Article pubs.acs.org/Organometallics
Magnetic Circular Dichroism and Density Functional Theory Studies of Iron(II)-Pincer Complexes: Insight into Electronic Structure and Bonding Effects of Pincer N‑Heterocyclic Carbene Moieties Tessa M. Baker,† Teresa L. Mako,‡ Aristidis Vasilopoulos,‡ Bo Li,‡ Jeffery A. Byers,*,‡ and Michael L. Neidig*,† †
Department of Chemistry, University of Rochester, Rochester, New York 14627, United States Department of Chemistry, Boston College, Chestnut Hill, Massachusetts 02467, United States
‡
S Supporting Information *
ABSTRACT: Iron complexes containing pincer ligands that incorporate N-heterocyclic carbene (NHC) moieties are of significant interest in organometallic catalysis in order to generate more oxidatively robust complexes that may exhibit novel catalytic properties. In order to define the effect that introducing NHC moieties into pincer ligands has on electronic structure and bonding in iron(II)-pincer complexes, MCD and DFT studies of (iPrCDA)FeBr2, (iPrPDI)FeBr2, and (iPrCNC)FeBr2 were performed. These studies quantify the electronic structures and bonding interactions as a function of pincer ligand variation. They also demonstrate that the observed ligand fields (and, hence, spin states) directly correlate to the increased Fe−C bonding and pincer-donating abilities that result from introducing NHC moieties into the pincer ligand. However, the net donor abilities of the pincers and the strength of the Fe-pincer interaction do not directly correlate to the number of NHC moieties present, but instead are determined to be due to differences in Fe−C and overall Fe-pincer bonding as a result of the position of the NHC moieties in the pincer ligand and the overall geometric constraints of the pincer architecture.
■
ligands (abbreviated here as CNC ligands).37 In this work, Mössbauer, X-ray absorption, and density functional theory (DFT) studies were combined with structural characterization to demonstrate that (iPrCNC)Fe(N2)2 and the related dimethylaminopyridine (DMAP) adduct (iPrCNC)Fe(DMAP)(N2) are best described as complexes with redox-noninnocent ligands with the iPrCNC pincer acting as a classical π-acceptor ligand. In addition, a recent study by Kühn and co-workers investigated how NHC and pyridine change the iron(II) center in low-spin iron complexes with tetradentate pincer ligands.38 Cyclic voltammetry carried out on these complexes revealed that the oxidation potentials of the complexes were linearly correlated with the number of NHC moieties coordinated to iron. A particularly interesting NHC-containing pincer ligand is bis(amidinato)-N-heterocyclic carbene ligands (referred to here as carbenodiamidine, or CDA, ligands), which, when bound to iron, have been shown to exhibit enhanced reactivity in certain reactions compared to the analogous PDI complexes of iron.33,39 Magnetic measurements carried out on (iPrCDA)FeCl2 also revealed that the complex exhibits an S = 1 ground state at low temperature, in contrast to the S = 2 ground state
INTRODUCTION N-Heterocyclic carbene (NHC)-containing pincers ligands have been of recent interest in organometallic catalysis in order to generate more oxidatively robust pincer complexes and impart new catalytic reactivity. For example, second- and thirdrow transition metal complexes of NHC-containing pincer ligands have been shown to be highly effective in a wide variety of catalytic reactions1−13 including C−H functionalization,14−16 hydroamination,17 and hydrogenation.18−20 While less common, first-row transition metal complexes with NHCcontaining pincer ligands have also proven to be capable of both stoichiometric and catalytic reactivity.21−30 In particular, NHC-containing pincer complexes of iron have been shown to successfully catalyze reactions including hydrogenation31 and polymerization.32,33 Due to the important catalytic properties of iron-pincer complexes, electronic structure studies of these systems have been widely explored,34−37 particularly where ligand redox noninnocence can occur such as in iron complexes containing bis(imino)pyridine ligands (also known as pyridyl diimine, or PDI, ligands). By contrast, detailed electronic structure studies of NHC-containing pincer complexes of iron have been less common. One recent study was reported by Chirik and coworkers, who investigated the electronic structure of iron complexes containing bis(N-heterocyclic carbene)pyridine © XXXX American Chemical Society
Received: August 12, 2016
A
DOI: 10.1021/acs.organomet.6b00651 Organometallics XXXX, XXX, XXX−XXX
Article
Organometallics for (iPrPDI)FeX2 (X = Cl, Br).40 Moreover, (iPrCNC)FeBr2 is also S = 2 despite the fact that it contains two NHC moieties in the pincer ligand compared to one NHC moiety in iPrCDA. Clearly, a more detailed understanding of how the introduction of NHC moieties into pincer ligand architectures affect electronic structure, bonding, and, hence, reactivity is needed in order to explain this trend and to facilitate the rational design of new iron-pincer complexes with targeted electronic properties. In the present study, magnetic circular dichroism (MCD) spectroscopy is utilized in combination with DFT studies to further define the ligand field and spin-state differences in (iPrCDA)FeBr2 relative to (iPrPDI)FeBr2 and (iPrCNC)FeBr2 (Figure 1). Combined, these studies provide quantitative
ments reported for (iPrCDA)FeCl2, (iPrCDA)FeBr2 has an S = 1 ground state at 5 K. The 5 K, 7 T UV−vis MCD spectrum of (iPrCDA)FeBr2 (Figure 2D) contains a large number of transitions within the accessible experimental range from ∼16 000 to ∼28 000 cm−1 of charge-transfer character as well as a high-energy LF transition as expected for such an S = 1 complex and predicted by TD-DFT (see SI). However, due to the significant number of transitions in the UV−visible region, this LF band cannot be unambiguously identified. By contrast to (iPrCDA)FeBr2, the 5 K, 7 T NIR MCD spectrum of (iPrPDI)FeBr2 (Figure 2B) exhibits only two LF transitions, at 0.5 for the LF band at 10 400 cm−1) assigned as an MLCT transition from Fe d to PDI π* based on TD-DFT calculations (see SI). The 5 K, 7 T UV−vis MCD spectrum (Figure 2E) contains multiple chargetransfer transitions from ∼14 300 to ∼29 000 cm−1 comprising both MLCT and LMCT transitions based on TD-DFT (see SI). MCD studies of (iPrCNC)FeBr2 were performed to obtain insight into the ligand field and electronic structure of this S = 2 complex compared to that determined for S = 2 (iPrPDI)FeBr2 and S = 1 (iPrCDA)FeBr2. The 5 K, 7 T NIR MCD spectrum of (iPrCNC)FeBr2 (Figure 2C) contains two LF transitions, a lowenergy transition at 7710 cm−1 and an additional broad transition at ∼13 400 cm−1. Notably, the observed LF transitions for (iPrCNC)FeBr2 correspond to a larger overall ligand field splitting (10Dq ≈ 10 555 cm−1) than that observed for (iPrPDI)FeBr2 (10Dq ≤ 7700 cm−1), indicating that iPrCNC is a stronger field ligand than iPrPDI in the five-coordinate (pincer)FeBr2 complexes. The saturation magnetization data for (iPrCNC)FeBr2 collected at 7634 cm−1 (Figure 2I) are wellfit to an S = 2 −ZFS non-Kramers doublet model with groundstate spin-Hamiltonian parameters of δ = 1.6 ± 0.2 cm−1 and g|| = 9.0 ± 0.2 with D = −11 ± 2 cm−1 and |E/D| = 0.23 ± 0.02. The 5 K, 7 T UV−vis MCD spectrum of (iPrCNC)FeBr2 (Figure 2F) contains multiple transitions from ∼17 000 to ∼27 000 cm−1 best described as predominately Fe d to CNC π* MLCT transitions from TD-DFT (see SI) with additional LMCT character present for transitions beyond 27 000 cm−1. Electronic Structure Calculations of (Pincer)FeBr2 Complexes. In order to obtain more detailed insight into the electronic structure and bonding differences in (iPrCDA)FeBr2, (iPrPDI)FeBr2, and (iPrCNC)FeBr2, spin-unrestricted DFT calculations were performed. Geometry optimizations using B3LYP/def2-TZVP and a THF solvent model were performed starting from the reported crystal structure
Figure 1. Structures of (iPrCDA)FeBr2, (iPrPDI)FeBr2, and (iPrCNC)FeBr2.
insight into the effect of NHC moieties in pincer ligands on iron-pincer bonding and electronic structure, which highlights the importance of the position of the NHC moiety as well as the geometric constraints of the pincer architectures in modulating iron-pincer bonding.
■
RESULTS AND ANALYSIS MCD Characterization of (Pincer)FeBr2 Complexes. MCD spectroscopic studies of (iPrCDA)FeBr2, (iPrPDI)FeBr2, and (iPrCNC)FeBr2 were performed to probe the effects of variations in NHC moieties of the pincer ligand (a central NHC, no NHC, or two peripheral NHCs) on the d-orbital splittings and overall electronic structure of the resulting (pincer)FeBr2 complexes. The 5 K, 7 T, near-infrared (NIR) MCD spectrum of (iPrCDA)FeBr2 (Figure 2A) contains five ligand-field (LF) transitions between ∼7000 and ∼13 500 cm−1. This large number of LF transitions in the near-infrared region exceeds the number expected for a high-spin iron(II) (S = 2 complex) from LF theory and, instead, is consistent with the expanded number of transitions expected for an intermediate-spin (S = 1) iron(II) complex. Consistent with the observed LF transitions, the saturation magnetization behavior of (iPrCDA)FeBr2 collected at 9217 cm−1 is well-fit to an S = 1 positive zero-field split (+ZFS) non-Kramers doublet model with ground-state spin Hamiltonian parameters of D = 20 ± 4 cm−1 and |E/D| = 0.24 ± 0.03 with giso = 2.05 ± 0.10 (Figure 2G). Thus, consistent with previous magnetic measureB
DOI: 10.1021/acs.organomet.6b00651 Organometallics XXXX, XXX, XXX−XXX
Article
Organometallics
Figure 2. MCD spectra of (pincer)FeBr2 complexes. 5 K, 7 T near-infrared MCD spectra of (A) (iPrCDA)FeBr2, (B) (iPrPDI)FeBr2, and (C) (iPrCNC)FeBr2. 5 K, 7 T UV−vis MCD spectra of (D) (iPrCDA)FeBr2, (E) (iPrPDI)FeBr2, and (F) (iPrCNC)FeBr2. Gaussian fits are shown as dashed lines. VTVH-MCD data (dots) and fit (lines) of (G) (iPrCDA)FeBr2 collected at 9217 cm−1, (H) (iPrPDI)FeBr2 collected at 13 889 cm−1, and (I) (iPrCNC)FeBr2 collected at 7634 cm−1. All spectra were collected in 1:1 THF/2-MeTHF.
Table 1. Comparison of Calculated and Experimental Structural Parameters for (Pincer)FeBr2 Complexes
Fe(CDA)Br2 Fe−1 (Å) Fe−2 (Å) Fe-3 (Å) Fe−Br′ (Å) Fe−Br″ (Å) 1−Fe−3 (deg) Br′−Fe−Br″ (deg)
Fe(PDI)BrF2
Fe(CNC)Br2
exptl
calcd S = 1
calcd S = 2
exptl
calcd
exptl
calcd
2.015(2) 1.811(3) 2.004(2) 2.4775(5) 2.4215(5) 145.38(9) 106.44(2)
2.05 1.83 2.04 2.57 2.46 147.0 107.1
2.24 2.04 2.24 2.55 2.45 140.7 116.0
2.205(2) 2.073(2) 2.209(2) 2.4692(6) 2.3999(6) 141.13(8) 116.054(18)
2.21 2.10 2.22 2.54 2.45 139.6 114.6
2.193(10) 2.211(8) 2.166(10) 2.5065(18) 2.4208(17) 141.0 (4) 108.19(7)
2.19 2.28 2.17 2.59 2.49 138.4 111.4
(iPrCDA)FeBr2 and (iPrCNC)FeBr2 (see SI for additional FMO depictions including those for (iPrPDI)FeBr2). For (iPrCDA)FeBr2, the α FMOs exhibit dominant Fe d character in the following orbitals: occupied α169, α170, α171, and α182 as well as the unoccupied α184 orbital, assigned as dyz, dxz, dxy, dz2, and dx2−y2, respectively. In the β manifold, the FMOs exhibit dominant Fe d character in the occupied β179 and β180 orbitals assigned as dxy and dxz, respectively, as well as the unoccupied β182, β183, and β184 orbitals assigned as dyz, dz2, and dx2−y2, respectively. For (iPrCDA)FeBr2, dx2−y2 is highest in energy and unoccupied in both the α and β manifolds due to the direct σ* interaction between the N, C, and N ligating atoms of the iPrCDA pincer ligand and dx2−y2 iron-based orbitals (Figure 3). The increase in energy of dx2−y2 due to this σ* interaction with iPrCDA contributes to the increased splitting in energy of the Eg-derived orbitals that leads to intermediate S =
coordinates, yielding good agreement between experiment and theory (Table 1) for both S = 2 (iPrPDI)FeBr2 and (iPrCNC)FeBr2. For (iPrCDA)FeBr2, both S = 1 and S = 2 models were calculated, with the S = 1 model giving good agreement with the experimental structural parameters (Table 1), consistent with the previous identification that this complex is S = 1 from MCD. For all three complexes, evaluation of molecular orbital (MO) character and energies was subsequently conducted from the optimized geometries using spinunrestricted B3LYP/def2-TZVP. The calculated MO energy diagrams for (iPrCDA)FeBr2 (S = 1), (iPrPDI)FeBr2, and (iPrCNC)FeBr2 are given in Figure 3. For consistency with S = 1 (iPrCDA)FeBr2, both the α and β orbitals are given for all three compounds. Select frontier molecular orbital (FMO) depictions are also given in Figure 3 for the α FMOs of dominant d-orbital character for C
DOI: 10.1021/acs.organomet.6b00651 Organometallics XXXX, XXX, XXX−XXX
Article
Organometallics
Figure 3. Calculated FMO diagrams for (pincer)FeBr2 complexes. Selected α-orbital depictions are given for S = 1 (iPrCDA)FeBr2 and S = 2 (iPrCNC)FeBr2. Additional orbital depictions for each complex are given in the SI.
1. It is notable that in (iPrCDA)FeBr2 a significant Fe−C πinteraction involving β180 (dxz) is also present (see SI). These findings are consistent with the bonding picture previously reported for (iPrCDA)FeCl2.40 In contrast to (iPrCDA)FeBr2, (iPrPDI)FeBr2 and (iPrCNC)FeBr2 exhibit similar overall MO diagrams to each other in both the α and β manifolds, where dx2−y2 is the highest energy unoccupied d orbital for both complexes. Furthermore, analogous to (iPrCDA)FeBr2, the dx2−y2 orbital in both (iPrPDI)FeBr2 and (iPrCNC)FeBr2 is highest in energy due to the direct σ* interaction between the iron dx2−y2 orbital and iPr PDI and iPrCNC pincers, respectively (Figure 3). However, an additional σ interaction involving α161 (dxy) and both NHC donors is also present in (iPrCNC)FeBr2, which is notably absent from (iPrPDI)FeBr2. Mayer bond order (MBO) and charge donation analyses were conducted for S = 1 (iPrCDA)FeBr2 as well as S = 2
(iPrPDI)FeBr2 and (iPrCNC)FeBr2 to further probe the nature of the metal−ligand bonding (Table 2). Consistent with the low-spin state observed for (iPrCDA)FeBr2, this complex has the largest MBO of the series for the sum of the Fe−N and Fe−C MBOs of the ligating atoms of the pincer (MBO = 2.198). The larger bond order for this compound is primarily due to the large MBO for the Fe−C bond of 1.059, which is by far the largest MBO observed for an Fe−L bond in this series. By comparison, (iPrCNC)FeBr2 has the second largest summed pincer Fe−L MBO of 1.542, due to the large Fe−C MBOs (0.594 and 0.657) compared to the Fe−Nimine MBOs in (iPrPDI)FeBr2 (0.393 and 0.397). Interestingly, the MBO analysis of (iPrCNC)FeBr2 demonstrates a substantially reduced Fe−Npyr MBO to the central N of the pincer (0.291) compared to the Fe−N MBOs for all the Fe−N interactions in (iPrPDI)FeBr2. Importantly, the order of the MBOs for the Fe−pincer interactions (iPrCDA > iPrCNC > iPrPDI) correlates well to the order of the ligand-field splittings observed in these D
DOI: 10.1021/acs.organomet.6b00651 Organometallics XXXX, XXX, XXX−XXX
Article
Organometallics Table 2. Mayer Bond Order and Charge Donation Analyses for (Pincer)FeBr2 Complexesa charge donation analysis (α + β)
Mayer bond order complex
Fe−L
Fe−Br
donation: (pincer → FeBr2) −
back-donation: (FeBr2 → pincer) −
net charge donation to FeBr2
( CDA)FeBr2
1
0.572 1.059 0.567
0.609 0.599
1.613 e (α: 0.872 e−, β: 0.741 e−)
0.910 e (α: 0.187 e−, β: 0.723 e−)
0.703 e−
(iPrCDA)FeBr2
2
0.366 0.775 0.356
0.688 0.593
0.954 e− (α: 0.370 e−, β: 0.584 e−)
0.366 e− (α: 0.108 e−, β: 0.258 e−)
0.588 e−
(iPrPDI)FeBr2
2
0.393 0.447 0.397
0.584 0.711
0.806 e− (α: 0.323 e−, β: 0.483 e−)
0.309 e− (α: 0.074 e−, β: 0.235 e−)
0.497 e−
(iPrCNC)FeBr2
2
0.594 0.291 0.657
0.666 0.520
1034 e− (α: 0.455 e−, β: 0.579 e−)
0.268 e− (α: 0.096 e−, β: 0.172 e−)
0.766 e−
iPr
a
spin
Numbers in bold indicate the central atom in the pincer ligand that is bound to the metal complex.
( iPr PDI)FeBr 2 , further supporting the increased β dπ population in S = 1 as the origin of the significant backdonation in this complex. Due to much more significant backdonation present in S = 1 (iPrCDA)FeBr2 compared to S = 2 (iPrCNC)FeBr2, the latter complex exhibits the largest net charge donation from pincer to FeBr2 in the series.
complexes by MCD spectroscopy, which ultimately results in (iPrCDA)FeBr2 being an intermediate spin complex, while (iPrPDI)FeBr2 and (iPrCNC)FeBr2 are high spin complexes. Charge donation analysis indicates that iPrCDA is the strongest donating pincer of the series (1.613 e−), with iPrCNC the second most donating (1.034 e−) and iPrPDI being the least donating (0.806 e−). This ordering is consistent with the increased Fe−pincer summed MBO for ( iPr CNC)FeBr 2 compared to ( iPr PDI)FeBr2 due to the strong Fe−C interactions in (iPrCNC)FeBr2 compared to the corresponding Fe−N interactions in (iPrPDI)FeBr2. While back-donations from FeBr2 to the pincer ligands in both (iPrCNC)FeBr2 and (iPrPDI)FeBr2 are similarly small, significant back-bonding is observed for (iPrCDA)FeBr2 (0.910 e−) that predominately derives from an increase in the β back-donation. This finding is consistent with the increased occupied β dπ-orbital population in S = 1 (iPrCDA)FeBr2 compared to the S = 2 complexes. In fact, the highest occupied β d-orbital (dxz) in (iPrCDA)FeBr2 is involved in π-back-bonding to the CDA pincer in this complex. Analysis of the bonding in a theoretical S = 2 (iPrCDA)FeBr2 model compared to S = 2 (iPrPDI)FeBr2 and (iPrCNC)FeBr2 provides further insight into the bonding differences in iPrCDA compared to iPrPDI and iPrCNC that lead to an S = 1 ground state for (iPrCDA)FeBr2. Notably, for the S = 2 series the Fe−C bond in (iPrCDA)FeBr2 has the highest MBO (0.775) of any of the Fe−L bonds in the S = 2 complexes (Table 2). While reduced compared to the corresponding Fe−C MBO in S = 1 (iPrCDA)FeBr2 due to the partial occupation of the σ* x2−y2 d orbital in the S = 2 model, it is clear that the carbene moitety of iPr CDA provides for a much stronger bonding interaction than any other ligating atom in the S = 2 series. Analysis of the calculated MO diagram for S = 2 (iPrCDA)FeBr2 (see SI) demonstrates that, similar to the S = 1 complex, the central carbene of iPrCDA is involved in a σ*-bonding interaction with dx2−y2. While the overall d-orbital ordering is similar between all three S = 2 complexes, this strong donor interaction in (iPrCDA)FeBr2 results in an increased energy of α dx2−y2 compared to ( iPrPDI)FeBr2 and (iPrCNC)FeBr2, which significantly contributes to (iPrCDA)FeBr2 favoring an S = 1 ground state. Lastly, charge donation analysis of the S = 2 (iPrCDA)FeBr2 model (Table 2) indicates a similar degree of back-donation to that observed in S = 2 (iPrCNC)FeBr2 and
■
DISCUSSION Rational catalyst development with NHC-containing pincer complexes of iron(II) requires a fundamental understanding of how introducing NHC moieties into pincer ligands affects electronic structure and bonding in their resulting iron(II) complexes. Further motivated by the unusual intermediate spin states observed in (iPrCDA)FeCl2 and (iPrCDA)FeBr2, MCD and DFT investigations were performed in this study in order to define the ligand and bonding contributions that determine electronic structure in S = 1 (iPrCDA)FeBr2 versus S = 2 (iPrPDI)FeBr2 and S = 2 (iPrCNC)FeBr2. One key question is the quantitative effects that σ-donation has on electronic structure of iron(II) complexes when NHC moieties are introduced to replace nitrogen donor atoms in pincer ligands. For the S = 2 complexes, (iPrPDI)FeBr2 and (iPrCNC)FeBr2, it is observed that the presence of two NHC moieties in the iPrCNC pincer ligand leads to a larger ligand field for (iPrCNC)FeBr2 compared to (iPrPDI)FeBr2 (10Dq ≈ 10 555 and ∼7700 cm−1, respectively). The increase in the ligand field for (iPrCNC)FeBr2 directly correlates to the increased donor ability of the NHC moieties in the iPrCNC ligand, as both MBO and charge donation analysis indicate a substantial increase in σ donation upon NHC substitution for imine nitrogens in iPrCNC versus iPrPDI. For both complexes, the highest unoccupied β d-orbital is dx2−y2, which is involved in σ* interaction with the pyridine and imine moieties (3 N) of the iPrPDI pincer or the NHC (2 C) and pyridine (1 N) moieties of the iPrCNC ligand. Importantly, charge donation analysis indicates similar π-back-bonding is present in both (iPrPDI)FeBr2 and (iPrCNC)FeBr2, and, hence, the dominant contribution to the ligand-field variation in these complexes is the difference in σ-donor abilities between these two pincer ligands. Nevertheless, the increased ligand field observed in (iPrCNC)FeBr2 remains insufficient to promote an intermediate-spin (S = 1) ground state. The increased σ-donation E
DOI: 10.1021/acs.organomet.6b00651 Organometallics XXXX, XXX, XXX−XXX
Article
Organometallics observed upon replacing the pyridine moiety in iPrPDI with an NHC moiety so as to form iPrCDA complexes yields a significant increase in the energy of dx2−y2 and, hence, a larger overall ligand field due to the stronger Fe−C interaction in (iPrCDA)FeBr2. In fact, this Fe−C interaction is the strongest Fe−L interaction observed across this series of complexes, as evidenced by it exhibiting the highest MBO (Fe−C, 1.059), consistent with the short Fe−C bond observed experimentally and the strongest pincer σ-donation to iron (1.613 e− in (iPrCDA)FeBr2 vs 1.034 e− and 0.806 e− in (iPrCNC)FeBr2 and (iPrPDI)FeBr2, respectively). An interesting observation in NHC-containing pincer complexes of iron(II) is that the ultimate spin state of a complex (S = 1 vs S = 2) does not simply correlate to the number of NHC moieties in the pincer ligand. This fact is true despite the findings from these spectroscopic studies, which demonstrate that introduction of an NHC moiety results in strong Fe−pincer interactions and increased σ-donation to iron. In this study, this effect is muddled when comparing (iPrCDA)FeBr2 and (iPrCNC)FeBr2, where the former contains a single NHC moiety resulting in an S = 1 ground state, whereas the latter contains two NHC moieties yet remains S = 2. Evaluation of the MBO and charge donation analysis provides insight into the origins of these differences, namely, that the single NHC in iPrCDA is a much stronger σ-donor and results in a uniquely strong Fe−C interaction across this series of complexes. Importantly, this Fe−C interaction also has the highest Fe−L MBO interaction when comparing a theoretical S = 2 (iPrCDA)FeBr2 model to S = 2 (iPrPDI)FeBr2 and (iPrCNC)FeBr2. Ultimately, it is this unique Fe−C bonding interaction accessible with the iPrCDA ligand that leads to (iPrCDA)FeBr2 favoring an S = 1 ground state. One contribution that results in this strong Fe−C bonding interaction with the iPrCDA ligand is likely the position of the NHC moieties in the pincer ligands. For example, the central position of the NHC in (iPrCDA)FeBr2 not only leads to a significant σ* interaction with dx2−y2 but also results in an Fe−C π interaction with dxz that is not present between iron and the two NHC moieties in (iPrCNC)FeBr2. An additional contribution to the reduced σ-donating capabilities of the pincer ligand in (iPrCNC)FeBr2 is likely geometric in nature. Specifically, this complex displays the weakest Fe−L bond to a pincer atom by MBO analysis of any complex in the series investigated, as the central N ligand of iPrCNC is much more weakly bound than that in the iPrPDI complex (0.291 vs 0.447, respectively). This weakened central Fe−N interaction in (iPrCNC)FeBr2 may be the consequence of the geometric constraints of the pincer ligand, which might limit the ability to form strong interactions to both the flanking NHC and central pyridine moieties. The net effect is a reduction in the total pincer donation to iron. An alternative explanation for the differences between iPr CNC and iPrCDA ligands is that the former contains unsaturated NHCs, while the latter contains a saturated NHC. Although there is not a homologue available to directly study this effect in this series, previous studies of fourcoordinate S = 2 iron(II)-NHC complexes indicated no significant difference in ligand field or electronic structure as a function of NHC backbone saturation.42 Hence, the degree of NHC saturation is unlikely to be a significant contributor to the electronic structure differences observed between (iPrCNC)FeBr2 and (iPrCDA)FeBr2. Thus, these results indicate that the design of NHC-containing pincers to target specific electronic
structures in their iron adducts must take into account not only the number of NHC moieties but also their overall orientation in the pincer and the geometric constraints imposed by the pincer architecture.
■
CONCLUSION Due to the continuing importance of NHC-containing pincer complexes of iron(II) in catalysis,43−45 MCD and DFT studies were utilized to obtain detailed insight into electronic structure and bonding in a series of structurally related complexes: (iPrCDA)FeBr2, (iPrPDI)FeBr2, and (iPrCNC)FeBr2. These studies provide the first direct and quantitative evaluations of the comparative electronic structures and bonding present. When combined with MBO and charge donation analysis investigations, a direct correlation was found between increasing Fe−L bonding and pincer-donating abilities when NHC moieties are introduced in the pincer ligand. However, the net donor abilities of the pincer ligands and the strength of the Fe−pincer interaction do not directly correlate to the number of NHC moieties present. Equally, if not more important, is the orientation of the NHC moiety and the geometric constraints of the pincer architecture in modulating iron-pincer bonding. These findings illustrate the importance of understanding not only the inherent σ-donating characteristics of an NHC moiety but also its orientation in a ligand architecture so as to maximize metal−ligand interactions. Such criteria are important to keep in mind when designing ligands to fully take advantage of the electronic features of the NHC functional group.
■
EXPERIMENTAL SECTION
General Considerations. All reagents were purchased from commercial sources. All air- and moisture-sensitive synthetic manipulations were performed using an MBraun inert atmosphere (N2) drybox or by standard Schlenk techniques.46 All preparations of spectroscopy samples were conducted in an MBraun inert atmosphere (N2) drybox equipped with a direct liquid nitrogen inlet line. Anhydrous solvents were further dried using activated alumina, 4 Å molecular sieves and stored under an inert atmosphere over molecular sieves or were dried by passage over a column of alumina under an atmosphere of argon. NMR spectra were recorded on a Varian spectrometer operating at 500 MHz for 1H NMR or 125 MHz for 13C NMR. The line listings for the NMR spectra of diamagnetic complexes are reported as chemical shift in ppm (number of protons, splitting pattern, coupling constant). The line listing for the NMR spectrum of the paramagnetic complex is reported as chemical shift in ppm (number of protons, peak width at half height in Hz). IR spectra were recorded on a Bruker ALPHA Platinum ATR infrared spectrometer. High-resolution mass spectra were obtained at the Boston College Mass Spectrometry Facility. Elemental analyses were conducted at Robertson Microlit Laboratories. The iron complexes (iPrPDI)FeBr247 and (iPrCNC)FeBr248 were prepared according to previously reported procedures. (iPrCDA)FeBr2 was prepared similarly to the previous reported synthesis of (iPrCDA)FeCl2 with slight modifications that are elucidated in the Synthetic Methods section. Synthetic Methods. N-(2,6-Diisopropylphenyl)acetimidoyl Bromide. This reaction was performed on a Schlenk line and open to a mercury bubbler to avoid overpressurization of the reaction vessel. In a 250 mL, two-neck round-bottom flask equipped with a stir bar and a 180° joint, 2,6-lutidene (2.78 mL, 23.9 mmol) was added to a solution of N-(2,6-diisopropylphenyl) acetamide (5.00 g, 22.8 mmol) in dichloromethane (22.8 mL, 1.0 M). The clear, pale yellow solution was cooled to 0 °C, and oxalyl bromide (2.25 mL, 23.9 mmol) was added as rapidly as possible while accounting for the vigorous gas evolution of the reaction. The resulting cloudy, orange solution was stirred at 0 °C for 30 min. The solvent was removed and the remaining F
DOI: 10.1021/acs.organomet.6b00651 Organometallics XXXX, XXX, XXX−XXX
Article
Organometallics yellow-orange solids were taken into a nitrogen-filled glovebox. Hexanes (100 mL) was added to the solid, and the precipitate was removed by filtering through Celite. The solvent was removed from the filtrate, and the resulting oil was dissolved in 100 mL of pentane. After a second filtration through Celite, the solvent was removed to give a yellow oil. The crude reaction mixture was purified under inert conditions by Kugelrohr distillation at 1 × 10−4 Torr and 150 °C. The product was obtained with a small amount (