Our research is centered on the design and organometallic chemistry of N-heterocyclic carbenes (NHCs) and their applications particularly in homogeneous catalysis with focus on C-C and C-heteroatom bond formation. We are especially interested in the following sub-areas:
Click directly on the sub-areas to show/hide the more details.
Click directly on the sub-areas to show/hide the more details.
1. Mixed NHC/co-ligand complexes:
NHC complexes bearing halido ligands have been investigated in great detail. On the other hand, NHC complexes bearing other anionic or neutral co-ligands remain relatively rare, although such complexes may exhibit interesting properties due to mutual interactions of the co-ligand in question with the ancillary carbenes. In this respect, we have reported a novel class of mixed dicarboxylato-bis(carbene) complexes of palladium(II). In these complexes, the Pd-carboxylato moiety can be efficiently stabilized by either mono- or dicarbenes and therefore prevent reductive decomposition and auto-ionization (ligand disproportionation) processes – two important side-reactions that lead to catalysts deactivation of phosphine-carboxylato complexes. The preferably cis-configured complexes are also highly active in the Mizoroki-Heck reaction.
We have also studied mixed NHC/phosphine and mixed NHC/N-donor complexes. It is interesting to note that the former prefer a cis-arrangement due to the transphobia effect, whereas the latter are found exclusively in trans-position.
We have also studied mixed NHC/phosphine and mixed NHC/N-donor complexes. It is interesting to note that the former prefer a cis-arrangement due to the transphobia effect, whereas the latter are found exclusively in trans-position.
2. Functionalization of NHCs:
NHCs can be modified in many ways by introducing functionalities at the nitrogen atoms of the N-heterocyclic ring. Various donor-functionalized NHCs, their complexes and their use in catalysis have been explored. The functionality offers both the possibility for hemilabile coordination and the opportunity to immobilize the resulting catalysts on polymer resins. N-functionalization with N-, O- and P-donor groups is relatively common, whereas NHCs bearing S-donors are rare. Sulfur-functionalized NHCs are potentially hemilabile ligands, which may be beneficial for certain types of catalytic applications. As a contribution to this little explored field, we have recently communicated the hemilabile coordination behavior of a methyl-aryl thioether-functionalized imidazolin-2-ylidene ligand. More recently, we have also described a Pd(II) complex bearing thiophene-functionalized benzimidazolin-2-ylidene ligands and its catalytic activiity in aequous Suzuki-Miyaura reactions.
3. Non-classical carbenes:
One major drive in carbene chemistry is the development of novel non-classical carbenes. In addition to the classical C2 bound systems (“normal” carbenes, type A; Figure 1), NHC chemistry has recently been extended to complexes of “abnormal” carbenes, which contain an unusual C4/C5 coordination (type B). The latter contain a carbenoid center that is adjacent to only one nitrogen heteroatom. More recently, Raubenheimer and co-workers reported complexes bearing N-heterocyclic carbene ligands with a 6-membered ring, in which the carbenoid carbon is distant from the heteroatom (type C). Computational studies on these remote N-heterocyclic carbenes (rNHC) derived from pyridine or quinoline suggested an even stronger σ-donor ability compared to their well-known normal NHC counterparts, which may be beneficial for certain types of catalytic reactions. Despite these promising properties, rNHCs have not attracted the same degree of attention as common NHCs yet. As a contribution to this little-explored field, we have introduced the first pyrazole-derived remote NHC (type D) complexes of Pd(II).
4. Different types of C-H⋅⋅⋅M interactions:
In general, there are three forms of C-H⋅⋅⋅M interactions including i) agostic, ii) anagostic (or preagostic) and iii) hydrogen bond. The first (i) is usually referred to a 3c-2e interaction resulting in a pronounced highfield chemical shift of the corresponding H atom. Agostic interactions are often found in d6 complexes with small M-H distances of ~1.8-2.2 Å and C-H⋅⋅⋅M angles of ~90-130°. Hydrogen bonds (iii) on the other hand are 3c-4e interactions with an almost linear geometry accompanied by a downfield 1H NMR shift of the participating H atom. Rare anagostic (or preagostic) interactions (ii) occurring in square planar d8 systems are of broad general interest due to possible implications for the mechanism of C-H activation. They are in between the former two with characteristic downfield shifts of the anagostic protons, H-M distances of ~2.3-2.9 Å and C-H⋅⋅⋅M angles usually in the range of 130-170°. The origin of these interactions is still under debate and may involve donation of filled dz2 or dxz/yz-orbitals of the metal center into the C-H sigma* orbital. Nevertheless, anagostic interactions are generally weak and it is pointless to discuss such interactions, when downfield shift of <1 ppm are observed.
In our research, we have generally observed anagostic interactions in d8 complexes of the 1,3-diisopropylbenzimidazolin-2-ylidene ligand. A comparison among four of such isoelectronic and isostructural complexes shows a decrease of anagostic interactions in the order Ni(II) > Pt(II) > Pd(II) >> Au(III). Interestingly, this observation may be related to the Shannon effective ionic radii of these d8 ions in square planar geometry, which increase in the order 0.49 Å {Ni(II)} < 0.60 Å {Pt(II)} < 0.64 Å {Pd(II)} < 0.68 Å {Au(III)}. The strong donation of the carbene ligand to the smallest Ni(II) ion would result in the least Lewis acidic metal center, which translates into the observed strongest anagostic interaction. The most Lewis acidic Au(III) complex, on the other hand, does not show any C-H⋅⋅⋅M interactions in solution at all as corroborated by a slight upfield-shift of the corresponding C-H proton.
5. Catalytic Organic Transformations:
N-heterocyclic carbenes have made major impact as ancillary ligands in transition-mediated catalysis. Due to their nucleophilic character NHCs are also often directly employed as organocatalysts. We have employed some of our complexes as catalyst precursors for C-C coupling reactions such as Mizoroki-Heck, aqueous Suzuki-Miyaura or room temperature Kumada-Corriu.
In particular we have studied the influence of trans- versus cis-Pd(II) NHC-complexes on the catalyst initiation in the Mizoroki-Heck reaction. A detailed kinetic study revealed that the cis-bis(carbene) complex gave a fast conversion, whereas its trans-isomer is much slower showing an induction period.
This surprising result is worthy of comment and suggests that reduction of Pd(II) to Pd(0) and thus formation of the catalytically active species occurs faster in the cis-isomer. This reductive process can already be accomplished by traces of formic acid in DMF. The figure below depicts a feasible pathway which includes ligand substitution (a), CO2 elimination (b) and reductive elimination (c) affording the active Pd(0) species, which subsequently enters the catalytic cycle. A facile dissociation of the iodo ligand as a possible introductory step is favored in the cis-isomer, in which the strong trans-effect of the carbene ligand weakens the Pd-I bond. On the other hand, such a Pd-I bond cleavage is less favored in the trans-isomer and may account for the observed induction period, in which a trans-cis isomerization might occur to facilitate Pd-I bond cleavage prior to the reduction step.
In a separate study we investigated the influence of co-ligands in the Mizoroki-Heck reaction. Cis-bis(carbene) Pd(II) complexes 1-3 bearing different carboxylato ligands have been chosen for that purpose. As depicted in the in the following table, all three complexes show a very similar activity for a range of substrates. It seems that the labile carboxylato ligand has no or only very little influence on the catalytic activity. This is expected when the anionic ligand does not participate in the catalytic cycle, which is in line with the standard model of the Heck reaction with the 14e- species[Pd0L2] (L = PR3 or here NHC) as the key intermediate.
Interestingly, when 4-chlorobenzaldehyde was employed in the reaction, the three complexes show different catalytic activities. Complex 1 was most effective and gave quantitative yield, whereas 2 and 3 activated the aryl chloride slightly less efficiently with good yields of 77% and 85%, respectively (entries 7-9). These surprising results and the superiority of the acetato-complex 1 may be explained using the Amatore-Jutand model, which proposes the 16e- species [Pd0L2(OAc)]- as a key intermediate. This model emphasizes the important ability of the acetato ligand in stabilizing active Pd0-species and explains indirectly the advantage of NaOAc as a base in the Mizoroki-Heck reaction. This ability is probably less pronounced for the fluorocarboxylato ligands in 2 and 3 due to their weaker donor-strength.
In particular we have studied the influence of trans- versus cis-Pd(II) NHC-complexes on the catalyst initiation in the Mizoroki-Heck reaction. A detailed kinetic study revealed that the cis-bis(carbene) complex gave a fast conversion, whereas its trans-isomer is much slower showing an induction period.
This surprising result is worthy of comment and suggests that reduction of Pd(II) to Pd(0) and thus formation of the catalytically active species occurs faster in the cis-isomer. This reductive process can already be accomplished by traces of formic acid in DMF. The figure below depicts a feasible pathway which includes ligand substitution (a), CO2 elimination (b) and reductive elimination (c) affording the active Pd(0) species, which subsequently enters the catalytic cycle. A facile dissociation of the iodo ligand as a possible introductory step is favored in the cis-isomer, in which the strong trans-effect of the carbene ligand weakens the Pd-I bond. On the other hand, such a Pd-I bond cleavage is less favored in the trans-isomer and may account for the observed induction period, in which a trans-cis isomerization might occur to facilitate Pd-I bond cleavage prior to the reduction step.
In a separate study we investigated the influence of co-ligands in the Mizoroki-Heck reaction. Cis-bis(carbene) Pd(II) complexes 1-3 bearing different carboxylato ligands have been chosen for that purpose. As depicted in the in the following table, all three complexes show a very similar activity for a range of substrates. It seems that the labile carboxylato ligand has no or only very little influence on the catalytic activity. This is expected when the anionic ligand does not participate in the catalytic cycle, which is in line with the standard model of the Heck reaction with the 14e- species[Pd0L2] (L = PR3 or here NHC) as the key intermediate.
Interestingly, when 4-chlorobenzaldehyde was employed in the reaction, the three complexes show different catalytic activities. Complex 1 was most effective and gave quantitative yield, whereas 2 and 3 activated the aryl chloride slightly less efficiently with good yields of 77% and 85%, respectively (entries 7-9). These surprising results and the superiority of the acetato-complex 1 may be explained using the Amatore-Jutand model, which proposes the 16e- species [Pd0L2(OAc)]- as a key intermediate. This model emphasizes the important ability of the acetato ligand in stabilizing active Pd0-species and explains indirectly the advantage of NaOAc as a base in the Mizoroki-Heck reaction. This ability is probably less pronounced for the fluorocarboxylato ligands in 2 and 3 due to their weaker donor-strength.
6. Determination of ligand donor strentghs:
The properties and reactivities of metal complexes are mainly determined by the electronic and steric effects of their ligands. The most common method to determine the donor strengths of ligands is Tolman’s electronic parameter (TEP), which compares the CO vibrational modes of [Ni(CO)3PR3] complexes. The second major electronic parameter developed by Lever, is based on electrochemical E0 values of a redox couple (e.g. RuII/III) bearing the ligands of interest (LEP). Although both methods provided valuable insights for the characterization of ligands, they are encumbered by complicated synthetic or relatively uncommon analytical routes. TEP determination requires utmost caution in the handling of the extremely toxic gas [Ni(CO)4], whereas LEP is hampered by the use of less common electrochemical apparatus and irreversible redoxchemistry. Furthermore, only few ligands appear in both TEP and LEP series, making a direct comparison difficult. In analogy to phosphines, the electronic properties of N-heterocyclic carbenes (NHCs) are generally determined by measuring the CO stretch in low valent complexes such as [Ni(CO)3NHC] or [MX(CO)2(NHC)] (M = IrI or RhI, X = halide) as a modification of TEP. Here, it is assumed that the amount of π-back-donation to CO is directly related to the donor strength of the NHC, although non-negligible π-back-donation from the electron rich metal center to the NHC competing with that to the CO gave inconsistent results. To overcome these limitations, we have introduced a convenient, safe and non-destructive new electronic parameter for the unified evaluation and comparison of Werner-type and organometallic ligands such as phosphines and NHCs, which is based on a 13C NMR spectroscopic evaluation of a carbonyl-free mixed NHC/co-ligand PdII system.
The method is based on the sensitivity of the carbene signal to the Lewis acidity of the metal center, which in turn is influenced by the co-ligands. We have therefore anticipated that mixed-ligand complexes of the type trans-[PdBr2(iPr2-bimy)L] bearing various co-ligands L should enable the direct comparison of Werner-type and organometallic ligands on the same scale by determining the iPr2-bimy carbene shift. The trans-standing iPr2-bimy probe is hereby chosen, because it is least influenced by any steric effects and hence enables the most rigorous assessment of the electronic properties of the ligand in question. The synthesis of the respective iPr2-bimy/L mixed-ligand complexes is straightforward and involves a common bridge-cleavage reaction of the easily available dimeric complex [PdBr2(iPr2-bimy)]2 (1) with two equivalents of the ligand in question L. The evaluation of overall 25 ligands showed that a stronger donor would lead to an enhanced downfield shift of the iPr2-bimy carbene signal, whereas a relative highfield shift is observed with poorer donors. Using our methodology differences due to backbone and substituent effects can be evaluated on a finer level.
The method is based on the sensitivity of the carbene signal to the Lewis acidity of the metal center, which in turn is influenced by the co-ligands. We have therefore anticipated that mixed-ligand complexes of the type trans-[PdBr2(iPr2-bimy)L] bearing various co-ligands L should enable the direct comparison of Werner-type and organometallic ligands on the same scale by determining the iPr2-bimy carbene shift. The trans-standing iPr2-bimy probe is hereby chosen, because it is least influenced by any steric effects and hence enables the most rigorous assessment of the electronic properties of the ligand in question. The synthesis of the respective iPr2-bimy/L mixed-ligand complexes is straightforward and involves a common bridge-cleavage reaction of the easily available dimeric complex [PdBr2(iPr2-bimy)]2 (1) with two equivalents of the ligand in question L. The evaluation of overall 25 ligands showed that a stronger donor would lead to an enhanced downfield shift of the iPr2-bimy carbene signal, whereas a relative highfield shift is observed with poorer donors. Using our methodology differences due to backbone and substituent effects can be evaluated on a finer level.
