Molecular Adsorbates Switch on Heterogeneous Catalysis: Induction

Jun 21, 2017 - Dedicated to Professor Teruaki Mukaiyama in celebration of his 90th birthday (Sotsuju). Financial support by the FCI (J.B.E.), DFG (Lei...
0 downloads 0 Views 1MB Size
Communication pubs.acs.org/JACS

Molecular Adsorbates Switch on Heterogeneous Catalysis: Induction of Reactivity by N‑Heterocyclic Carbenes Johannes B. Ernst,† Christian Schwermann,‡ Gen-ichi Yokota,§ Mizuki Tada,§,∥,⊥ Satoshi Muratsugu,*,§ Nikos L. Doltsinis,*,‡ and Frank Glorius*,† †

Organisch-Chemisches Institut, Westfälische Wilhelms-Universität Münster, Corrensstrasse 40, 48149 Münster, Germany Institute for Solid State Theory and Center for Multiscale Theory and Computation, Westfälische Wilhelms-Universität Münster, Wilhelm-Klemm-Str. 10, 48149 Münster, Germany § Department of Chemistry, Graduate School of Science, ∥Research Center for Materials Science (RCMS), and ⊥Integrated Research Consortium on Chemical Sciences (IRCCS), Nagoya University, Furo-cho, Chikusa, Nagoya, Aichi 464-8602, Japan ‡

S Supporting Information *

represent a powerful class of ligands in homogeneous catalysis, as organic ligands. A prerequisite for the development of more efficient heterogeneous catalysis is the understanding of the binding, mobility,5d and electronic effect of NHCs on heterogeneous catalysts. As a prototypical system, we decided to investigate the influence of NHCs as ligands on the activity of an aluminasupported palladium catalyst (Pd/Al2O3). To be able to investigate the activating effect of the NHC, we chose the hydrogenolysis of bromobenzene as our benchmark reaction (Scheme 1), since the activation of bromobenzene should be, according to the principles of homogeneous catalysis, facilitated in the presence of electron-donating ligands. NHC coordination should, in this case, lead to a more electron-rich heterogeneous catalyst, which should result in a lower barrier

ABSTRACT: We report the N-heterocyclic carbene (NHC)-induced activation of an otherwise unreactive Pd/Al2O3 catalyst. Surface analysis techniques demonstrate the NHC being coordinated to the palladium particles and affecting their electronic properties. Ab initio calculations provide further insight into the electronic effect of the coordination with the NHC injecting electron density into the metal nanocluster thus lowering the barrier for bromobenzene activation. By this NHC modification, the catalyst could be successfully applied in the Buchwald− Hartwig amination of aryl chlorides, bromides, and iodides. Various heterogeneity tests could additionally show that the reaction proceeds via a heterogeneous active species.

N-Heterocyclic carbenes (NHCs) are a unique class of ligands which is widely applied in the fields of coordination chemistry and homogeneous catalysis. They are characterized by their electron-rich nature and their superb ability to form strong bonds to metals. In general, they act as strong σ-donor and weak π-acceptor ligands and are capable of significantly influencing the reactivity and selectivity of homogeneous transition metal catalysts.1−3 Recently, the use of NHCs for the stabilization of transition metal based nanoparticles (NPs) and surfaces has emerged.4,5 The groups of Chechik and Fairlamb6 and Tilley7 stabilized Au-NPs by coordinating NHCs and the groups of Glorius,8−10 Ravoo,9 Chaudret,10,11 and others12 applied NHCs to catalytically active NP-systems such as Pd, Pt, Ru, and Cu. The applicability of NHC stabilized AuNPs in biological systems could be recently demonstrated by Johnson et al.13 Furthermore, Richeter et al. investigated the different binding modes of NHCs to Au-NPs by DFTcalculations14 and the groups of Yang and Chang applied NHCs as ligands for Au-NPs in electrochemistry.15 The application of organic ligands to electronically modify heterogeneous catalysts and its application in catalysis is an emerging field and has been studied for amines,16 thiols and phosphines.17 However, no change in the inherent reactivity or the induction of a new kind of reactivity could be realized.18 To accomplish selective activation of a previously inactive heterogeneous catalyst, we envisioned applying NHCs, which © 2017 American Chemical Society

Scheme 1. Effect of the NHC-Modification on the Hydrogenolysis of Bromobenzenea

a

Yields determined by GC-FID analysis.

Received: May 17, 2017 Published: June 21, 2017 9144

DOI: 10.1021/jacs.7b05112 J. Am. Chem. Soc. 2017, 139, 9144−9147

Communication

Journal of the American Chemical Society

attachment of IPr on Pd.9c FT-IR peaks in the C−H stretching region (2969, 2935, and 2876 cm−1, Figure S8) and the weight loss attributed to IPr determined by TGA (3.9 wt %, Figure S9) further support IPr being attached to the surface. Additionally, TEM and STEM-EDS analyses of IPr@Pd/Al2O3 and Pd/ Al2O3 revealed that the NHC modification did not lead to a significant change in the morphology of the Pd-NP or the Al2O3 support (Figures S10, S11). To investigate the effect of the NHC modification on the electronic situation of Pd, we conducted Pd 3d XPS analysis of IPr@Pd/Al2O3, ICy@Pd/ Al2O3, and unmodified Pd/Al2O3 and compared the corresponding Pd 3d binding energies (Figures 1, S6, Table S19). The Pd 3d peaktop energies of IPr@Pd/Al2O3 (339.9 eV (Pd 3d3/2) and 334.7 eV (Pd 3d5/2)) are shifted to lower binding energies compared to those of Pd/Al2O3 (340.3 eV (Pd 3d3/2) and 335.0 eV (Pd 3d5/2)), suggesting the donation of electron density from IPr to Pd. The corresponding peaktop energies for ICy@Pd/Al2O3 (340.0 eV (Pd 3d3/2) and 334.8 eV (Pd 3d5/2)) are located between the binding energies of IPr@Pd/Al2O3 and Pd/Al2O3 suggesting the donation of electron density from ICy to Pd, however, in contrast to coordination chemistry, less than IPr.20 These results experimentally support our hypothesis of an electronically modified heterogeneous catalyst by NHC coordination. To investigate the influence of the NHC, especially with aromatic N-substituents, on the electronic character of the NP, we conducted DFT calculations of Pd13-nanoclusters in the presence of different NHCs as our model system (Figure 2 I,

for the cleavage of the C−Br bond of bromobenzene compared to the unmodified catalyst allowing for improved catalyst performance. Interestingly, we were able to observe a significant accelerating effect of NHCs with aromatic N-substituents being the most effective ligands (IPr, IMes). NHCs bearing alkyl Nsubstituents (ICy, IMe) were less efficient in the promotion of the hydrogenolysis, however, still superior to unmodified Pd/ Al2O3 (Scheme 1). Inspired by the observed activating effect of the NHC coordination, we went on to study the structure of the IPrmodified catalyst in detail to gain a deeper understanding of the interaction and effect of the NHC coordination. To determine whether the NHC is binding via the carbene carbon atom, we performed 13C-SS-NMR experiments using 13C-labeled IPrmodified Pd/Al2O3 (13C-labeled IPr@Pd/Al2O3) and compared the chemical shift of the carbene carbon atom with 13Clabeled IPr-modified Al2O3 (13C-labeled IPr@Al2O3) and unlabeled IPr-modified Pd/Al2O3 (unlabeled IPr@Pd/Al2O3). We observed a distinct peak at 163 ppm (Figure 1), which was

13

Figure 2. Localized molecular orbitals of IMes-stabilized Pd13. (I) σBond between carbene carbon and Pd. (II) π-Bond between carbene carbon and Pd. (III) Bonding combination of aromatic π- and Pd 4dorbital.

13

Figure 1. (A) C SS CP-MAS NMR-spectra of (a) C-labeled IPr@ Pd/Al2O3, (b) unlabeled IPr@Pd/Al2O3, and (c) 13C-labeled IPr@ Al2O3. (B) Normalized Pd 3d XPS spectra of (a) IPr@Pd/Al2O3 (red dashed line), (b) ICy@Pd/Al2O3 (green dashed line), (c) Pd/Al2O3 (blue dashed line), and (d) IPr@Al2O3.

II). Geometry optimizations performed at the PBE/def2SVP level21 show that all NHCs bind via the carbene carbon atom to the nanocluster. In addition, the aromatic N-substituents of IMes and IPr bind to the nanocluster via their delocalized πorbitals (Figure 2 III). Due to the coordination of the NHCs to the nanocluster, the ionization potentials (IPs), obtained by ΔSCF calculations, are decreased significantly, in particular for NHCs bearing aromatic N-substituents (Tables 1, S22, S23). The trend toward a reduced IP depending on the NHC coordination represents the

shifted upfield from that of the corresponding free IPr (220 ppm, Figure S4) and in the range of NHCs bound to transition metals.1,4 Considering that in this region no peak was observed for 13C-labeled IPr@Al2O3, and the fact that the shift is very similar to NHC modified Ru/K−Al2O3 catalysts,8b we attribute this peak to IPr binding via the carbene carbon atom to the Pdcatalyst. The peak at 156 ppm can also be attributed to IPr being attached to Pd. Two peaks at 146 and 133 ppm were observed, respectively. The peak at 146 ppm can be attributed to the aromatic region of IPr as the comparison with the peaks of 13C-labeled IPr@Al2O3 at 145 ppm and unlabeled IPr@Pd/ Al2O3 at 146 ppm matches the observed peak. The peak at 133 ppm is mainly attributed to the reprotonated carbene carbon atom of IPr and subsequent adsorption of the Lewis basic aromatic N-substituents onto the Lewis acidic support.19 Peaks at 131 and 126 ppm can be assigned to the aromatic ring of IPr. Furthermore, a clear N 1s XPS signal at 400 eV was observed for IPr@Pd/Al2O3 (Figure S5) which was shifted to lower binding energies compared to IPr@Al2O3, supporting the

Table 1. Calculated IP and Activation Barriers (ΔE#) for the Hydrogenolysis of Bromobenzene on Pd13-Clusters with Different NHC Adsorbates

Pd13 IMe ICy IMes IPr 9145

IP (eV)

ΔE# (kJ/mol)

6.39 5.78 5.64 5.39 5.16

16.50 12.32 12.27 8.12 6.66 DOI: 10.1021/jacs.7b05112 J. Am. Chem. Soc. 2017, 139, 9144−9147

Communication

Journal of the American Chemical Society Scheme 2. Substrate Scope of the Reactiona

XPS-results of a decreased binding energy upon NHC coordination obtained for Pd 3d. Subsequently, relaxed potential hypersurface scans were performed for the C−Br bond of adsorbed bromobenzene to determine the effect of the NHC coordination on the barrier for the activation of bromobenzene (Figures S12, S13). The resulting dissociation barriers correlate with the IP and confirm the important role of the aromaticity of the N-substituents observed in the hydrogenolysis of bromobenzene. In accordance with the effect in the hydrogenolysis, coordination of IPr and IMes leads to a smaller activation barrier ΔE# for the cleavage of the C−Br bond of bromobenzene compared to NHCs bearing alkyl Nsubstituents which are, nevertheless, superior to the unmodified Pd-cluster (Tables 1, S24). Based on the structural analysis and the theoretical results showing that the electronic effect of NHCs leads to a lower barrier for the activation of aryl halides, we decided to apply our concept to a more demanding reaction. As there are various examples in the literature employing ligand-free supported Pdcatalysts in Suzuki−Miyaura coupling reactions,22,23 we decided to focus on the Pd-catalyzed Buchwald−Hartwig amination as our reaction of choice, as this reaction is widely applied in academia and industry, however, generally using highly sophisticated homogeneous catalyst systems.2,24 It is important to note that in the absence of NHCs no product formation could be observed. The application of IPr as ligand led to the formation of the desired product in moderate yield. Additionally, we determined a strong dependence of the applied NHC, as only IPr and its related analogues IPr* and SIPr were able to induce product formation though other NHCs, which are effective in homogeneous Buchwald−Hartwig aminations, did not lead to product formation (ItBu, IAd, IMes).2 After optimizing the reaction conditions, we were able to determine the substrate scope of the reaction (Scheme 2). Substitution in the ortho-, meta-, and para-position is tolerated on the aryl halide (3a−f) as well as on the corresponding aniline (3h−j) moiety. Furthermore, heteroaryl bromides (3g), secondary aniline derivatives (3k), and alkyl amines (3l) led to product formation in moderate to good yields. Aryl chlorides and aryl iodides are suitable coupling partners as well. Fortunately, the catalyst could be prepared in advance, as pretreatment of the catalyst with a solution of free IPr and subsequent application under the reaction conditions provided the desired product 3a in a similar yield suggesting that there is no difference between in situ generation and separate preparation of the NHC modified catalyst. To gain a deeper understanding of whether or not the active species is heterogeneous in nature, we performed a variety of experiments.25 Addition of Hg, which should result in the formation of a catalytically inactive amalgam, led to an immediate inhibition of the reaction. Hot filtration of the reaction suspension to remove the insoluble heterogeneous catalyst also led to the inhibition of the reaction. Finally, the three-phase test, for which one of the substrates is immobilized on a solid support, supports the presence of a heterogeneous catalytically active species, since no product formation could be observed under the reaction conditions.26 By contrast, the corresponding homogeneous catalytic system afforded the expected product under the conditions of the heterogeneity tests. Furthermore, we observed a significant dependence on the loading of the NHC. Higher loadings resulted in a slower and less productive reaction suggesting a heterogeneous active species, as larger amounts of active sites are blocked due to a

a

Reactions were performed on a 0.4 mmol scale with 1.2 equiv of amine. Isolated yields of product are reported. (α) Preformed catalyst was used. (β) Chlorobenzene or (γ) iodobenzene were used as aryl halide.

higher NHC loading (Figure S1). TXRF-results also indicated a heterogeneous active species, as the Pd-concentration in solution was below