aryl iodide substrates with electronwithdrawing substituents (i.e., p-cyano), aryl migration to the CO ligand becomes slower, and this allows base-assisted amine attack on CO early in the catalytic cycle. From the Pd amide aryl complex, carbon–carbon bond formation can easily occur to yield the monocarbonylation product and end the reaction. To the best of our knowledge, this is the first time that a convincing explanation has been provided for understanding the origin of selectivity between di- and monocarbonylation.
Experimental Section Computational details All calculations were carried out by using the Gaussian 09 package of programs[21] with the M06 functional of Truhlar et al.[22] Palladium and iodine atoms were described by using the LANL2DZ effective core potential for their inner-shell electrons, and its associated
10987
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper double-z basis set for the outer electrons.[23] Additionally, for the iodine atom a d-type polarisation function was included (exponent = 0.289).[24] For C, N, H and O atoms the 6-31(d,p) basis set was used.[25] The structures of all species and transition states were optimised without symmetry constraint by using the Berny algorithm implemented in Gaussian 09.[26] All minima and transition states were confirmed by performing frequency calculations. Solvent effects of toluene (Tol, e = 2.3741) were introduced into the optimised vacuum geometries by using the self-consistent reaction field approach, by means of the integral equation formalism polarisable continuum model (IEFPCM).[27] Geometry optimisations were carried out in vacuum. Free energy corrections were calculated by employing Equation (1).
DGTol ¼ DE Tol þ ðDGGas DE Gas Þ
ð1Þ
Acknowledgements We thank the Ministerio de Economia y Competividad (MINECO) of Spain (projects CTQ2011-29054-C02-01, CTQ201127033, CTQ2010-14938/BQU, CTQ2011-22872BQU, Ramon y Cajal fellowship to C. Godard) and the Generalitat de Catalunya (2009SGR116, 2009SGR259, 2009SGR746 and 2009SGR462, XRQTC and ICIQ Foundation) for financial support. Keywords: carbonylation · chemoselectivity · density functional calculations · palladium · reaction mechanisms [1] Modern Carbonylation Methods, (Ed.: L. Kollr), Wiley-VCH, 2008. [2] H. des Abbayes, J. Y. Salan, Dalton Trans. 2003, 1041 – 1052. [3] a) E. R. Murphy, J. R. Martinelli, N. Zaborenko, S. L. Buchwald, K. F. Jensen, Angew. Chem. 2007, 119, 1764 – 1767; Angew. Chem. Int. Ed. 2007, 46, 1734 – 1737; b) A. Yamamoto, Bull. Chem. Soc. Jpn. 1995, 68, 433 – 446; c) F. Ozawa, T. Yamamoto, A. Yamamoto, J. Synth. Org. Chem. Jpn. 1985, 43, 442 – 452; d) F. Ozawa, H. Soyma, T. Yamamoto, A. Yamamoto, Tetrahedron Lett. 1982, 23, 3383 – 3386. [4] a) S. Schreiber, Science 1991, 251, 283 – 287; b) D. C. N. Swindells, P. S. White, J. A. Findlay, Can. J. Chem. 1978, 56, 2491 – 2492; c) H. Tanaka, A. Kuroda, H. Marusawa, H. Hatanaka, T. Kino, T. Goto, M. Hashimoto, T. Taga, J. Am. Chem. Soc. 1987, 109, 5031 – 5033; d) J. Gante, Angew. Chem. 1994, 106, 1780 – 1802; Angew. Chem. Int. Ed. Engl. 1994, 33, 1699 – 1720; e) Q. Guo, H.-T. Ho, I. Dicker, L. Fan, N. Zhou, J. Friborg, T. Wang, B. V. McAuliffe, H.-g. H. Wang, R. E. Rose, H. Fang, H. T. Scarnati, D. R. Langley, N. A. Meanwell, R. Abraham, R. J. Colonno, P.-f. Lin, J. Virol. 2003, 77, 10528 – 10536; f) H.-H. Otto, T. Schirmeister, Chem. Rev. 1997, 97, 133 – 172; g) T. Wang, Z. Zhang, O. B. Wallace, M. Deshpande, H. Fang, Z. Yang, L. M. Zadjura, D. L. Tweedie, S. Huang, F. Zhao, S. Ranadive, B. S. Robinson, Y.-F. Gong, K. Ricarrdi, T. P. Spicer, C. Deminie, R. Rose, H.-G. H. Wang, W. S. Blair, P.-Y. Shi, P.-f. Lin, R. J. Colonno, N. A. Meanwell, J. Med. Chem. 2003, 46, 4236 – 4239; h) Y. J. Yoo, D. H. Nam, S. Y. Jung, J. W. Jang, H. J. Kim, C. Jin, A. N. Pae, Y. S. Lee, Bioorg. Med. Chem. Lett. 2011, 21, 2850 – 2854. [5] a) J. Balogh, . Kuik, L. rge, F. Darvas, J. Bakos, R. Skoda-Fçldes, J. Mol. Catal. A 2009, 302, 76 – 79; b) Z. Szarka, . Kuik, R. Skoda-Fçldes, L. Kollr, J. Organomet. Chem. 2004, 689, 2770 – 2775; c) N. Tsukada, Y. Ohba, Y. Inoue, J. Organomet. Chem. 2003, 687, 436 – 443; d) Y. Ouzumi, T. Arii, T. Watanabe, J. Org. Chem. 2001, 66, 5272 – 527; e) Z. Szarka, R. Skoda-Fçldes, L. Kollr, Tetrahedron Lett. 2001, 42, 739 – 741. [6] a) D. U. Nielsen, K. Neumann, R. H. Taaning, A. T. Lindhardt, A. Modvig, T. Skrydstrup, J. Org. Chem. 2012, 77, 6155 – 6165; b) F. Ozawa, T. Sugimoto, Y. Yuasa, M. Santra, T. Yamamoto, A. Yamamoto, Organometallics 1984, 3, 683 – 692. [7] M. Iizuka, Y. Kondo, Chem. Commun. 2006, 1739 – 1741. Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
[8] V. de La Fuente, C. Godard, E. Zangrando, C. Claver, S. Castillon, Chem. Commun. 2012, 48, 1695 – 1697. [9] M. Papp, R. Skoda-Fçldes, J. Mol. Catal. A 2013, 378, 193 – 199. [10] S. Zheng, Y. Wang, C. Zhang, J. Liu, C. Xia, Appl. Organomet. Chem. 2014, 28, 48 – 53. [11] a) A. Yamamoto, F. Ozawa, K. Osakada, L. Huang, T. Son, N. Kawasasku, M- K. Doh, Pure Appl. Chem. 1991, 63, 687 – 696; b) L. Huang, F. Ozawa, A. Yamamoto, Organometallics 1990, 9, 2603 – 2611; c) F. Ozawa, H. Soyama, H. Yanagihara, I. Aoyama, H. Takino, K. Izawa, T. Yamamoto, A. Yamamoto, J. Am. Chem. Soc. 1985, 107, 3235 – 3245. [12] F. Ozawa, N. Kawasaki, H. Okamoto, T. Yamamoto, A. Yamamoto, Oraganometallics 1987, 6, 1640 – 1651. [13] a) M. Garc a-Melchor, A. A. C. Braga, A. Lleds, G. Ujaque, F. Maseras, Acc. Chem. Res. 2013, 46, 2626 – 2634; b) M. P rez-Rodr guez, A. A. C. Braga, A. R. deLera, F. Maseras, R. Alvarez, P. Espinet, Organometallics 2010, 29, 4983 – 4991; c) B. Fuentes, M. Garc a-Melchor, A. Lleds, F. Maseras, J. A. Casares, G. Ujaque, P. Espinet, Chem. Eur. J. 2010, 16, 8596 – 8599; d) M. P rez-Rodr guez, A. A. C. Braga, M. Garc a-Melchor, M. H. P rez-Temprano, J. A. Casares, G. Ujaque, A A. R. deLera, R. Alvarez, F. Maseras, P. Espinet, J. Am. Chem. Soc. 2009, 131, 3650 – 3657; e) A. A. C. Braga, G. Ujaque, F. Maseras, Organometallics 2006, 25, 3647 – 3658; f) A. A. C. Braga, N. H. Morgon, G. Ujaque, F. Maseras, J. Am. Chem. Soc. 2005, 127, 9298 – 9307. [14] a) M. Barsukova, N. V. Izarova, R. N. Biboum, B. Keita, L. Nadjo, V. Ramachandran, N. S. Dalal, N. S. Antonova, J. J. Carb, J. M. Poblet, U. Kortz, Chem. Eur. J. 2010, 16, 9076 – 9085; b) E. Serrano, R. Navarro, T. Soler, J. J. Carb, A. Lleds, E. P. Urriolabeitia, Inorg. Chem. 2009, 48, 6823 – 6834; c) V. Lillo, E. Mas-Marza, A. M. Segarra, J. J. Carb, C. Bo, E. Peris, E. Fernndez, Chem. Commun. 2007, 32, 3380 – 3382; d) E. Serrano, C. Valles, J. J. Carb, A. Lleds, T. Soler, R. Navarro, E. P. Urriolabeitia, Organometallics 2006, 25, 4653 – 4664. [15] a) K. Vikse, T. Naka, J. S. McIndoe, M. Besora, F. Maseras, ChemCatChem 2013, 5, 3604—3609; b) M. Garc a-Melchor, M. C. Pacheco, C. Njera, A. Lleds, G. Ujaque, ACS Catal. 2011, 1, 135 – 144; c) A. A. C. Braga, G. Ujaque, F. Maseras, Organometallics 2006, 25, 3647 – 3658; d) L. J. Goossen, D. Koley, H. L. Hermann, W. Thiel, Organometallics 2005, 24, 2398 – 2410. [16] a) Y. Hu, J. Liu, Z. L, X. Luo, H. Zhang, Y. Lan, A. Lei, J. Am. Chem. Soc. 2010, 132, 3153 – 3158; b) S. A. Macgregor, G. W. Neave, Organometallics 2003, 22, 4547 – 4556; c) K. E. Frankcombe, K. J. Cavell, B. F. Yates, R. B. Knott, Organometallics 1997, 16, 3199 – 3206. [17] A. L. Casado, P. Espinet, Organometallics 1998, 17, 954 – 959. [18] a) S. M. A. Donald, S. A. Macgregor, V. Settels, D. J. Cole-Hamilton, G. R. Eastham, Chem. Commun. 2007, 0, 562 – 564; b) E. Zuidema, C. Bo, P. W. N. M. van Leeuwen, J. Am. Chem. Soc. 2007, 129, 3989 – 4000; c) P. W. N. M. van Leeuwen, M. A. Zuideveld, B. H. G. Swennenhuis, Z. Freixa, P. C. J. Kamer, K. Goubitz, J. Fraanje, M. Lutz, A. L. Spek, J. Am. Chem. Soc. 2003, 125, 5523 – 5539. [19] a) S. Aguado-Ullate, J. J. Carb, O. Gonzlez-del Moral, A. Mart n, M. Mena, J. M. Poblet, C. Santamar a, Inorg. Chem. 2011, 50, 6269 – 6279; b) E. Sicilia, N. Russo, J. Am. Chem. Soc. 2002, 124, 1471 – 1480. [20] For transition state TS4 b-5 b, we assumed that the acyl and iodide ligands are in cis positions to allow DBU to remove simultaneously the proton of the amine and the iodide ligand. Consequently, TS4 b-5 b, in fact, connects to the intermediate cis-[Pd(DBU)(CO){C(O)Ph)}I] (4 b). This species is only about 1 kcal mol1 higher in energy than 4 a and we expect easy trans-to-cis isomerisation for these four-coordinate PdII complex (see refs. [16c] and [17] and NMR data in ref. [8]). For the sake of simplicity, we did not add intermediate 4 b to the energy profiles. Analogous reasoning can be applied to transition state TS2 c-3 c. [21] Gaussian 09, Revision A.02, M. J. Frisch, 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, 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, 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. Moro-
10988
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
[22] [23] [24]
[25]
kuma, 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, and D. J. Fox, Gaussian Inc., Wallingford, CT, 2009. Y. Zhao, D. Truhlar, Theor. Chem. Acc. 2008, 120, 215 – 241. P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299 – 310. A. Hçllwarth, M. Bçhme, S. Dapprich, A. W. Ehlers, A. Gobbi, V. Jonas, K. F. Kçhler, R. Stegmann, A. Veldkamp, G. Frenking, Chem. Phys. Lett. 1993, 208, 237. a) M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M. S. Gordon, D. J. Defrees, J. A. Pople, J. Chem. Phys. 1982, 77, 3654; b) W. J. Hehre, R.
Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
Ditchfield, J. A. Pople, J. Chem. Phys. 1972, 56, 2257; c) P. C. Hariharan, J. A. Pople, Theor. Chim. Acta. 1973, 28, 213. [26] P. Pulay, G. Fogarasi, J. Chem. Phys. 1992, 96, 2856 – 2860. [27] B. Mennucci, E. Canc s, J. Tomasi, J. Phys. Chem. B 1997, 101, 10506 – 10517.
Received: April 10, 2014 Published online on July 22, 2014
10989
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
& Reaction Mechanisms
Pd-Catalysed Mono- and Dicarbonylation of Aryl Iodides: Insights into the Mechanism and the Selectivity Victor M. Fernndez-Alvarez ,[a, b] Vernica de la Fuente,[a] Cyril Godard,[a] Sergio Castilln,[c] Carmen Claver,[a] Feliu Maseras ,*[b, d] and Jorge J. Carb*[a]
Abstract: The mechanism of the experimentally reported phosphine-free palladium-catalysed carbonylation of aryl iodides with amines in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as base was investigated at the DFT level. Paths were identified for both di- and monocarbonylation, and the calculated selectivity for three different substrates was in agreement with experiment. In dicarbonylation yielding a-ketoamides, formation of the second carbon–carbon bond occurs through reductive elimination in the Pd acyl amide intermediate after DBU-assisted nucleophilic attack of an amine at a terminal CO ligand. This path yields the major product with iodobenzene and the almost
exclusive product with p-methoxyiodobenzene. Two different possible pathways yield the monocarbonylated amide product. In one of them, which affords the minor product for iodobenzene, base-assisted nucleophilic attack of the amine takes place on a Pd-bound acyl ligand. For substrates with electron-withdrawing substituents, such as p-cyanoiodobenzene, aryl migration to the CO ligand is disfavoured, and this allows base-assisted amine attack at a terminal CO ligand early in the catalytic cycle. From the resulting Pd amide aryl complex, the subsequent reductive elimination occurs easily, and monocarbonylation becomes favoured.
Introduction The carbonylation of organic substrates mediated by transition metal complexes is a convenient method for forming carbon– carbon bonds and incorporating carbonyl functional groups.[1] A fascinating synthetic possibility is the sequential incorporation of two molecules of carbon monoxide, which can lead to the synthesis of valuable products.[2] In particular, the Pd-catalysed dicarbonylation of aryl halides provides a direct method for the production of a-ketoamides (Scheme 1, left),[2, 3] which [a] V. M. Fernndez-Alvarez , V. de la Fuente, Dr. C. Godard, Prof. Dr. C. Claver, Dr. J. J. Carb Departament de Qumica Fsica i Inorgnica Universitat Rovira i Virgili C/Marcel.li Domingo s/n, 43007 Tarragona (Spain) Fax: (+ 34) 977-559-563 E-mail: [email protected] [b] V. M. Fernndez-Alvarez , Prof. Dr. F. Maseras Institute of Chemical Research of Catalonia (ICIQ) Av Paı¨sos Catalans 16, 43007 Tarragona, Catalonia (Spain) Fax: (+ 34) 977-920-231 E-mail: [email protected] [c] Prof. Dr. S. Castilln Departament de Qumica Analtica i Orgnica Universitat Rovira i Virgili C/Marcel.li Domingo s/n, 43007 Tarragona (Spain) [d] Prof. Dr. F. Maseras Departament de Qumica Universitat Autnoma de Barcelona 08193 Bellaterra, Catalonia (Spain) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/chem.201403022. Chem. Eur. J. 2014, 20, 10982 – 10989
Scheme 1. Competing reactions involved in the carbonylation process.
are precursors of biologically active compounds and several families of heterocycles.[4] Most of the reported Pd catalysts are modified with phosphine ligands,[5, 6, 7] which limit their application due to inherent problems with oxidation. One of the most efficient phosphine-based methods for dicarbonylation was discovered by Kondo and co-workers,[7] who used a Pd/PtBu3 catalyst in the presence of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as base. This reaction produced a-ketoamides selectively without the need for high CO pressures. Recently, some of us reported the first efficient phosphine-free Pd catalyst system for double carbonylation of aryl iodides.[8] In the presence of DBU, the reaction produced excellent conversions and selectivities for a wide range of aryl iodides and amine nucleophiles under atmospheric CO pressure. The phosphine-free protocol was then extended to supported catalysis by Papp and Skoda-Fçldes.[9] Another alternative to phosphines is the use of N-heterocyclic carbenes, which catalysed the dicarbonylation of aryl iodides under mild conditions.[10] The reaction mechanism of Pd/phosphine-catalysed monoand dicarbonylation has been experimentally studied.[11] In the dicarbonylation process the initial steps of the reaction are common to other Pd-mediated carbonylation reactions involv-
10982
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper ing an oxidative addition of the aryl iodide to Pd0 species followed by CO insertion into the Pd–aryl bond. The process proceeds by coordination of CO, nucleophilic attack of the amine at the terminal CO ligand to form a Pd acyl carbamoyl species and reductive elimination to give the a-ketoamide product. The mechanism involving insertion of CO into the Pd acyl species has also been considered.[10, 11a] Dicarbonylation takes place under conditions very similar to those used for monocarbonylation, and the two processes can compete (Scheme 1). Indeed, this procedure cannot be used for aryl iodides with electron-withdrawing substituents, because the selectivity shifts towards the monocarbonylation products.[5d, 8] It seems also clear from previous reports that the nature of the base is crucial in these processes, and that a large excess of base is always required for dicarbonylation.[5c,d, 8] The mechanism of monocarbonylation is controversial, and a different mechanism has been proposed when alcohols are used as nucleophiles instead of amines.[11, 12] A better understanding of the molecular basis of the chemoselectivity appears to be the key to the development of more efficient and selective processes. As a part of our ongoing projects to understand theoretically the role of Pd complexes in CC bond-forming reactions and other aspects of Pd chemistry,[13, 14] we report herein a detailed mechanistic study of the di- and monocarbonylation of different aryl iodides with amines catalysed by the Pd/DBU system. In the phosphine-free palladium-catalysed dicarbonylation process, DBU is fundamental, since it plays the role of ligand, base and eventually nucleophile.[8] Our goals were to characterise the catalytic cycle in which DBU is the ligand, to determine the role of the base and to identify the factors that favour the incorporation of two CO molecules (dicarbonylation) instead of one CO molecule (monocarbonylation) into an organic substrate. Herein, we focus on the reaction between PdL2 (L = DBU, CO) complexes, NH3 and iodobenzene derivatives. We studied three different substrates: iodobenzene, pmethoxyiodobenzene and p-cyanoiodobenzene. This set of substrates is particularly intriguing, because their dicarbonylation:monocarbonylation ratios are 76:24, 98:2 and 1:99, respectively.[8]
Results and Discussion Initially, we analysed the dicarbonylation mechanism shown in Scheme 2, which was previously proposed and based on experimental findings.[8] The mechanism can be divided into several steps: 1) oxidative addition of iodobenzene to 1 a to yield 2 a, 2) migration of the phenyl group to the terminal CO ligand to form Pd acyl complex 3 a, 3) substitution of a DBU ligand by a second equivalent of CO to produce 4 a, 4) nucleophilic attack of ammonia on coordinated CO assisted by a base to yield 5 a and 5) reductive elimination to form the a-ketoamide and restore the catalyst 1 a. Figures 1 and 2 show the structures of the key species, and Figure 3 depicts the overall freeenergy profile of the reaction. Then, we determined the mechanism for monocarbonylation in order to identify the factors governing the selectivity of the process. Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
Scheme 2. Studied mechanism for the dicarbonylation of aryl iodide by the Pd/DBU system with ammonia as nucleophile and DBU as base.
The first carbon–carbon bond-formation step Formation of the spectroscopically characterised Pd acyl complex [PdI(DBU)2{C(O)aryl}] (3 a) involves oxidative addition of aryl iodide to Pd0 and Ph migration to the terminal CO ligand. Both elementary reactions have been studied by computational methods.[15, 16] Therefore, these steps were not analysed in detail, but the eventual consequences of using DBU ligands are discussed. Because the reaction medium contains a small excess of both CO and DBU, several Pd0 species could coexist: [Pd(DBU)2] (1 a’), [Pd(DBU)(CO)] (1 a) and [Pd(CO)2] (1 a’’). Their relative free energies in toluene are + 25.3, 0.0 and + 5.4 kcal mol1, respectively. Thus, we selected the most stable Pd0 species 1 a as the starting point for the catalytic reaction. First, the active species 1 a reacts with iodobenzene to form the cis oxidative-addition product [Pd(DBU)(CO)(Ph)I] (2 a). As previously observed, the CI cleavage leads to a cis PdII complex.[17] The calculated transition state for oxidative addition of aryl iodide lies 21.6 kcal mol1 above the reactants, and the process is slightly endothermic by 0.9 kcal mol1. For the oxidative addition of PhI to [Pd(PH3)2], the calculated free energy barrier in CH2Cl2 solvent at a similar level of theory (B3LYP/PCM) is somewhat lower (17.0 kcal mol1),[15b] that is, introduction of p-acidic ligands such as CO somewhat disfavours the CI bond activation. Migration of the phenyl group to the carbonyl ligand and coordination of an additional DBU ligand forms cis-[Pd(DBU)2{C(O)Ph}I] (3 a), which has remarkable thermodynamic stability (14.4 kcal mol1 below the reactants). The corresponding trans isomer is almost isoenergetic (0.7 kcal mol1 above), because the two large DBU ligands in cis positions adopt an alternate conformation to avoid mutual steric hindrance (see Figure 1). This result is in agreement with the equilibrium between cis and trans species observed by NMR spectroscopy.[8] For phenyl migration we considered a three-centre mechanism[16b] and located transition state TS2 a-3 a with an energy
10983
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Scheme 3. Possible CO ligand exchange in complex 3 a. Free energies in kcal mol1.
Figure 1. Selected molecular structures and main geometric parameters of the first part of the cycle: the transition states for Ph migration to terminal CO (TS2 a-3 a) and for replacement of DBU by free CO (TS3 a-4 a) and the intermediate cis-[Pd(DBU)2{C(O)Ph}I] 3 a. Hydrogen atoms omitted for clarity. Distances in ngstrçm.
barrier of 11.1 kcal mol1 and the phenyl group migrating in the plane of the complex (see Figure 1). Both the energy and the geometry closely resemble those recently calculated for trans-[PdPh(CO)(OEt)(PMe3)] by Lei et al.[16a] The intermediate 3 a is not only thermodynamically stable (DGTol = 14.4 kcal mol1), but its reverse energy barrier is also quite high (25.4 kcal mol1). Thus, the CO migratory insertion step can be classified as an irreversible step in the catalytic cycle. In summary, calculations show that the formation of Pd acyl species by oxidative addition and subsequent CO migratory insertion is not energetically demanding and leads to a thermodynamically stable complex. This result is in full agreement with previous experimental findings.[8] The second carbon–carbon bond-formation step: dicarbonylation Using the spectroscopically characterised Pd acyl complex as a reliable stepping stone, we analysed the mechanism of a-ketoamide formation from complex 3 a. This intermediate can react with a second CO molecule to form a cationic complex by substitution of the iodide ligand (4 a’) or a neutral complex by replacement of a DBU ligand (4 a) (Scheme 3). The substitution of DBU is thermodynamically favoured over that of the iodide ligand by 10.4 kcal mol1, and the calculated energy barrier is also lower ( 2 kcal mol1). Thus, we assumed that the reaction proceeds through the neutral acyl carbonyl palladium complex 4 a, and that acyl and carbonyl ligands are in cis arrangement to allow the second carbonylation of the aryl substrate. In any case, cis-to-trans isomerisation in four-coordinate PdII complexes is known to be an easy process.[16c, 17] Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
Scheme 4. Possible mechanism proceeding from a Pd acyl carbonyl species. i) Second CO insertion into a Pd ketoacyl species and ii) nucleophilic attack of NH3. Free energies in kcal mol1.
Once the acyl carbonyl Pd species is formed, we considered two possibilities (Scheme 4): 1) another CO insertion to form the Pd ketoacyl species 5 a’ and 2) nucleophilic attack of NH3 at the terminal CO ligand to form a Pd acyl amide species. The migratory insertion of CO into the Pd–acyl bond of 4 a has a significant free-energy barrier of 22.5 kcal mol1 (19.2 kcal mol1 for 4 a’). Alternatively, the free ammonia can attack the carbonyl ligand to form Pd amide complex 5 a. Since ammonia and primary amines are not strong nucleophiles, we considered that the attack is assisted by free DBU, which deprotonates the nucleophile to yield complex 5 a and DBUH + I (Scheme 4). We found a transition state TS4 a-5 a involved in the process. Note that dicarbonylation only occurs in the presence of a base.[8] In addition, the protonated DBU takes the iodide as counterion, yielding the neutral acyl amide complex 5 a (Scheme 4 and Figure 2). A similar process involving the deprotonation of an alkyne by an external pyrrolidine base has been proposed for the Pd-catalysed Sonogashira reaction.[15b] The estimated energy barrier of the concerted process (16.0 kcal mol1, from 4 a to TS4 a-5 a) is considerably lower than that of the insertion of CO into the Pd–acyl bond (by 6.5 kcal mol1). Thus, the base-assisted nucleophilic attack of the amine is likely to occur prior to formation of the second carbon–carbon bond. These facts are in agreement with the experimentally proposed mechanism for Pd/phosphine catalyst systems.[11] Finally, the Pd acyl amide intermediate 5 a undergoes reductive elimination via the transition state TS5 a-6 a (10.2 kcal mol1) to give the final a-ketoamide product and regenerate
10984
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper none of the calculated energy barriers, that is, for ammonia attack or reductive elimination, is high. This is also in agreement with experiments, since no intermediate was detected by NMR spectroscopy when the reaction was carried out in the presence of amine due to the rapidity of the reaction on the NMR timescale.[8] Thus, we can conclude that the mechanism proposed in Scheme 2 is a plausible reaction pathway fully consistent with experimental results. The competitive monocarbonylation reaction
Figure 2. Molecular structures and main geometric parameters of selected species involved in the dicarbonylation mechanism: the TS for the base-assisted nucleophilic attack of ammonia on the terminal CO ligand (TS4 a-5 a), the intermediate [Pd(DBU){C(O)Ph}{C(O)NH2}] (5 a), and the TS for the reductive elimination step (TS5 a-6 a). Hydrogen atoms omitted for clarity. Distances in ngstrçm.
the Pd0 species. In the overall mechanism for dicarbonylation (see Figure 3, solid line), the most stable intermediate is Pd acyl complex 3 a. This is consistent with the previous experimental finding that, in the absence of amine, only the intermediate 3 a was detected.[8] Moreover, once 3 a is formed,
To understand the origin of selectivity in this phosphine-free Pd system, it is mandatory to also examine the mechanism of the monocarbonylation reaction. For example, in the case of iodobenzene as substrate, the incomplete selectivity for the dicarbonylation product (76 %)[8] indicated that a competitive monocarbonylation pathway is operative and is close in energy. Taking the calculated catalytic cycle as reference, the monocarbonylation route may separate at different points depending on how and where the nucleophilic attack occurs (Figures 3–5). Thus we analysed three possible routes involving base-assisted attack of the amine: 1) at Pd in 3 a, 2) at the acyl ligand in 4 a and 3) at the terminal CO ligand in 2 a. The first route is an intramolecular stepwise mechanism involving formation of amido complex [Pd(C(O)Ph)(DBU)2(NH2)] (7 a) and subsequent reductive elimination of the NH2 and acyl ligands. This mechanism resembles that proposed for the methanolysis of palladium–acyl bonds.[18] However, intermediate 7 a lies 28.1 kcal mol1 above 3 a and is significantly higher in energy than the specific transition states for dicarbonylation. The ni-
Figure 3. Gibbs free-energy profile for the di- (continuous line) and monocarbonylation (dashed line, mechanism b) of iodobenzene (R = H) with ammonia as nucleophile. Values for the other aryl iodides in parentheses (R = p-OMe) and in brackets (R = p-CN). Energies in kcal mol1. Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
10985
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 4. Molecular structures and main geometric parameters of selected species involved in possible monocarbonylation mechanisms: the amido complex 7 a, the TS for nucleophilic attack on the acyl ligand (TS4 b-5 b), for p-cyanoiodobenzene the TS for nucleophilic attack on the terminal CO ligand (TS2 c-3 c) and Pd aryl amide complex 3 c. Hydrogen atoms omitted for clarity. Distances in ngstrçm.
trogen atom in 7 a shows a hybridised environment (see Figure 4) due to the repulsive interaction between the filled palladium dp orbital and lone-pair pp orbital of the N atom, in contrast to the strong metal–amido bonding interaction found for early transition metals with vacant d orbitals.[19] This would indicate that monocarbonylation is not competitive, which is not true for iodobenzene as substrate; consequently, we can rule out Pd amide formation. Another possibility often proposed for this type of reactions involves direct intermolecular attack of the amine at the acyl ligand (mechanism b, dashed lines in Figure 3). This attack leads to dissociation of the resulting amide species from the Pd0 centre. Similarly to the process from 4 a to 5 a, we located a transition state (TS4 b-5 b in Figure 4) in which free DBU assists nucleophilic attack of the amine by deprotonation and takes the iodide in cis position to form DBUH + I and regenerate the [Pd(DBU)(CO)] catalyst.[20] The TS4 b-5 b structure lies 22.5 kcal mol1 above the most stable intermediate, namely, Pd acyl complex 3 a, that is, 5.4 kcal mol1 higher than that for NH3 attack at the terminal CO ligand (TS4 a-5 a) and only 2.7 kcal mol1 higher than the overall barrier for the second carbon–carbon bond-formation step (TS5 a-6 a). In this case, it is difficult to reproduce the product ratio quantitatively because the transition states are quite different and therefore have little error cancelation. Nevertheless, the results are in good agreement with the experimentally observed high selectivity for dicarbonylation, albeit with non-negligible formation of the monocarbonylation product. The selectivity is very sensitive to the aryl substituent. In particular, p-methoxyiodobenzene (98 % for the a-ketoamide) and Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
p-cyanoiodobenzene (99 % for the amide) showed extreme selectivity.[8] Figure 3 shows the Gibbs free energies for these two selected substrates and compares them with iodobenzene. For the key selectivity-determining process, the amine attack, we observed that the energy barrier for the dicarbonylation route (3 a!TS4 a-5 a) remains very similar upon substitution: 17.1, 17.1 and 16.7 kcal mol1 for p-cyano-substituted, unsubstituted and p-methoxy-substituted aryls. On the other hand, the barrier for the monocarbonylation route (3 a!TS4 b-5 b) varies significantly: 21.1, 22.5 and 30.1 kcal mol1, respectively. As expected, the presence of a strong electron-donor group, such as methoxyl, deactivates the acyl carbon atom for the nucleophilic attack and results in a high energy barrier for monocarbonylation (30.1 kcal mol1). This result explains the excellent selectivity for the dicarbonylation product observed for p-methoxyiodobenzene. For p-cyanoiodobenzene, the attack at the terminal CO ligand and the dicarbonylation route are still preferred with respect to attack to at the acyl ligand and monocarbonylation route by about 5 kcal mol1, in disagreement with experiments. Therefore, we examined whether the selectivity can be also decided early in the catalytic cycle, and for this purpose, we analysed the route for monocarbonylation that starts from compound 2 a. This route (Figure 5, mechanism c, dashed lines) includes nucleophilic attack of the amine at the terminal CO ligand of the Pd phenyl carbonyl complex, followed by reductive elimination of the amide and aryl ligands. For iodobenzene, formation of Pd acyl complex 3 a is faster than DBU-assisted NH3 attack; the energy difference between the transition states is 4.7 kcal mol1. Comparison with the other two aryl substrates shows that the most substrate-sensitive step is the CO insertion (2 a!TS2 a-3 a) with energy barriers of 21.8, 10.2 and 7.9 kcal mol1 for p-cyano, no and p-methoxy substituents. This can be easily understood if the reaction is regarded as a migration of the aryl group to a terminal electrophilic CO ligand, which is therefore disfavoured by the presence of electron-withdrawing substituents. Thus, p-cyanoiodobenzene has the largest energy barrier, and the monocarbonylation route becomes clearly favoured (DG° = 21.8 versus 15.2 kcal mol1). This mechanism explains the experimental observations for the Pd/DBU system, and is in accordance with the proposed intermediates for monocarbonylation by Pd/phosphine catalysts.[11] To complete the description of mechanism c, we note that the aryl amide intermediate 3 c is neither thermodynamically nor kinetically stable. However, its direct barrier (3 c!TS3 c-5 b) to yield the amide product is lower ( 6 kcal mol1) than its reverse barrier ( 9 kcal mol1), and overall mechanism c is less energetically demanding than the dicarbonylation mechanism. In summary, the full study of monocarbonylation reaction suggests that for iodobenzene the observed by-product is formed by mechanism b, amine attack on the acyl group, whereas for p-cyanoiodobenzene mechanism c, that is, reductive elimination from the aryl amide complex, operates.
10986
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
Figure 5. Gibbs free-energy profile for the di- (continuous line) and monocarbonylation (dashed line, mechanism c) of p-cyanoiodobenzene (R = p-CN in brackets) with ammonia as nucleophile. Values of the other aryl iodides (R = H) and in parentheses (R = p-OMe). Energies in kcal mol1.
Conclusion The mechanism of the dicarbonylation of aryl iodides with the Pd/DBU catalyst system was characterised by DFT methods. The process can be described as follows: 1) oxidative addition of aryl iodide to the Pd0 catalyst 1 a’ to form the PdII species 2 a, 2) migration of the aryl group to the terminal electrophilic CO ligand to produce Pd acyl intermediate 3 a 3) replacement of a DBU ligand by CO to yield 4 a, 4) DBU-assisted nucleophilic attack of amine at a terminal CO to form the Pd acyl amide species 5 a and 5) reductive elimination of the amide and acyl ligands to yield the a-ketoamide and restore the initial catalyst. The mechanism proposed in this work for dicarbonylation shows neither deep wells nor high energy barriers, that is, it is a plausible pathway for the observed reaction. The lowestenergy intermediates correspond to Pd acyl species, which were spectroscopically characterised. The base DBU enhances the nucleophilicity of the amine through its deprotonation, which facilitates its attack at a terminal carbonyl ligand. The fact that toluene is the best solvent for the process is in agreement with a reaction mechanism in which all intermediates are neutral. To understand the selectivity, we need to consider two different scenarios. For aryl iodide substrates with electron-donor substituents (i.e., p-methoxy), the first carbon–carbon bondChem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
formation step by aryl migration to the CO ligand is fast and results in preferred formation of dicarbonylation products. The monocarbonylation reaction can take place if the acyl ligand is not deactivated enough toward the nucleophilic base-assisted attack of the amine (mechanism b). In such cases (i.e., iodobenzene), the CO and acyl ligands compete for the amine and end the reaction before the second carbon–carbon bond is formed. In contrast, for aryl iodide substrates with electronwithdrawing substituents (i.e., p-cyano), aryl migration to the CO ligand becomes slower, and this allows base-assisted amine attack on CO early in the catalytic cycle. From the Pd amide aryl complex, carbon–carbon bond formation can easily occur to yield the monocarbonylation product and end the reaction. To the best of our knowledge, this is the first time that a convincing explanation has been provided for understanding the origin of selectivity between di- and monocarbonylation.
Experimental Section Computational details All calculations were carried out by using the Gaussian 09 package of programs[21] with the M06 functional of Truhlar et al.[22] Palladium and iodine atoms were described by using the LANL2DZ effective core potential for their inner-shell electrons, and its associated
10987
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper double-z basis set for the outer electrons.[23] Additionally, for the iodine atom a d-type polarisation function was included (exponent = 0.289).[24] For C, N, H and O atoms the 6-31(d,p) basis set was used.[25] The structures of all species and transition states were optimised without symmetry constraint by using the Berny algorithm implemented in Gaussian 09.[26] All minima and transition states were confirmed by performing frequency calculations. Solvent effects of toluene (Tol, e = 2.3741) were introduced into the optimised vacuum geometries by using the self-consistent reaction field approach, by means of the integral equation formalism polarisable continuum model (IEFPCM).[27] Geometry optimisations were carried out in vacuum. Free energy corrections were calculated by employing Equation (1).
DGTol ¼ DE Tol þ ðDGGas DE Gas Þ
ð1Þ
Acknowledgements We thank the Ministerio de Economia y Competividad (MINECO) of Spain (projects CTQ2011-29054-C02-01, CTQ201127033, CTQ2010-14938/BQU, CTQ2011-22872BQU, Ramon y Cajal fellowship to C. Godard) and the Generalitat de Catalunya (2009SGR116, 2009SGR259, 2009SGR746 and 2009SGR462, XRQTC and ICIQ Foundation) for financial support. Keywords: carbonylation · chemoselectivity · density functional calculations · palladium · reaction mechanisms [1] Modern Carbonylation Methods, (Ed.: L. Kollr), Wiley-VCH, 2008. [2] H. des Abbayes, J. Y. Salan, Dalton Trans. 2003, 1041 – 1052. [3] a) E. R. Murphy, J. R. Martinelli, N. Zaborenko, S. L. Buchwald, K. F. Jensen, Angew. Chem. 2007, 119, 1764 – 1767; Angew. Chem. Int. Ed. 2007, 46, 1734 – 1737; b) A. Yamamoto, Bull. Chem. Soc. Jpn. 1995, 68, 433 – 446; c) F. Ozawa, T. Yamamoto, A. Yamamoto, J. Synth. Org. Chem. Jpn. 1985, 43, 442 – 452; d) F. Ozawa, H. Soyma, T. Yamamoto, A. Yamamoto, Tetrahedron Lett. 1982, 23, 3383 – 3386. [4] a) S. Schreiber, Science 1991, 251, 283 – 287; b) D. C. N. Swindells, P. S. White, J. A. Findlay, Can. J. Chem. 1978, 56, 2491 – 2492; c) H. Tanaka, A. Kuroda, H. Marusawa, H. Hatanaka, T. Kino, T. Goto, M. Hashimoto, T. Taga, J. Am. Chem. Soc. 1987, 109, 5031 – 5033; d) J. Gante, Angew. Chem. 1994, 106, 1780 – 1802; Angew. Chem. Int. Ed. Engl. 1994, 33, 1699 – 1720; e) Q. Guo, H.-T. Ho, I. Dicker, L. Fan, N. Zhou, J. Friborg, T. Wang, B. V. McAuliffe, H.-g. H. Wang, R. E. Rose, H. Fang, H. T. Scarnati, D. R. Langley, N. A. Meanwell, R. Abraham, R. J. Colonno, P.-f. Lin, J. Virol. 2003, 77, 10528 – 10536; f) H.-H. Otto, T. Schirmeister, Chem. Rev. 1997, 97, 133 – 172; g) T. Wang, Z. Zhang, O. B. Wallace, M. Deshpande, H. Fang, Z. Yang, L. M. Zadjura, D. L. Tweedie, S. Huang, F. Zhao, S. Ranadive, B. S. Robinson, Y.-F. Gong, K. Ricarrdi, T. P. Spicer, C. Deminie, R. Rose, H.-G. H. Wang, W. S. Blair, P.-Y. Shi, P.-f. Lin, R. J. Colonno, N. A. Meanwell, J. Med. Chem. 2003, 46, 4236 – 4239; h) Y. J. Yoo, D. H. Nam, S. Y. Jung, J. W. Jang, H. J. Kim, C. Jin, A. N. Pae, Y. S. Lee, Bioorg. Med. Chem. Lett. 2011, 21, 2850 – 2854. [5] a) J. Balogh, . Kuik, L. rge, F. Darvas, J. Bakos, R. Skoda-Fçldes, J. Mol. Catal. A 2009, 302, 76 – 79; b) Z. Szarka, . Kuik, R. Skoda-Fçldes, L. Kollr, J. Organomet. Chem. 2004, 689, 2770 – 2775; c) N. Tsukada, Y. Ohba, Y. Inoue, J. Organomet. Chem. 2003, 687, 436 – 443; d) Y. Ouzumi, T. Arii, T. Watanabe, J. Org. Chem. 2001, 66, 5272 – 527; e) Z. Szarka, R. Skoda-Fçldes, L. Kollr, Tetrahedron Lett. 2001, 42, 739 – 741. [6] a) D. U. Nielsen, K. Neumann, R. H. Taaning, A. T. Lindhardt, A. Modvig, T. Skrydstrup, J. Org. Chem. 2012, 77, 6155 – 6165; b) F. Ozawa, T. Sugimoto, Y. Yuasa, M. Santra, T. Yamamoto, A. Yamamoto, Organometallics 1984, 3, 683 – 692. [7] M. Iizuka, Y. Kondo, Chem. Commun. 2006, 1739 – 1741. Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
[8] V. de La Fuente, C. Godard, E. Zangrando, C. Claver, S. Castillon, Chem. Commun. 2012, 48, 1695 – 1697. [9] M. Papp, R. Skoda-Fçldes, J. Mol. Catal. A 2013, 378, 193 – 199. [10] S. Zheng, Y. Wang, C. Zhang, J. Liu, C. Xia, Appl. Organomet. Chem. 2014, 28, 48 – 53. [11] a) A. Yamamoto, F. Ozawa, K. Osakada, L. Huang, T. Son, N. Kawasasku, M- K. Doh, Pure Appl. Chem. 1991, 63, 687 – 696; b) L. Huang, F. Ozawa, A. Yamamoto, Organometallics 1990, 9, 2603 – 2611; c) F. Ozawa, H. Soyama, H. Yanagihara, I. Aoyama, H. Takino, K. Izawa, T. Yamamoto, A. Yamamoto, J. Am. Chem. Soc. 1985, 107, 3235 – 3245. [12] F. Ozawa, N. Kawasaki, H. Okamoto, T. Yamamoto, A. Yamamoto, Oraganometallics 1987, 6, 1640 – 1651. [13] a) M. Garc a-Melchor, A. A. C. Braga, A. Lleds, G. Ujaque, F. Maseras, Acc. Chem. Res. 2013, 46, 2626 – 2634; b) M. P rez-Rodr guez, A. A. C. Braga, A. R. deLera, F. Maseras, R. Alvarez, P. Espinet, Organometallics 2010, 29, 4983 – 4991; c) B. Fuentes, M. Garc a-Melchor, A. Lleds, F. Maseras, J. A. Casares, G. Ujaque, P. Espinet, Chem. Eur. J. 2010, 16, 8596 – 8599; d) M. P rez-Rodr guez, A. A. C. Braga, M. Garc a-Melchor, M. H. P rez-Temprano, J. A. Casares, G. Ujaque, A A. R. deLera, R. Alvarez, F. Maseras, P. Espinet, J. Am. Chem. Soc. 2009, 131, 3650 – 3657; e) A. A. C. Braga, G. Ujaque, F. Maseras, Organometallics 2006, 25, 3647 – 3658; f) A. A. C. Braga, N. H. Morgon, G. Ujaque, F. Maseras, J. Am. Chem. Soc. 2005, 127, 9298 – 9307. [14] a) M. Barsukova, N. V. Izarova, R. N. Biboum, B. Keita, L. Nadjo, V. Ramachandran, N. S. Dalal, N. S. Antonova, J. J. Carb, J. M. Poblet, U. Kortz, Chem. Eur. J. 2010, 16, 9076 – 9085; b) E. Serrano, R. Navarro, T. Soler, J. J. Carb, A. Lleds, E. P. Urriolabeitia, Inorg. Chem. 2009, 48, 6823 – 6834; c) V. Lillo, E. Mas-Marza, A. M. Segarra, J. J. Carb, C. Bo, E. Peris, E. Fernndez, Chem. Commun. 2007, 32, 3380 – 3382; d) E. Serrano, C. Valles, J. J. Carb, A. Lleds, T. Soler, R. Navarro, E. P. Urriolabeitia, Organometallics 2006, 25, 4653 – 4664. [15] a) K. Vikse, T. Naka, J. S. McIndoe, M. Besora, F. Maseras, ChemCatChem 2013, 5, 3604—3609; b) M. Garc a-Melchor, M. C. Pacheco, C. Njera, A. Lleds, G. Ujaque, ACS Catal. 2011, 1, 135 – 144; c) A. A. C. Braga, G. Ujaque, F. Maseras, Organometallics 2006, 25, 3647 – 3658; d) L. J. Goossen, D. Koley, H. L. Hermann, W. Thiel, Organometallics 2005, 24, 2398 – 2410. [16] a) Y. Hu, J. Liu, Z. L, X. Luo, H. Zhang, Y. Lan, A. Lei, J. Am. Chem. Soc. 2010, 132, 3153 – 3158; b) S. A. Macgregor, G. W. Neave, Organometallics 2003, 22, 4547 – 4556; c) K. E. Frankcombe, K. J. Cavell, B. F. Yates, R. B. Knott, Organometallics 1997, 16, 3199 – 3206. [17] A. L. Casado, P. Espinet, Organometallics 1998, 17, 954 – 959. [18] a) S. M. A. Donald, S. A. Macgregor, V. Settels, D. J. Cole-Hamilton, G. R. Eastham, Chem. Commun. 2007, 0, 562 – 564; b) E. Zuidema, C. Bo, P. W. N. M. van Leeuwen, J. Am. Chem. Soc. 2007, 129, 3989 – 4000; c) P. W. N. M. van Leeuwen, M. A. Zuideveld, B. H. G. Swennenhuis, Z. Freixa, P. C. J. Kamer, K. Goubitz, J. Fraanje, M. Lutz, A. L. Spek, J. Am. Chem. Soc. 2003, 125, 5523 – 5539. [19] a) S. Aguado-Ullate, J. J. Carb, O. Gonzlez-del Moral, A. Mart n, M. Mena, J. M. Poblet, C. Santamar a, Inorg. Chem. 2011, 50, 6269 – 6279; b) E. Sicilia, N. Russo, J. Am. Chem. Soc. 2002, 124, 1471 – 1480. [20] For transition state TS4 b-5 b, we assumed that the acyl and iodide ligands are in cis positions to allow DBU to remove simultaneously the proton of the amine and the iodide ligand. Consequently, TS4 b-5 b, in fact, connects to the intermediate cis-[Pd(DBU)(CO){C(O)Ph)}I] (4 b). This species is only about 1 kcal mol1 higher in energy than 4 a and we expect easy trans-to-cis isomerisation for these four-coordinate PdII complex (see refs. [16c] and [17] and NMR data in ref. [8]). For the sake of simplicity, we did not add intermediate 4 b to the energy profiles. Analogous reasoning can be applied to transition state TS2 c-3 c. [21] Gaussian 09, Revision A.02, M. J. Frisch, 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, 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, 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. Moro-
10988
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Paper
[22] [23] [24]
[25]
kuma, 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, and D. J. Fox, Gaussian Inc., Wallingford, CT, 2009. Y. Zhao, D. Truhlar, Theor. Chem. Acc. 2008, 120, 215 – 241. P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299 – 310. A. Hçllwarth, M. Bçhme, S. Dapprich, A. W. Ehlers, A. Gobbi, V. Jonas, K. F. Kçhler, R. Stegmann, A. Veldkamp, G. Frenking, Chem. Phys. Lett. 1993, 208, 237. a) M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley, M. S. Gordon, D. J. Defrees, J. A. Pople, J. Chem. Phys. 1982, 77, 3654; b) W. J. Hehre, R.
Chem. Eur. J. 2014, 20, 10982 – 10989
www.chemeurj.org
Ditchfield, J. A. Pople, J. Chem. Phys. 1972, 56, 2257; c) P. C. Hariharan, J. A. Pople, Theor. Chim. Acta. 1973, 28, 213. [26] P. Pulay, G. Fogarasi, J. Chem. Phys. 1992, 96, 2856 – 2860. [27] B. Mennucci, E. Canc s, J. Tomasi, J. Phys. Chem. B 1997, 101, 10506 – 10517.
Received: April 10, 2014 Published online on July 22, 2014
10989
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
