ARTICLE pubs.acs.org/Organometallics
Mapping Palladium Reduction by Carbon Monoxide in a Catalytically Relevant System. A Novel Palladium(I) Dimer Fabio Ragaini,*,† Heros Larici,† Martino Rimoldi,† Alessandro Caselli,† Francesco Ferretti,† Piero Macchi,b,c and Nicola Casatic †
Dipartimento di Chimica Inorganica, Metallorganica e Analitica “Lamberto Malatesta” and ISTM-CNR, Universita degli Studi di Milano, Via Venezian 21, 20133 Milano, Italy b Department of Chemistry and Biochemistry, University of Bern, Freiestrasse 3, CH3012 Bern, Switzerland c Dipartimento di Chimica Strutturale e Stereochimica Inorganica, Universita degli Studi di Milano, Via Venezian 21, 20133 Milano, Italy
bS Supporting Information ABSTRACT: Reaction of neutral palladium(II) complexes with chelating nitrogen ligands of the phenanthroline family had been earlier found to proceed through the formation of a CO adduct, which is then reduced to another observable complex before decomposing to metallic palladium. We have now extended this study and completely characterized by single-crystal X-ray diffraction one member of this class of compounds. The intermediate is an unprecedented type of palladium(I) dimer with two bridging COs. The same complex could also be obtained by a conproportionation reaction of a Pd0 with a PdII complex. The picture of the reactivity of neutral palladium(II) complexes with phenanthroline ligands in a CO atmosphere was completed by the identification of two byproducts of the main reaction.
’ INTRODUCTION In recent years, an enormous number of new applications of palladium-based homogeneous catalysts have been reported. However, palladium applications to carbonylation reactions have increased at a much slower pace. This is surely mostly due to the notorious easiness with which metallic palladium is generated from palladium(II) complexes under a CO atmosphere. Understanding the early stages of palladium reduction and aggregation is essential in designing more effective catalytic systems, but only in a few cases have intermediate complexes in the reduction process been unequivocally characterized. We have recently reported that reversible coordination of CO to palladium complexes of the kind Pd(RPhen)X2 (1, RPhen = 1,10phenanthroline or a substituted phenanthroline, X = COOMe or a carboxy group) to initially give an adduct of composition Pd(RPhen)X2(CO) (2) is an essential step in the palladium/ phenanthroline-catalyzed carbonylation of nitroarenes to carbamates and ureas.1 Although only kinetic evidence could be obtained for these adducts in several cases, use of 2,9-dimethylphenanthroline (neocuproine, Neoc) allowed the spectroscopic observation of these compounds and the isolation of single crystals of the chloride analogue Pd(Neoc)Cl2(CO) (2a), suitable for X-ray structure determination. The structure of this and related (vide infra) complexes is not discussed in the present paper. It is worth recalling that 2a is not a truly pentacoordinate complex, but is better viewed as square planar, with distortion toward a square pyramid due to a fifth interaction occurring with one of the nitrogen atoms of the neocuproine ligand. More recently, an X-ray diffraction experimental study on the electron density of this complex has been performed, and the results are r 2011 American Chemical Society
discussed elsewhere.2 None of the complexes of type 2 were indefinitely stable in solution, and they generally evolved to metallic palladium in a time frame ranging from less than one hour to a couple of days at room temperature. Only the chloride complex 2a was stable enough to allow the growth of single crystals at low temperature. However, during the decomposition, a third type of complex (3) was observed in all cases. Type 3 complexes are characterized by an IR absorption in the range 18771884 cm1, typical of bridging CO groups, indicating they are dimers or higher aggregates. No IR absorption is observed in the region of terminal CO groups. Despite the fact that no complex of this class could be isolated or even observed alone in solution, a low-temperature NMR study of the reaction of Pd(Neoc)(TMB)2 (1b) (TMB = 2,4,6-trimethylbenzoate) with CO allowed the attribution of a set of signals to the corresponding 3b complex. These signals indicate that it contains one neocuproine and one trimethylbenzoate group per palladium atom. The neocuproine ligand is in a symmetrical environment, although this may be due to fluxionality. Together, this information indicates that complexes 3 are palladium(I) complexes of general composition [Pd(Neoc)(X)]n(μ-CO)m, where n g 2 and m g 1, but nothing more than this can be said (Scheme 1).
’ RESULTS AND DISCUSSION Following the studies described in the Introduction, we have made more attempts to unveil the nature of complexes 3, with the idea that knowing the structure of an intermediate in an Received: February 7, 2011 Published: March 30, 2011 2385
dx.doi.org/10.1021/om200118v | Organometallics 2011, 30, 2385–2393
Organometallics Scheme 1
unwanted reaction may help in avoiding it. Since the chloride complex 1a has a lower tendency to evolve toward metallic palladium than all complexes having COOMe or a carboxylate as the anionic ligand, we decided to investigate this complex and its bromide (1c) and iodide (1d) analogues in more detail.3 We have previously reported that by placing a suspension of 1a in CH2Cl2 under a CO atmosphere at room temperature, the compound gradually dissolves and an absorption at 2133 cm1 appears in the infrared spectrum of the solution, due to Pd(Neoc)Cl2(CO) (2a). At the same time, a weaker absorption at 1875 cm1 also appeared. By prolonging the reaction time, we now observed that precipitation of a yellow solid starts after several hours. The IR spectrum of the solid, in Nujol, showed a strong absorbance at 1875 cm1, coincident with the band observed in solution for 3a, accompanied by a weak absorption at 1921 cm1 and a very weak one at 1846 cm1. The precipitation is complete in about four days, after which the solution has become colorless and no more absorptions are observed in the carbonyl region of the IR spectrum. The precipitate at this stage showed the two aforementioned bands, but a weak band was also observed at 2143 cm1. If the suspension was left in a CO atmosphere for two weeks, two new bands at 1900 (m) and 1966 (vw) cm1 also appeared, which were not observed for shorter reaction times. The 2143 cm1 absorption is very close to that reported in the literature for [NH2Et2][PdCl3(CO)] (2146 cm1)46 and may be attributed to the same anionic complex, whereas the last two bands and their relative intensity are consistent with those of [Pd2Cl4(μ-CO)2]2 (1906, 1966 cm1 in CH2Cl2 solution, 1896 and 1903 cm1 in Nujol for two independent molecules in the crystals of the Bu4Nþ salt).7 Once precipitated, 3a was insoluble in most solvents and only a little soluble in DMSO. An 1H NMR spectrum of the residue in this solvent showed the presence of only one set of signals due to a neocuproine ligand in a symmetrical environment. As for 3b, this may be due to fluxionality. Signals due to the minor complexes observed by IR could not be assigned because they are too weak. Attempts to purify 3a or to get crystals suitable for X-ray diffraction failed. Thus we started investigating the corresponding bromide complexes. Carbonylation of Pd(Neoc)Br2 (1c) proceeds analogously to that of 1a, and the initial Pd(Neoc)Br2(CO) 2c adduct (νCO = 2124 cm1) was even more stable. Indeed, no peak in the bridging CO region is initially observable in the IR spectrum of the solution, and only after several hours does a very weak peak at 1887 cm1 appear, attributable to 3c. This peak always remained very weak due to the very low solubility of 3c in CH2Cl2, which precipitated as the reaction proceeded. The reaction was continued at RT for several days,
ARTICLE
Table 1. Crystallographic Data for Compound 4c formula
C14H13N2, CBr3OPd
fw
583.37
cryst syst
triclinic
space group
P1 (No. 2)
a, b, c [Å] R, β, γ [deg]
7.4271(1), 9.2217(1), 13.5184(2) 83.746(1), 76.144(1), 71.549(1)
V [Å3]
852.18(2)
Z
2
D(calc) [g/cm3]
2.273
μ(Mo KR) [mm1]
8.123
F(000)
552
cryst size [mm]
0.25 0.14 0.12
Data Collection temperature (K)
100(2)
radiation [Å]
Mo KR 0.71069
θ min., max. [deg]
1.5, 48.5
hkl
15: 15; 19: 18; 28: 27
tot, uniq data, Rint
49 480, 15 557, 0.025
obsd data [I > 2.0σ(I)]
11 846
Refinement Nref, Npar R1, wR2, S
15 557, 203 0.0575, 0.1039, 1.02
R1, wR2 (I > 2.0σ(I))
0.0364, 0.0937
max. and av shift/error
0.012, 0.001
min. and max. resd dens [e Å3]
3.15, 2.90
after which the solution was colorless and no carbonyl absorption was observable by IR. The yellow residue obtained showed in the IR spectrum in Nujol the same pattern as that observed for 3a, that is, an intense IR absorption at 1878 cm1, accompanied by a weak absorptions at 1922 cm1 and a very weak one at 1846 cm1. The elemental analysis of the solid agrees with a composition of the [Pd(Neoc)Br(CO)]n type. Again this is consistent with a dimer of composition [Pd(Neoc)Br]2(μ-CO)2, but higher aggregates cannot be excluded. During the reaction, samples were withdrawn periodically and layered with hexane. From these attempts, crystals of 2c and of [PdBr3(CO)][NeocH] (4c) suitable for X-ray diffraction were grown. The X-ray structure of 2c resembles that previously reported for the corresponding chloride analogue 1c and will be described in a forthcoming paper together with an electron density study of type 2 halide complexes including the X-ray structure of the iodide analogue (2d) (vide infra). The structure of the anionic part of 4c (Figure 1) is very similar to that previously reported for the anionic part of [PdBr3(CO)][NBu4]5 or [PdCl3(CO)][NBu4]5 and not worth further discussion. However, this structural determination unambiguously shows that compounds of composition [PdX3(CO)] are indeed formed under these conditions and supports the assignment of the 2143 cm1 absorption observed in the case of the chloride complex to [PdCl3(CO)]. Compound 4c should be formed only in very small amounts because no IR absorption around 2120 cm1 (as expected from the literature for a solution in CH2Cl2)5 could be observed either in solution during the reaction or in the Nujol mull of the precipitated product. The iodide complex Pd(Neoc)I2 (1d) was also successfully carbonylated to Pd(Neoc)I2(CO) (2d) (ν(CO) = 2101 cm1 in CH2Cl2). Complex 2d is the most stable in the series, and only a 2386
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics
ARTICLE
Scheme 2. Possible Structures for Complex 2e
Figure 1. Ortep drawing of [PdBr3(CO)][NeocH] (4c) with ellipsoids drawn at the 30% probability level (H atoms are simple spheres).
very weak absorption at 1889 cm1 attributable to 3d was observable even after several days standing in a CO atmosphere. Compounds 2a,c,d were also characterized by 1H NMR spectroscopy. However, at atmospheric pressure the carbonylation reaction proceeds slowly and some starting material was still observable when formation of byproducts began, especially in the case of 1a. In order to get clean spectra, a few milligrams of each of the starting palladium complexes was placed in a test tube with a stirring bar and treated with 1 mL of CDCl3. The test tubes were placed inside an autoclave and stirred under 30 bar of CO for 2 h at RT. The autoclave was then vented, and the 1H NMR spectrum of the three solutions immediately recorded. This procedure resulted in a fast enough carbonylation to avoid byproduct formation, and clean spectra of the adducts 2a,c,d were obtained. They all show signals attributable to a single neocuproine ligand in a symmetrical environment, evidencing that the compounds are fluxional at room temperature, as the two NPd distances are not equivalent in the solid state. After recording the 1H NMR spectra, the three solutions were also analyzed by IR spectroscopy, to confirm that the compounds of which the NMR spectrum had been recorded are indeed 2a,c,d. The IR spectra confirm the identity of the obtained compounds, although some decomposition is already evident in the spectrum of the chloride derivative. It is worth evidencing the different chemical stability of the three CO adducts. The chloride adduct 2a is only moderately stable and fully decomposes in solution at room temperature in several hours. Indeed, the single crystals of this compound employed in the previously published paper had to be grown at 50 °C.1 On the other hand, the bromide derivative 2c is stable under a CO atmosphere at room temperature for several days. The iodide compound 2d is the most stable. Its solutions can be evaporated in vacuo, and the compound redissolved under a dinitrogen atmosphere without any evidence of CO loss. The higher stability of the iodide derivative may be important to further study the reactivity of this kind of compound. It is worth mentioning that this stability order is opposite of that observed for several halocarbonyl derivatives of palladium (and even of platinum and gold) not containing other donating ligands.5,8,9 The reasons for this reversal of stability may be multifold and cannot be discussed here, but this observation may be relevant to the catalytic behavior of complexes of this class. Despite several attempts, we could not grow single crystals of any compounds of type 3, but a solution to this problem was found using 6,60 -dimethyl-2,20 -bipyridine (Me2Bipy) as a ligand in place of neocuproine. The use of Me2Bipy was originally motivated by the desire to better understand the nature of the longer PdN interaction in complex 2a. Indeed, in the case of
neocuproine this interaction may be a true bonding interaction, albeit very weak, or otherwise the nitrogen atom may be held close to palladium just because of the rigidity of the ligand itself. On the other hand, Me2Bipy is more flexible, and if no PdN bond were present, then the ligand should tilt for steric reasons (structure 2e00 in Scheme 2). The iodide complex was selected because it gives the most stable adduct of type 2 in the case of neocuproine. Complex 2e proved to be less stable than its neocuproine analogues. Despite that its formation could be clearly observed in solution by IR (νCO = 2100 cm1) when conducting the carbonylation of 1e in CH2Cl2 under the same conditions as for the other complexes, the absorption remained weak, suggesting that the equilibrium between 1e and 2e is not completely shifted to the right at atmospheric CO pressure. Formation of 3e (νCO = 1889 cm1) was also evident after just 30 min. Since all attempts to grow crystals of 2e from several solutions obtained by slight variations of the reaction conditions failed, we considered an alternative high CO pressure strategy. In order to shift the 1e2e equilibrium to the right and at the same time grow crystals under these conditions, we placed some 1e in a test tube with a little magnetic stirrer and dissolved it in a small amount of CH2Cl2. The test tube was placed inside an autoclave (a drilled aluminum block was used as an internal support), and the rest of the autoclave was flooded with hexane. The autoclave was charged with CO (30 bar), and the solution stirred for 2 h at RT, a time deemed to be sufficient to convert all 1e into 2e under these conditions. At this point stirring was interrupted and the autoclave was left undisturbed for 15 days, to allow diffusion of hexane vapors into the test tube. After this time, the autoclave was vented and the formed solid examined. The residue was composed of a mixture of 2e and 3e (by IR), although some metallic palladium was also clearly visible. Though the solid was mostly present as a powder, some single crystals were present, suitable for X-ray diffraction. The structure was solved and it is shown in Figure 2: it clearly corresponds to a member of the class of compounds 3, as also confirmed by an IR spectrum measured by ATR on a single crystal (1924 (mw), 1875 (s), 1850 (vw) cm1). In accord with all the collected data, 3e is a palladium(I) complex having only one chelating ligand and one halide group per palladium atom, with two bridging COs. To the best of our knowledge, no complex with this composition has been previously reported in the literature,10,11 and for sure none have been characterized by X-ray diffraction. Given the close similarity in the IR spectra of all compounds 3, it is clear that all members of the series have the same dimeric structure, and therefore higher aggregates need not to be invoked. At this stage, the two most intense absorptions in the IR spectra of compounds 3 can be respectively assigned to the asymmetric (around 1880 cm1) and symmetric (around 1920 cm1) stretching modes of the two 2387
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics
ARTICLE
Table 3. Principal Bond Distances in 3ea
a
Figure 2. Ortep drawing of [Pd(Me2Bipy)I]2(μ-CO)2 (3e) with ellipsoids drawn at the 30% probability level (H atoms are simple spheres).
Table 2. Crystallographic Data for Compound 3e formula
C26H24I2N4O2Pd2
fw
891.09
cryst syst
monoclinic
space group
P21/c (No. 14)
a, b, c [Å]
14.078(4) 10.466(3) 19.686(5)
R, β, γ [deg] V [Å3]
90 90.602(3) 90 2900.4(14)
Z
4
D(calc) [g/cm3]
2.041
μ(Mo KR) [ mm1 ]
3.398
F(000)
1688
cryst size [mm]
0.2 0.15 0.1
Data Collection temperature (K) radiation [Å]
293(2) Mo KR 0.71069
θ min., max. [deg]
2.1, 29.3
hkl
19: 19 ; 14: 13 ; 26: 25
tot, uniq data, Rint
25 041, 7353, 0.052
obsd data [I > 2.0σ(I)]
5018
Refinement Nref, Npar
7353, 325
R1, wR2, S R1, wR2 (I > 2.0σ(I))
0.0659, 0.0978, 1.01 0.0371, 0.0852
max. and av shift/error
0.003, 0.00
min. and max. resd dens [e Å3]
0.54, 1.06
carbonyl groups, while the very weak absorption around 1845 cm1 is likely due to the isotopic 13CO content. The structure of 3e is characterized by a quite short Pd1Pd2 distance (2.6704(9) Å; on average, structures containing PdPd bonds have a distance of 2.822(3) Å). Around each Pd atom, there are five coordination bonds, two of them with CO bridges, one with the terminal I, and two with the chelating Me2Bipy.
bond
distance (Å)
Pd1Pd2
2.6704(9)
Pd1I1
2.7012(9)
Pd1N1
2.196(3)
Pd1N10
2.381(3)
Pd1C21
2.141(6)
Pd1C22
1.909(4)
Pd2I2
2.6995(10)
Pd2N31 Pd2N40
2.200(4) 2.395(4)
Pd2C21
1.928(4)
Pd2C22
2.068(4)
In parentheses estimated uncertainties are given.
The CO bridges are quite asymmetric, as evident from the uneven PdC distances (see Table 3) and the PdCO angles (see Figure 2). The longer PdC distances are opposed to PdI bonds. The chelating Me2Bipy is also asymmetrically coordinated, with N1/N31 occupying the site pseudosymmetric to iodide, whereas N10/N40 atoms occupy an unusual fifth coordination, at a PdN distance of ca. 0.2 Å longer. As a result, 3e differs substantially from some classical M2L6L0 2 species (L being a terminal ligand and L0 a bridging ligand), where the metal is in a pseudo-octahedral coordination (with an empty site) and the MM bond is along the pseudo-3-fold axis. As additional proof, the planes including the two bridging CO's form an angle of just 36°, instead of 60°. The structure is then a (severe) distortion from a planar dimer, which is known for the only dinuclear palladium(I) complexes featuring the PdI(μ-CO)2PdI unit characterized so far, namely, [NR4]2[Pd2Cl4(CO)2] (R = Bu7 or H12), [PPN]2[Pd2Br4(CO)2]13 (PPNþ = (Ph3P)2Nþ), and [Pd2(CO)2][SO3F]2.14 It is quite instructive that the first of these compounds and some of its substituted derivatives15 are diamagnetic. This was initially attributed to the formation of a direct PdPd bond, but theoretical calculations show that this is not the case, and coupling occurs through the bridging ligand orbitals, even though the two palladium atoms are at a distance compatible with a direct metalmetal bond,15,16 similar to that of 3e. Searching in the literature, the structure of bis((μ2-2,4,6trimethylphenylisocyanido)-(hydrogen tris(pyrazol-1-yl)borato)palladium(I)) also presents pentacoordinated Pd(I) atoms;17 however the fifth coordination of a pyrazolyl is much weaker than the fifth coordination given by N10/N30 in 3e, and many other features make that structure more similar to a planar dimer (for example the two phenylisocyanido bridges lie in the same plane, at variance from CO's in 3e). Therefore, 3e is somewhat unique by virtue of the relatively strong fifth coordination. Noteworthy, the asymmetry of Pd1N1/Pd1N10 or Pd2N31/Pd2N40 bonds is much smaller than that observed in the structures of compounds 2 that we have characterized so far. Compounds 3e and 3c are too unsoluble to allow the recording of an NMR spectrum, but the 1H NMR spectra of 3a in DMSO (this paper) and 3b in CDCl31 show that they are also diamagnetic. It is important to recall that 6,60 -dimethyl-2,20 -bipyridine, neocuproine, and its subtituted derivatives have been shown to lead to higher catalytic activities with respect to phenanthroline 2388
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics Scheme 3. General Reaction Scheme
when employed as ligands for several palladium-catalyzed reactions, including the Heck reaction,18 the oxidative Heck arylation of olefins by boronic acids,1921 the production of hydrogen peroxide,22,23 and especially the alcohol oxidation by dioxygen.2428 The reason with general validity has not been unveiled for this higher reactivity; however the most frequent explanation advanced in the literature2228 is that the two ortho methyl groups inhibit or at least disfavor the formation of catalytically inactive complexes, and some experimental support has been gained for this hypothesis in some cases.29 However, the X-ray structure of 3e demonstrates that such steric hindrance does not inhibit dimerization, at least when reduction of palladium(II) to palladium(I) occurs. In fact, the two ligands assume a conformation such that the angle between their planes is sufficiently large (30°) to minimize the methylmethyl repulsion (the shorter C---C distance is 3.71 Å). Thus, a reduced tendency to dimerize cannot be the general reason for better catalytic performances. This is also in agreement with the results by Tsuji and coworkers,30,31 who reported an enhanced stability for pyridine and bipyridine palladium complexes where the steric hindrance is due to much larger substituents. Complexes 3 are clearly intermediates in the reduction by CO of palladium(II) complexes to metallic palladium, a very important process whose details are however little understood. In the case of the synthesis of [Pd2Cl4(CO)2]2, it was proposed that the reduction of the initial palladium(II) complex is due to phosgene elimination,15,32 whereas when acetate is the counteranion, elimination of CO2 and acetic anhydride has been observed.33 On the other hand, reduction by CO/H2O is a general pathway for the reduction of a variety of metals. As far as the mechanism of formation of complexes 3 is concerned, we have no evidence for the formation of phosgene (IR absorption 1810 cm1 in CH2Cl2 solution34) or its brominated analogue. Moreover no carboxylate is present in several cases, and in the only case in which a carboxylate is present and the organic products of the reduction were identified, that is, in the case of 3b (2,4,6-trimethylbenzoate is the counteranion), the observed coproduct was trimethylbenzoic acid and not the corresponding anhydride.1 Since monoelectronic reduction of palladium is very unlikely under the conditions of this study, we consider that the most likely route for the formation of 3 is that 2 is initially reduced to a palladium(0) complex by reaction with CO and trace amounts of water that cannot be completely removed from the system even if CO was passed through molecular sieves before reaching the reaction flask. The so-formed complex
ARTICLE
would be trapped by still unreacted 2 to afford the dimeric 3. Formation of a small amount of CO2 could indeed be observed by IR. The intentional addition of small amounts of water to the reaction solution also resulted in a strong acceleration of the reduction process, but in this case the reaction could not be stopped at the stage of the dimeric complexes and extensive formation of metallic palladium was observed. Scheme 3 accounts for the formation of all observed products. Coordination of CO to 1 generates 2. Reduction of the latter by CO/H2O generates an equivalent of acid. Its formation accounts for the production of 4 by reaction with 2, although the very small amount of 4 observed (and only when X = Cl or Br) indicates that this process occurs only to a small extent. Reaction of the formed halogenidric acid with 3 may account for the formation of [Pd2X4(CO)2]2. This reaction should be even more difficult since it has been observed only for X = Cl, only when the reaction mixture was left to stand for a prolonged time, and well after the formation of 3 had been completed. On the other hand, even in the case in which X is a carboxylate (apart from trimethylbenzoate, we recall that the same general reactivity was observed in the case of X = acetate, pivalate, triflouroacetate, and C(O)OMe,1 although the reaction could not be stopped at the stage of 3 in any of these cases and the minor byproducts were not observed) we could not observe any further intermediate in the reduction of 3 to metallic palladium. In particular, under our conditions no evidence could be gained for the formation of a neocuproine analogue of the known [Pd4(Phen)4(CO)2]4þ,35 which has a higher aggregation, although the formal oxidation state of palladium is still þ1. Neither could complexes of the kind [L2Pd(μ-H)(μ-CO)PdL2]þ (which are known for L2 = Phen23,36or Bipy37) be detected. A few precedents exist for the formation of palladium(I) dimeric complexes by a well-defined reaction of two different monomeric precursors, but they all involve hydride complexes as intermediates38,39 and the resulting complexes have a different structure. However, the conproportionation of a palladium(0) and a palladium(II) complex has been proposed as a possible pathway for the formation of a palladium(I) dimer not containing a bridging hydride in at least one case.40 In order to get support for the conproportionation step in our mechanism, we attempted to obtain complexes of type 3 by reaction of the two independently prepared components. Unfortunately, palladium(0) complexes of the type Pd(L)(CO)n (L = Neoc or Me2Bipy) are not stable. As a matter of fact, no compound of composition Pd0(L)(CO)n with L = chelating nitrogen ligand has ever been reported in the literature. As the closest precursors, complexes of the type Pd(L)(dba) can be easily prepared by reaction of Pd(dba)2 with one equivalent of the desired ligand.41,42 In our hands, Pd(Neoc)(dba) and Pd(Me2Bipy)(dba) evolved to metallic palladium in a few seconds when exposed to a CO atmosphere in solution, likely through the intermediate formation of the desired palladium(0) carbonyl complex. To avoid as much as possible the direct aggregation of the palladium(0) carbonyl complex, the following general procedure was devised. A solution of 2 in CH2Cl2 was prepared in the usual way by carbonylation of 1. In a separate flask, a solution of the corresponding Pd(L)(dba) complex was prepared. The latter solution was then slowly added to the former by means of a syringe pump. This way the palladium(0) carbonyl complex is formed in low concentration and in the presence of a larger amount (at least initially) of the palladium(II) one, thus maximizing the chances of conproportionation with respect to self-aggregation. 2389
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics Scheme 4. Synthesis of the Pd(I) Dimer 4e by a Conproportionation Reaction
This strategy was first applied to the synthesis of 3e. The compound was indeed formed and precipitated out of the reaction mixture, in accord with its insolubility. Despite the precautions employed, some metallic palladium was also formed, but the IR spectrum of the formed precipitate closely matched that of the single crystals employed in the diffraction study. The precipitated complex was completely insoluble in any solvent, and a separation of the coprecipitated metallic palladium was not possible. The elemental analysis of the obtained solid was far lower than required, but the ratio between the carbon, nitrogen, and hydrogen percentages was very close to that calculated for 3e, supporting the idea that the solid is essentially composed by the latter complex and metallic palladium. The same strategy was then applied to the preparation of 3a. An analogous result was obtained. The IR spectrum (in Nujol) of the obtained precipitate matches that of 3a obtained by simple carbonylation reaction, but the formation of metallic palladium could not be avoided, and an analytically pure compound could not be obtained. However, again the ratio between the carbon, nitrogen, and hydrogen percentages was very close to that calculated for pure 3a.
’ CONCLUSIONS In this paper we have shown that the formation of the previously reported Pd(Neoc)Cl2(CO) (2a) is indeed a general reaction pathway by preparing its higher halide homologues and a related 6,60 -dimethyl-2,20 -bipyridine complex. Most importantly, the identity of the next observable stage during the decomposition to metallic palladium, a palladium(I) dimer, has been unequivocally identified, and evidence for a conproportionation mechanism leading to its formation has been obtained. Thus two early stage intermediates in the reduction of a palladium(II) complex by CO have been characterized for the same system. This may help in designing new ligands that bear the right type of steric hindrance to prevent aggregation. The halogenidric acid formed during the reaction has also been found to originate two minor byproducts, [Pd(CO)3X] and [Pd2X4(CO)2]2, thus completing the picture of the reactivity of the investigated compounds in a CO atmosphere. ’ EXPERIMENTAL SECTION General Procedures. Unless otherwise stated, all reactions and manipulations were conducted under a dinitrogen atmosphere. Carbon monoxide employed in the reaction at atmospheric pressure was passed though molecular sieves before reaching the reaction flask, but no purification could be done on the CO employed in the high-pressure
ARTICLE
reactions. All solvents were dried by standard procedures and distilled under dinitrogen immediately before use. All glassware and magnetic stirring bars used in catalytic reactions were kept in an oven at 120 °C for at least two hours and allowed to cool under vacuum before use. 2,9Dimethyl-1,10-phenanthroline was purchased as a hydrate. It was dried by dissolving it in CH2Cl2, drying the resulting solution with Na2SO4, filtering the suspension under dinitrogen, and evaporating in vacuo the filtered solution. It was then stored under dinitrogen. It can be weighed in the air without problems, but must be stored in an inert atmosphere if water uptake is to be avoided. Pd(Neoc)Cl2,1,43 Pd(Neoc)Br2,43 Pd(Neoc)I2,43 and Pd(dba)244,45 were prepared and characterized according to the procedures reported in the literature. All other chemicals were purchased from Aldrich, Acros, or Alfa Aesar and used as received. CDCl3 was purified by passing through a short column of basic alumina that had been previously dried by heating in vacuo by a heating gun (to remove all acidic impurities and most of the water). The so-purified solvent was degassed and stored over activated molecular sieves under dinitrogen and in the dark. d6-DMSO was distilled over CaH2, degassed, and stored over activated molecular sieves under dinitrogen. NMR spectra were recorded on a Bruker AC300 FT or on an Avance Bruker DPX300 spectrometer. Unless otherwise noted, IR spectra were recorded on a Varian Scimitar FTS 1000 FT-IR spectrophotometer. Elemental analyses were recorded on a PerkinElmer 2400 CHN elemental analyzer. Synthesis of Pd(Me2Bipy)I2 (1e). The complex was prepared by adapting a procedure reported for the corresponding phenanthroline and bipyridine complexes.46 Pd(OAc)2 (111 mg, 0.494 mmol), NaI (179 mg, 1.20 mmol), and 6,60 -dimethylbipyridine (Me2Bipy, 129 mg, 0.700 mmol) were placed in a 50 mL Schlenk flask under dinitrogen, and methanol (10 mL) was added. The suspension was stirred for 30 min at RT. The black solid was separated by filtration in the air with a Buchner, washed with a 1:1 MeOH/H2O mixture, and dried in vacuo (254 mg, 0.467 mmol, 94.4% yield). 1H NMR (CDCl3): δ 8.2 (d, 2H, J = 7,8 Hz), 7.7 (pt, 2H, J = 7,7 Hz), 7.17 (d, 2H, J = 7,7 Hz), 2.65 (s, 6H) ppm. Anal. Calcd for C12H12I2N2Pd: C, 26.47; H, 2.22; N, 5.15. Found: C, 26.10; H, 1.91, N, 5.22. Carbonylation Reactions at Atmospheric Pressure. Pd(Neoc)Cl2 (1a). The reaction has been performed several times, varying the sampling time and the total reaction time. A reaction aimed at isolating Pd(Neoc)Cl2(CO) (2a) has been described in a previous paper.1 In a typical preparation, Pd(Neoc)Cl2 (13.9 mg, 3.60 102 mmol) was placed in a Schlenk tube under a CO atmosphere and suspended in CH2Cl2 (5 mL). The initially yellow suspension gradually turned into a yellow solution. Stirring was continued for 4 days, after which a pale yellow precipitate was separated by centrifugation from the colorless solution and washed twice with CH2Cl2. IR (Nujol): 2143 (vw; the intensity of this band relative to the others is variable from one preparation to the other), 1921 (w), 1875 (s), 1846 (vw) cm1 (see also Figure S1, Supporting Information). Attempts to remove the impurity failed, and the elemental analysis of the solid is out of range for [Pd(Neoc)Cl]2(μ-CO)2 (3a). Anal. Calcd for C30H24Cl2N4O2Pd2: C, 47.64; H, 3.20; N, 7.41. Found: C, 46.31; H, 3.47; N, 7.18. The precipitate is insoluble in CDCl3, but soluble enough in DMSO to allow recording of a 1H NMR spectrum. 1H NMR (d6-DMSO, RT): δ 8.74 (d, 2H, H4, H7 or H3, H8, J = 8.26 Hz); 8.17 (s, 2H, H5 and H6); 8.02 (d, 2H, H3, H8 or H4, H7, J = 8.30 Hz); 3.22 (s, 6H, CH3). The reaction was also performed with the same amounts, but for 14 days instead of 4. The IR spectrum (Nujol) of the formed precipitate showed, in addition to those mentioned above, two additional bands at 1900 (ms) and 1966 (w) cm1 (see also Figure S2, Supporting Information). Pd(Neoc)Br2 (1c). The reaction has been performed several times, varying the sampling time and the total reaction time. In a typical 2390
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics preparation, Pd(Neoc)Br2 (1c, 41.0 mg, 8.64 102 mmol) was placed in a Schlenk tube under a CO atmosphere and suspended in CH2Cl2 (5 mL). An IR absorption immediately start to grow at 2120 cm1 and the initially orange suspension gradually turned into an orange solution. Stirring was continued for 5 days. During this time, samples of the solution were withdrawn and layered with hexane. From these attempts, single crystals suitable for X-ray diffraction analysis were obtained for Pd(Neoc)Br2(CO) (2c) and [PdBr3(CO)][NeocH] (4c). After the allotted time was finished, a yellow precipitate was separated by centrifugation from the almost colorless solution and washed twice with CH2Cl2. The solid was analytically pure [Pd(Neoc)Br]2(μ-CO)2 (3c). Anal. Calcd for C30H24Br2N4O2Pd2: C, 42.63; H, 2.86; N, 6.63. Found: C, 42.75; H, 3.25; N, 6.43. IR (Nujol): 1922 (w), 1877 (s), 1847 (vw) cm1 (see also Figure S3, Supporting Information). Pd(Neoc)I2 (1d). Complex 1d was also carbonylated according to the protocol described for 1a,c. An IR absorption immediately started to grow at 2103 cm1, due to 2d. A very weak absorption is also observable after some time at 1889 cm1, due to 3d. However, after having reached a maximum after several hours, in correspondence with the complete dissolution of the initially dark violet suspension, the intensity of the former absorption did not decrease even after several days, and no precipitate formed. Drying in vacuo the solution did not result in CO loss, and the dark violet residue was redissolved in CH2Cl2 and layered with hexane to give crystals of 2d suitable for X-ray diffraction. Pd(Me2Bipy)I2 (1e). Complex 1e (20.0 mg, 3.01 102 mmol) was also carbonylated in CH2Cl2 (5 mL) according to the protocol described for 1a,c. A IR absorption immediately started to grow at 2100 cm1, due to 2d. A very weak absorption was also observable after 30 min at 1889 cm1. However, the intensity even of the former band remained weaker than in the other cases, and the initial suspension did not completely dissolve before formation of metallic palladium started to be observable after several hours. Carbonylation Reactions under High CO Pressure. The reactions were conducted in parallel in three 10 mm wide 40 mm high test tubes, each having a magnetic stirring bar, which were located in the holes of an aluminum block designed to fit a 200 mL stainless steel autoclave. The block and the test tubes were placed inside a Schlenk tube with a wide mouth, and compounds 1a,c,d (1.8 102 mmol each) and dry and degassed CDCl3 (1 mL to each test tube) were added under dinitrogen. Each tube was closed with a screw cap with a glass wool-filled open mouth that allows gaseous reagents to exchange, and the block was rapidly transferred to the autoclave. The autoclave was purged from air, charging CO at 30 bar and discharging to 2 bar one time before performing the reaction. CO (30 bar) was then charged at room temperature, and stirring was continued for 2 h at RT. After this time, the autoclave was vented, the solutions were rapidly transferred to three NMR tubes under a CO atmosphere, and the 1H NMR spectra recorded immediately. After the NMR spectrum had been recorded, an IR spectrum of the same solution was also recorded, to confirm that the two spectra are attributed to the same species. In all cases, a single species was observable by 1H NMR. The IR spectra of the carbonylated samples of 1c,d also showed the exclusive presence of 2c,d (the single band is only slightly shifted on passing from CH2Cl2 to CDCl3 as solvent), although some decompositions is evident in the spectrum of the 2a. This must have occurred in the time frame from the recording of the NMR and IR spectra, since the NMR spectrum is clearly due to a single species. 2a: 1H NMR (CDCl3, RT): δ 8.34 (d, 2H, H4, H7 or H3, H8, J = 8.4 Hz); 7.86 (s, 2H, H5 and H6); 7.55 (d, 2H, H3, H8 or H4, H7, J = 8.4 Hz); 3.27 (s, 6H, CH3) (see also Figure S4, Supporting Information). IR (CDCl3): ν(CO) 2133 cm1 (see also Figure S5, Supporting Information).
ARTICLE
2c: 1H NMR (CDCl3, RT): δ 8.29 (d, 2H, H4, H7 or H3, H8, J = 8.2 Hz); 7.83 (s, 2H, H5 and H6); 7.71 (d, 2H, H3, H8 or H4, H7, J = 8.2 Hz); 3.46 (s, 6H, CH3) (see also Figure S6, Supporting Information). IR (CDCl3): ν(CO) 2124 cm1 (see also Figure S7, Supporting Information). 2d: 1H NMR (CDCl3, RT): δ 8.28 (d, 2H, H4, H7 or H3, H8, J = 8.2 Hz); 7.84 (s, 2H, H5 and H6); 7.69 (d, 2H, H3, H8 or H4, H7, J = 8.2 Hz); 3.32 (s, 6H, CH3) (see also Figure S8, Supporting Information). IR (CDCl3): ν(CO) 2105 cm1 (see also Figure S9, Supporting Information). Crystallization of 3e under High CO Pressure. The reaction has been performed in the same apparatus described for the other highpressure carbonylation, but using an open test tube. Compound 1e (7.0 mg, 1.3 102 mmol) was placed in a test tube that can fit one of the holes of the aluminum block, and CH2Cl2 (2 mL) was added. The rest of the autoclave was flooded with hexane (30 mL), taking care that no hexane entered the test tube. The autoclave was charged with CO (30 bar), and the solution stirred for 2 h at RT. At this point stirring was interrupted and the autoclave was left undisturbed for 15 days, to allow diffusion of hexane vapors into the test tube. After this time, the autoclave was vented and the formed solid examined. A few crystals of 3e suitable for X-ray diffraction analysis were present, together with powdery material and metallic palladium. The IR spectrum of a single crystal was recorded on an Thermo-Nicolet Avatar 360 FT-IR equipped with an ATR accessory for the recording of the spectra of single crystals. IR (ATR): ν(CO) 1924 (m), 1875 (s), 1849 (mw) cm1 (see also Figure S10, Supporting Information). Synthesis of 3a,e by a Conproportionation Reaction. 3e: In a Schlenk flask under a dinitrogen atmosphere, Pd(Me2Bipy)(dba) was prepared according to a procedure reported in the literature.41,42 Pd(dba)2 (158 mg, 0.276 mmol) was dissolved in CH2Cl2 (15 mL), and a solution of Neoc (57.4 mg, 0.276 mmol) in CH2Cl2 (4 mL) was slowly added by an equilibrated dropping funnel. The violet suspension turned into a deep red solution, which was stirred for a further 45 min to complete the reaction. In a separate Schlenk flask under a CO atmosphere, [Pd(Me2Bipy)I2] (1e) (100.0 mg, 0.184 mmol) was suspended in CH2Cl2 (10 mL) and stirred for 5 min. The Pd(Me2Bipy)(dba) solution was transferred to a syringe and slowly added to the other flask by means of a syringe pump. An immediate reaction was observed as the two solutions contacted each other, leading to the formation of an almost black precipitate. Stirring under CO was continued for 2 h after the end of the addition, after which a red solution and a black precipitate were present. The IR spectrum of the solution showed a band at 2100 cm1 due to 2e. The IR spectrum (in Nujol) of the precipitate is indistinguishable from that of 3e obtained by the method described before (there is a 1 cm1 shift on passing from the ATR measurement used for the single crystal to the KBr pellet employed for the present reaction). Anal. Calcd for C26H24I2N4O2Pd2: C, 35.04; H, 2.71; N, 6.29. Anal. Calcd for C26H24I2N4O2Pd2 3 7Pd: C, 19.09; H, 1.48; N, 3.42. Found: C, 18.64; H, 1.50; N, 3.17. Once precipitated, 3e was insoluble in any common organic solvent, and excess metallic palladium could not be removed. 3a: In a Schlenk flask under a CO atmosphere, [Pd(Neoc)Cl2] (1a) (25.6 mg, 0.066 mmol) was suspended in CH2Cl2 (6 mL) and stirred for 5 h. In the meanwhile, in a separate Schlenk flask under a dinitrogen atmosphere, Pd(Neoc)(dba) was prepared according to a procedure reported in the literature.41,42 Pd(dba)2 (38.2 mg, 0.066 mmol) was dissolved in CH2Cl2 (10 mL), and a solution of Neoc (13.8 mg, 0.066 mmol) in CH2Cl2 (2 mL) was slowly added by an equilibrated dropping funnel. The violet suspension turned into a deep red solution, which was stirred for a further 45 min to complete the reaction. The so-obtained solution was transferred to a syringe and slowly added to the first flask by means of a syringe pump. An immediate reaction was observed as the Pd(Neoc)(dba) solution contacted that containing [Pd(Neoc)Cl2(CO)], leading to the formation of a yellow-green precipitate, which was accompanied by some metallic palladium. The IR spectrum (in Nujol) of the 2391
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics precipitate is indistinguishable from that of 3a obtained by the method described before. Anal. Calcd for C30H24Cl2N4O2Pd2: C, 47.64; H, 3.20; N, 7.41. Anal. Calcd for C30H24Cl2N4O2Pd2 3 2Pd: C, 37.18; H, 2.50; N, 5.78. Found: C, 37.40; H, 2.70; N, 5.92. Once precipitated, 3a was insoluble in most organic solvents and only sparingly soluble in DMSO. Attempts to remove excess metallic palladium resulted in partial decomposition of the complex. X-ray Single-Crystal Structure Determination. Crystal samples of 3e and 4c were mounted in air on a glass fiber put on a goniometer head and then centered on the goniometer of a Bruker APEX II CCD diffractometer, equipped with an Oxford Cryosystem 700þ cryostream, with a generator operating at 50 kV and 30 mA. For species 4c, a more extensive data collection at low temperature was carried out with the intent to map the electron density distribution in NeocHþ, but the quality of the crystal was not enough for a more detailed modeling. Details of each data collection are given in Tables S1 and S2 of the Supporting Information. The raw integrated intensities of all data sets were corrected for absorption and diffraction anisotropies using SADABS.47 For 4c an analytical absorption correction was also possible. The structures were solved with direct methods using SIR9748 and refined based on fullmatrix least-squares on F2 with SHELX9749 within the WINGX package.50 Hydrogens were always modeled as riding on the corresponding carbon atoms, with a fixed isotropic thermal parameter. All other atoms were refined with anisotropic thermal parameters. Details of structure refinements are given in the Supporting Information.
’ ASSOCIATED CONTENT
bS
IR and 1H NMR spectra of complexes of type 2 and 3; crystallographic data for compounds 3e and 4a. This material is available free of charge via the Internet at http://pubs.acs.org. Supporting Information.
’ AUTHOR INFORMATION Corresponding Author
*Phone: (þ39)0250314373. Fax: (þ39)0250314405. E-mail:
[email protected].
’ ACKNOWLEDGMENT We thank the Italian MiUR (PRIN) for financial support. ’ REFERENCES (1) Ragaini, F.; Gasperini, M.; Cenini, S.; Arnera, L.; Caselli, A.; Macchi, P.; Casati, N. Chem.—Eur. J. 2009, 15, 8064–8077. (2) Macchi, P.; Casati, N.; Ragaini, F.; Sironi, A., manuscript in preparation (3) We have recently found that complexes 1a, 1c, and 1d exist in two different bond isomeric forms. These results will be described elsewhere, but the identity of the employed isomer is not influential to the present discussion, since both yielded the same products. (4) Calderazzo, F.; Dell’Amico, D. B. Inorg. Chem. 1981, 20, 1310–1312. (5) Andreini, B. P.; Dell’Amico, D. B.; Calderazzo, F.; Venturi, M. G. J. Organomet. Chem. 1988, 354, 369–380. (6) Browning, J.; Goggin, P. L.; Goodfellow, R. J.; Norton, M. G.; Rattray, A. J. M.; Taylor, B. F.; Mink, J. J. Chem. Soc., Dalton Trans. 1977, 2061–2067. (7) Goggin, P. L.; Goodfellow, R. J.; Herbert, I. R.; Orpen, A. G. J. Chem. Soc., Chem. Commun. 1981, 1077–1079. (8) Calderazzo, F. J. Organomet. Chem. 1990, 400, 303–320.
ARTICLE
(9) Belli Dell’Amico, D. B.; Labella, L.; Marchetti, F.; Samaritani, S. Coord. Chem. Rev. 2010, 254, 635–645. (10) For reviews on palladium carbonyl complexes see refs 8, 9, and 11. (11) Stromnova, T. A.; Moiseev, I. I. Russ. Chem. Rev. 1998, 67, 485–514. (12) Kulik, A. V.; Bruk, L. G.; Temkin, O. N.; Aleksandrov, G. G.; Nefedov, S. E.; Eremenko, I. L. Dokl. Chem. 2002, 383, 69–71. (13) Kulik, A. V.; Temkin, O. N.; Bruk, L. G.; Zavodnik, V. E.; Belsky, V. K.; Minin, V. V. Russ. Chem. Bull. 2005, 54, 1391–1397. (14) Wang, C.; Bodenbinder, M.; Willner, H.; Rettig, S.; Trotter, J.; Aubke, F. Inorg. Chem. 1994, 33, 779–786. (15) Baig, S.; Richard, B.; Serp, P.; Mijoule, C.; Hussein, K.; Guihery, N.; Barthelat, J. C.; Kalck, P. Inorg. Chem. 2006, 45, 1935–1944. (16) Kostic, N. M.; Fenske, R. F. Inorg. Chem. 1983, 22, 666–671. (17) Tanase, T.; Fukushima, T.; Nomura, T.; Yamamoto, Y.; Kobayashi, K. Inorg. Chem. 1994, 33, 32–39. (18) Cabri, W.; Candiani, I.; Bedeschi, A.; Santi, R. J. Org. Chem. 1993, 58, 7421–7426. (19) Enquist, P. A.; Lindh, J.; Nilsson, P.; Larhed, M. Green Chem. 2006, 8, 338–343. (20) Andappan, M. M. S.; Nilsson, P.; Larhed, M. Chem. Commun. 2004, 218–219. (21) Andappan, M. M. S.; Nilsson, P.; von Schenck, H.; Larhed, M. J. Org. Chem. 2004, 69, 5212–5218. (22) Bianchi, D.; Bortolo, R.; D’Aloisio, R.; Ricci, M. Angew. Chem., Int. Ed. 1999, 38, 706–708. (23) Querci, C.; D’Aloisio, R.; Bortolo, R.; Ricci, M.; Bianchi, D. J. Mol. Catal. A: Chem. 2001, 176, 95–100. (24) ten Brink, G.-J.; Arends, I. W. C. E.; Hoogenraad, M.; Verspui, G.; Sheldon, R. A. Adv. Synth. Catal. 2003, 345, 1341–1352. (25) Arends, I. W. C. E.; ten Brink, G.-J.; Sheldon, R. A. J. Mol. Catal. A: Chem. 2006, 251, 246–254. (26) Brink, G.-J. t.; Arends, Isabel, W. C. E.; Sheldon, Roger, A. Adv. Synth. Catal. 2002, 344, 355–369. (27) ten Brink, G. J.; Arends, I.; Hoogenraad, M.; Verspui, G.; Sheldon, R. Adv. Synth. Catal. 2003, 345, 497–505. (28) ten Brink, G. J.; Arends, I.; Sheldon, R. A. Science 2000, 287, 1636–1639. (29) Steric effects have also been generically proposed to affect the reactivity of the monomeric complex in a few cases, but the exact role of the steric hindrance has not been clarified. It is also possible that the presence of the ortho methyl groups inhibits the formation of electronically saturated and catalytically inactive monomeric complexes having two chelating ligands in the coordination sphere. However, to the best of our knowledge, this explanation has never been advanced in the literature for the ligands employed in the present study. (30) Iwasawa, T.; Tokunaga, M.; Obora, Y.; Tsuji, Y. J. Am. Chem. Soc. 2004, 126, 6554–6555. (31) Komano, T.; Iwasawa, T.; Tokunaga, M.; Obora, Y.; Tsuji, Y. Org. Lett. 2005, 7, 4677–4679. (32) Belli Dell’Amico, D.; Calderazzo, F.; Zandona, N. Inorg. Chem. 1984, 23, 137–140. (33) Stromnova, T. A.; Vargaftik, M. N.; Moiseev, I. I. J. Organomet. Chem. 1983, 252, 113–120. (34) McDermott, G. A.; Dorries, A. M.; Mayr, A. Organometallics 1987, 6, 925–931. (35) Vargaftik, M. N.; Stromnova, T. A.; Khodashova, T. S.; Porai-Koshits, M. A.; Moiseev, I. I. Koord. Khim. 1981, 7, 132–140. (36) Macchioni, A.; Romani, A.; Zuccaccia, C.; Guglielmetti, G.; Querci, C. Organometallics 2003, 22, 1526–1533. (37) Chan, S.; Siu-Ming, L.; Zhenyang, L.; Wing-Tak, W. J. Organomet. Chem. 1996, 510, 219–231. (38) Portnoy, M.; Frolow, F.; Milstein, D. Organometallics 1991, 10, 3960–3962. (39) Budzelaar, P. H. M.; Van Leeuwen, P. W. N. M.; Roobeek, C. F.; Orpen, A. G. Organometallics 1992, 11, 23–25. (40) Baya, M.; Houghton, J.; Konya, D.; Champouret, Y.; Daran, J.-C.; Almeida Le~ nero, K. Q.; Schoon, L.; Mul, W. P.; Oort, A. B. v.; 2392
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
Organometallics
ARTICLE
Meijboom, N.; Drent, E.; Orpen, A. G.; Poli, R. J. Am. Chem. Soc. 2008, 130, 10612–10624. (41) Ito, T.; Hasegawa, S.; Takahashi, Y.; Ishii, Y. J. Organomet. Chem. 1974, 73, 401–409. (42) Stahl, S. S.; Thorman, J. L.; Nelson, R. C.; Kozee, M. A. J. Am. Chem. Soc. 2001, 123, 7188–7189. (43) Plowman, R. A.; Power, L. F. Aust. J. Chem. 1971, 24, 303–308. (44) Ukai, T.; Kawazura, H.; Ishii, Y.; Bonnet, J. J.; Ibers, J. A. J. Organomet. Chem. 1974, 65, 253–266. (45) Rettig, M. F.; Maitlis, P. M. Inorg. Synth. 1977, 17, 134–137. (46) Aresta, M.; Giannoccaro, P.; Tommasi, I.; Dibenedetto, A.; Manotti Lanfredi, A. M.; Ugozzoli, F. Organometallics 2000, 19, 3879–3889. (47) Sheldrick, G. M. SADABS v 2.10; University of G€ ottingen: Germany, 2003. (48) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G. L.; Giacovazzo, C.; Guagliardi, A.; Moliterni, A. G. G.; Polidori, G.; Spagna, R. J. Appl. Crystallogr. 1999, 32, 115–119. (49) Sheldrick, G. M. SHELX97, Program for the Refinement of Crystal Structure; University of G€ottingen: Germany, 1997. (50) Farrugia, L. J. J. Appl. Crystallogr. 1999, 32, 837–838.
2393
dx.doi.org/10.1021/om200118v |Organometallics 2011, 30, 2385–2393
