Kathlyn L.
Fillman
a,
Jacob A.
Przyojski
b,
Malik H.
Al-Afyouni
a,
Zachary J.
Tonzetich
b and
Michael L.
Neidig
*a
aDepartment of Chemistry, University of Rochester, Rochester, New York 14627, USA. E-mail: neidig@chem.rochester.edu
bDepartment of Chemistry, University of Texas at San Antonio, San Antonio, Texas 78249, USA
First published on 10th November 2014
Iron salts and N-heterocyclic carbene (NHC) ligands is a highly effective combination in catalysis, with observed catalytic activities being highly dependent on the nature of the NHC ligand. Detailed spectroscopic and electronic structure studies have been performed on both three- and four-coordinate iron(II)–NHC complexes using a combined magnetic circular dichroism (MCD) and density functional theory (DFT) approach that provide detailed insight into the relative ligation properties of NHCs compared to traditional phosphine and amine ligands as well as the effects of NHC backbone structural variations on iron(II)–NHC bonding. Near-infrared MCD studies indicate that 10Dq(Td) for (NHC)2FeCl2 complexes is intermediate between those for comparable amine and phosphine complexes, demonstrating that such iron(II)–NHC and iron(II)–phosphine complexes are not simply analogues of one another. Theoretical studies including charge decomposition analysis indicate that the NHC ligands are slightly stronger donor ligands than phosphines but also result in significant weakening of the Fe–Cl bonds compared to phosphine and amine ligands. The net result is significant differences in the d orbital energies in four-coordinate (NHC)2FeCl2 complexes relative to the comparable phosphine complexes, where such electronic structure differences are likely a significant contributing factor to the differing catalytic performances observed with these ligands. Furthermore, Mössbauer, MCD and DFT studies of the effects of NHC backbone structure variations (i.e. saturated, unsaturated, chlorinated) on iron–NHC bonding and electronic structure in both three- and four-coordinate iron(II)–NHC complexes indicate only small differences as a function of backbone structure, that are likely amplified at lower oxidation states of iron due to the resulting decrease in the energy separation between the occupied iron d orbitals and the unoccupied NHC π* orbitals.
A wide variety of applications of iron–NHC complexes to catalytic transformations have begun to emerge, including applications in hydrosilylation, carbometallation, cyclization, aziridination and allylic substitution reactions, demonstrating the significant catalytic potential of iron–NHC complexes.7–9 Of particular note, the combination of simple iron salts and NHCs results in the in situ generation of catalytically active iron–NHC species for C–C cross-coupling reactions including aryl–aryl, alkyl–aryl and alkyl–alkyl couplings (Scheme 1). NHC ligands have been particularly successful for iron cross-coupling as the aryl–aryl system is currently the best performing iron-based system for heteroaryl coupling and the alkyl–alkyl system is currently the only iron-based cross-coupling system capable of C(sp3)–C(sp3) couplings in the presence of functional groups.
Within the catalytic applications of iron–NHCs, the nature of the specific NHC ligand that provides for the most effective catalysis varies significantly from one system to another. For example, the optimal NHC for aryl–aryl cross-coupling catalysis is SIPr where the corresponding, unsaturated IPr ligand was found to be much less effective (Scheme 1).23 By contrast, for the cross-coupling of non-activated chloroalkanes and aryl Grignards in Kumada-type couplings, the inverse of the NHC dependence on activity is observed with IPr outperforming SIPr.24 However, these dramatic disparities in reactivity due to differences in NHC backbone saturation are not observed in alkyl–alkyl cross-coupling where similar product yields can be obtained using either IMes or SIMes.25 In the aryl–alkyl cross-coupling system of Bedford and co-workers, NHCs lacking N-aryl substituents (Cy, tBu) outperformed IMes.26 Combined, these studies suggest that the NHC ring structure (i.e. saturated vs. unsaturated vs. substituted) and N-substitution may yield important differences in iron–NHC bonding, in situ iron–NHC speciation and, hence, iron–NHC reactivity.
Despite the observed dependence of catalytic performance on the NHC ligand structure, a detailed understanding of iron–NHC σ- and π-bonding and the effects of NHC ring perturbations on iron–NHC bonding is critically underdeveloped. In fact, detailed investigations of iron–NHC bonding have been limited to piano stool type iron(II) complexes with both cyclopentadienyl (Cp) and CO ligation where a combination of IR, electrochemical and theoretical methods suggested that the NHC ligand in these complexes can serve as both a σ-donor and moderate π-acceptor.27 However, fundamental insight into iron–NHC bonding in more electron deficient, high-spin iron complexes is lacking. This deficiency stands in stark contrast to precious metal NHC systems where IR studies of supporting CO ligands and DFT investigations have led to the general view of NHC ligands as strong σ-donors (stronger than phosphine ligands) and weak π-acceptors, where the extent of π-bonding is dependent on the nature of the metal and supporting ligands.28–43 While the general views of metal–NHC bonding from the precious metal systems currently drive much of the work in iron–NHC systems, rational catalyst development with iron–NHCs necessitates a fundamental understanding of the effects of NHC variations on electronic structure and bonding in paramagnetic iron systems that may be catalytically relevant. Such studies can also provide fundamental insight into the differences in electronic structures of iron–NHC and iron–phosphine complexes lacking CO ligation that may be relevant to catalysis, including iron-catalyzed cross-coupling. Importantly, iron–NHC bonding in high-spin systems may differ significantly compared to the low-spin, Cp and CO bound species more commonly investigated.
Towards this goal, an approach combining magnetic circular dichroism (MCD) studies and density functional theory (DFT) investigations of well-defined iron(II)–NHC complexes has been utilized to directly investigate electronic structure and bonding in high-spin iron(II)–NHC complexes. The results provide direct insight into the ligand-field strength of NHC ligands compared to amine and phosphine ligands, the effects of NHC ring variations on bonding and the extent of donation and back donation contributions to bonding in iron(II)–NHC complexes as a function of coordination number and geometry.
Further calculations of molecular orbitals (MOs) and TD-DFT used the B3LYP functional with the TZVP basis set on all atoms. MO compositions and analyses were calculated using the AOMix program.55,56 Atomic charges and spin densities were calculated using Mulliken population analysis (MPA). Orbitals from the Gaussian calculations were plotted with the ChemCraft program. TD-DFT was used to calculate the electronic transition energies and intensities of the 30–40 lowest-energy states. The analysis of the MO compositions in terms of fragment orbitals, Mayer bond orders, total overlap populations and the charge decomposition analysis (CDA)57,58 were performed using AOMix-FO.55 CDA and its applications have been previously described in detail by Gorelsky and co-workers.59,60
Spin unrestricted DFT calculations were used to further analyze the electronic structure of (IMes)2FeCl2. Geometry optimization with B3LYP/TZVP and the GD3BJ dispersion correction yielded overall structural features, bond lengths and angles in good agreement with those observed by crystallography (Table 1). The optimized (IMes)2FeCl2 complex is best described as a distorted tetrahedral complex with Fe–IMes bond lengths of 2.161 Å and 2.163 Å, Fe–Cl bond lengths of 2.334 Å, a C(IMes)–Fe–C(IMes) bond angle of 126.36° and a Cl–Fe–Cl bond angle of 107.45°. This optimized geometry correlates well with the literature structure, which has Fe–IMes bond lengths of 2.139 Å and 2.157 Å, Fe–Cl bond lengths of 2.310 Å and 2.292 Å, a C(IMes)–Fe–C(IMes) bond angle of 125.20° and a Cl–Fe–Cl bond angle of 106.66°. Both the experimental and computational studies of (IMes)2FeCl2 are indicative of a high-spin iron(II) complex (S = 2). The molecular orbitals and their corresponding energies as well as electronic transition energies were calculated from the optimized structure.
Complex | Fe–L1 (Å) | Fe–L2 (Å) | Fe–Cl1 (Å) | Fe–Cl2 (Å) | L1–Fe–L2 (°) | |||||
---|---|---|---|---|---|---|---|---|---|---|
Exp. | Calc. | Exp. | Calc. | Exp. | Calc. | Exp. | Calc. | Exp. | Calc. | |
(IMes)2FeCl2 | 2.139 | 2.161 | 2.157 | 2.163 | 2.310 | 2.334 | 2.292 | 2.334 | 125.20 | 126.36 |
(tmpn)FeCl2 | 2.140 | 2.187 | 2.140 | 2.187 | 2.271 | 2.296 | 2.239 | 2.275 | 99.35 | 97.04 |
(teeda)FeCl2 | 2.192 | 2.229 | 2.150 | 2.204 | 2.230 | 2.268 | 2.240 | 2.284 | 83.75 | 83.60 |
(PMe3)2FeCl2 | 2.430 | 2.449 | 2.427 | 2.445 | 2.240 | 2.280 | 2.235 | 2.278 | 102.78 | 104.43 |
(PPh3)2FeCl2 | 2.476 | 2.482 | 2.476 | 2.482 | 2.219 | 2.271 | 2.219 | 2.271 | 111.25 | 110.07 |
The electronic ground state of (IMes)2FeCl2 is described by the frontier molecular orbitals (FMOs) with focus on the unoccupied MOs, in conjunction with their occupied counterparts, to showcase the major contributions to bonding. The MOs and the corresponding energy diagram are shown in Fig. 2. In the β manifold, the highest occupied molecular orbital (HOMO) along with the unoccupied orbitals β-193, β-194, β-195, and β-208 are comprised mostly of Fe d contributions, slightly mixed with NHC and Cl orbital contributions. The Fe 3d orbitals listed in order of increasing energy are dx2−y2 (β-192), dz2 (β-193), dxz (β-194), dyz (β-195), dxy (β-208). In addition, there are FMOs that represent an occupied MO of both IMes (with backbone contributions) and Cl character (β-185), an occupied MO of Mes (NHC side chains) and Cl character (β-187), and an unoccupied MO with some d character as well as a σ bond with IMes (β-200).
Charge decomposition (CDA), fragment molecular orbital (FO) and Mayer bond order (MBO) analyses were completed for the (IMes)2FeCl2 complex. The MOs of a complex can be described as linear combinations of the occupied and unoccupied MOs of defined molecular fragments, termed fragment molecular orbitals or FOs. In CDA, donation and back donation between different molecular fragments can be evaluated based on the overlap and coefficients of the MO–LCFO matrix (where LCFO = linear combination of fragment molecular orbitals). In the case of (IMes)2FeCl2, CDA was completed to quantify the total charge donation and back donation between the Fe–Cl2 fragment and two IMes ligand fragments (i.e. 3 total fragments). When the Fe–Cl2 fragment and IMes ligand fragments are combined to form the complex, there is a mixing of the occupied fragment orbital of the donor with the unoccupied fragment orbital of the acceptor which allows for the transfer of electron density (charge donation) from the donor to the acceptor. From the CDA, it can be seen that there is a total charge donation (IMes2 → Fe–Cl2) of 0.954 electrons and total back donation (Fe–Cl2 → IMes2) of 0.391 electrons. This results in a net charge donation to Fe–Cl2 of 0.563 electrons. The FO analysis provides information about changes in occupancies of the fragment molecular orbitals of the donor and acceptor upon the complex formation so that the orbitals involved in the donation and back donation can be readily identified (i.e. the specific orbitals that change their occupancies). In the α manifold, the charge donation from the IMes ligand fragments to the Fe–Cl2 fragment occurs from the highest occupied fragment orbital (HOFO) of the IMes fragments, an IMes σ orbital, to two unoccupied fragment orbitals on the Fe–Cl2 fragment of iron 4s and 4p character (20% occupancy change per NHC ligand, see Fig. 3). A slightly larger occupancy change is observed for charge donation in the β manifold (22% per NHC ligand), where donation occurs again from the HOFO (IMes σ orbital) of the IMes fragments to several unoccupied FOs of the Fe–Cl2 fragment that have significant iron d character. In the β manifold, there is also a significant electronic polarization, particularly in the Fe–Cl2 fragment, which allows for a redistribution of charge from an occupied Fe–Cl2 fragment orbital with mostly d character to unoccupied fragment orbitals of the Fe–Cl2 fragment (see Fig. 3). Furthermore, the FO analysis indicates the presence of back donation in both the α and β manifolds from occupied Fe–Cl2 fragment orbitals to unoccupied IMes fragment orbitals (i.e. ∼9.6% and ∼10.1% occupancy changes for each IMes fragment in both the α and β manifolds, respectively). However, in contrast to the donation in this system, the occupancy changes corresponding to back donation are spread over many occupied Fe–Cl2 and unoccupied IMes FOs. For this distorted tetrahedral complex, the Mayer bond order (MBO) between Fe and each IMes is 0.760 and 0.767. In both cases, the β contribution to the bond order is greater than that of the α, 0.390 and 0.393 for β and 0.370 and 0.374 for α. When compared to the occupied α-spin MOs, the occupied β-spin MOs contribute more to the overall bond order due to a small increase in σ donation from the NHCs to the unoccupied β-spin orbitals on Fe.
Fig. 3 FO analysis of charge donation in (IMes)2FeCl2. Occupancy changes (α (red) and β (blue)) and orbital compositions are indicated for each FO. Arrows indicate donation to Fe–Cl2 FOs. |
The 5 K, 7 T NIR MCD spectra of the L2FeCl2 complexes investigated are given in Fig. 4 and the d–d transition energies and 10Dq(Td) values are summarized in Table 2. The diamine–iron(II)–dichloride complex (tmpn)FeCl2 (tmpn = N,N,N′,N′-tetramethylpropane-1,3-diamine) contains very low energy LF transitions at 5260 and 6140 cm−1 (10Dq(Td) = 5700 cm−1). Similarly, (teeda)FeCl2 (teeda = N,N,N′,N′-tetraethylethylenediamine) gives LF transitions at 5290 and 6310 cm−1 (10Dq(Td) = 5800 cm−1). Both NHC ligated tetrahedra gave slightly higher energy LF bands, where (ClIMes)2FeCl2 exhibited LF transitions at 5520 cm−1 and 6540 cm−1 (10Dq(Td) = 6030 cm−1), similar to those observed for (IMes)2FeCl2 (10Dq(Td) = 5980 cm−1). The phosphine–iron(II)–dichloride complexes yielded the highest energy ligand-field transitions, with (PPh3)2FeCl2 exhibiting LF bands at 5590 cm−1 and 7590 cm−1 (10Dq(Td) = 6590 cm−1) and (PMe3)2FeCl2 at 6340 cm−1 and 7600 cm−1 (10Dq(Td) = 6970 cm−1). In contrast to the NHC and diamine complexes, both phosphine species exhibit a pseudo-A term in their LF MCD transitions, where such pseudo-A terms in MCD (i.e. a pair of temperature-dependent C-terms with opposite sign) arise from spin orbit coupling (SOC) between two excited states that are close in energy to which two orthogonal transitions occur from a single ground state.64–66 While the physical origin of the pseudo-A term is currently elusive in the phosphine complexes as it requires a detailed understanding of the SOC mechanism, it is noteworthy that previous MCD studies of distorted tetrahedral bisphosphine complexes have also exhibited pseudo-A term LF transitions67 and the generality and origin of this behavior for tetrahedral phosphine complexes will be a focus of future study.
Complex | LF transitions | 10Dq(Td) |
---|---|---|
(tmpn)FeCl2 | 5260, 6140 cm−1 | 5700 cm−1 |
(teeda)FeCl2 | 5290, 6310 cm−1 | 5800 cm−1 |
(IMes)2FeCl2 | 5440, 6520 cm−1 | 5980 cm−1 |
(ClIMes)2FeCl2 | 5520, 6540 cm−1 | 6030 cm−1 |
(PPh3)2FeCl2 | 5590, 7590 cm−1 | 6590 cm−1 |
(PMe3)FeCl2 | 6340, 7600 cm−1 | 6970 cm−1 |
To obtain further insight into iron(II)–NHC bonding compared to phosphine and amine ligands, CDA and MBO analyses were completed for the L2FeCl2 complexes. The MBO between Fe and the NHCs in (IMes)2FeCl2 is the highest of the series (see Table 3). Notably, the MBOs between Fe and PPh3 or PMe3 are slightly smaller. Lower still are the MBOs for the Fe–diamine bonds in (tmpn)FeCl2 and (teeda)FeCl2, 0.362 and 0.362 for (tmpn)FeCl2 and 0.352 and 0.336 for (teeda)FeCl2. However, the MBO between Fe and Cl is much higher in both the phosphine and diamine complexes when compared to the NHC complex. This is indicative of a strong trans-type influence of the NHC ligand that causes a pronounced weakening of the Fe–Cl bond. This influence can be seen experimentally in the (IMes)2FeCl2 crystal structure where the Fe–Cl bonds are elongated (avg Fe–Cl = 2.30 Å) in comparison to the analogous bisphosphine (avg Fe–Cl = 2.23 Å) and diamine (avg Fe–Cl = 2.26 Å) complexes (see Table 1). The calculated total overlap population (TOP) between the Fe and Cl moieties in (IMes)2FeCl2 (avg Fe–Cl = 0.39) is significantly lower than those calculated for the bisphosphine and diamine complexes (avg Fe–Cl = 0.49 and 0.50, respectively). The amount of electron density between the Fe and Cl atoms in each complex displays an inverse relationship with the amount of electron density donated by each L-type ligand (Table 3). This correlation suggests that the strongest donor to the distorted tetrahedral Fe(II) center will cause the most pronounced weakening of the Fe–Cl bond. This rationalization using TOP values is consistent with the perturbation approach of Burdett and Albright,68 which shows that systems of lowered symmetry can, through orbital mixing, exhibit metal–ligand bond weakening effects similar to the trans influence typically seen in square-planar and octahedral complexes. The results of the CDA show that the complex with the highest net charge donation to the Fe–Cl2 fragment from the L-type ligand fragments is (IMes)2FeCl2 with a net charge donation of 0.563 electrons, followed by the two phosphine complexes, (PMe3)2FeCl2 and (PPh3)2FeCl2, with total net charge donations of 0.493 and 0.472 electrons, respectively. The diamine complexes, (tmpn)FeCl2 and (teeda)FeCl2, have comparatively lower total net charge donations of 0.275 and 0.250 electrons, respectively. In addition, there is significant back donation in (IMes)2FeCl2, 0.391 electrons, as well as (PMe3)2FeCl2 and (PPh3)2FeCl2, 0.222 and 0.294 electrons, respectively.
Complex | Mayer bond order | Charge decomposition analysis (α + β) | |||
---|---|---|---|---|---|
Fe–L | Fe–Cl | Donation: (L2 → Fe–Cl2) | Back donation: (Fe–Cl2 → L2) | Net charge donation to Fe–Cl2 | |
(IMes)2FeCl2 | 0.760 | 0.659 | 0.954 e− | 0.391 e− | 0.563 e− |
0.767 | 0.660 | ||||
(tmpn)FeCl2 | 0.362 | 0.752 | 0.427 e− | 0.152 e− | 0.275 e− |
0.362 | 0.792 | ||||
(teeda)FeCl2 | 0.352 | 0.763 | 0.424 e− | 0.174 e− | 0.250 e− |
0.336 | 0.797 | ||||
(PMe3)2FeCl2 | 0.644 | 0.776 | 0.715 e− | 0.222 e− | 0.493 e− |
0.641 | 0.774 | ||||
(PPh3)2FeCl2 | 0.663 | 0.799 | 0.766 e− | 0.294 e− | 0.472 e− |
0.663 | 0.799 |
Complex | Mössbauer | LF bands (cm−1) | ZFS parameters | |||
---|---|---|---|---|---|---|
δ (mm s−1) | ΔEQ (mm s−1) | D (cm−1) | |E|(cm−1) | |E/D| | ||
(IPr)Fe(CH2TMS)2 | 0.34 | 1.04 | 6660, 9260 | −20 ± 2 | 3.8 ± 0.5 | 0.19 ± 0.02 |
(SIPr)Fe(CH2TMS)2 | 0.35 | 1.12 | 6350, 9110 | −20 ± 2 | 4 ± 0.5 | 0.20 ± 0.02 |
(CIIPr)Fe(CH2TMS)2 | 0.33 | 1.08 | 6520, 9100 | −18 ± 2 | 3.4 ± 0.5 | 0.19 ± 0.02 |
Spin unrestricted DFT calculations were used to further analyze the electronic structures of (IPr)Fe(CH2TMS)2, (SIPr)Fe(CH2TMS)2, and (ClIPr)Fe(CH2TMS)2. Geometry optimizations with B3LYP/TZVP yielded overall structural features, bond lengths and angles in good agreement with those observed by crystallography (Table 5; note that no crystal structure is available for (ClIPr)Fe(CH2TMS)2). Both experimental and computational studies of the series of 3C NHC complexes are indicative of high-spin iron(II) complexes (S = 2). Details of the MO analyses are given in the ESI.† Importantly, CDA and MBO analyses were completed for the series of 3C Fe(II) NHC complexes (Table 6). From the CDA, it can be seen that the total donation from the NHC fragment to the Fe–(CH2TMS)2 fragment is the highest for (IPr)Fe(CH2TMS)2, suggesting that this complex is the strongest σ-donor of the series, followed by (SIPr)Fe(CH2TMS)2, and (ClIPr)Fe(CH2TMS)2 (see Table 6). While it can also be seen that (SIPr)Fe(CH2TMS)2 has the highest back donation of the series, the overall back donation to the NHC fragment (average = 0.076 electrons) for all three complexes is small when compared to the total donation to the Fe–(CH2TMS)2 fragment (average = 0.444 electrons). For (IPr)Fe(CH2TMS)2, the MBO between Fe and IPr is 0.667, with α and β contributions of 0.341 and 0.326, respectively. Similarly, the MBO between Fe and the NHC for (SIPr)Fe(CH2TMS)2 and (ClIPr)Fe(CH2TMS)2 are 0.664 and 0.625, respectively. The small disparities in the σ donation abilities of the NHC ligands in the 3C complexes derive from slightly different MBOs of the Fe–C(NHC) bonds and the C(NHC)–N(NHC) bonds. In (SIPr)Fe(CH2TMS)2, the C(NHC)–N(NHC) bonds are relatively strong (average MBO = 1.168) which weakens the Fe–C(NHC) bond, thereby making it a poorer σ donor. The C(NHC)–N(NHC) bond in (IPr)Fe(CH2TMS)2 is weaker than that of the SIPr complex (average MBO = 1.130), consequently causing a stronger Fe–C(NHC) bond and making the IPr complex an overall better σ donor. Overall, while slight differences in iron–NHC bonding are present as a function of the NHC backbone structure, both experimental and theoretical studies indicate that these differences are relatively small in this series of three-coordinate iron(II) complexes. In contrast to previous proposals that backbone substituents can greatly affect bonding interactions in Fe–NHC complexes,7 this result is consistent with our studies of 4C complexes where the electronic structure effects of IMes vs.ClIMes are minimal for S = 2 iron(II).
Complex | Fe–NHC (Å) | Fe–(CH2TMS)1 (Å) | Fe–(CH2TMS)2 (Å) | (CH2TMS)1–Fe–(CH2TMS)2 (°) | NHC–Fe–(CH2TMS)1 (°) | |||||
---|---|---|---|---|---|---|---|---|---|---|
Exp. | Calc. | Exp. | Calc. | Exp. | Calc. | Exp. | Calc. | Exp. | Calc. | |
(IPr)Fe(CH2TMS)2 | 2.164 | 2.182 | 2.060 | 2.096 | 2.062 | 2.096 | 123.09 | 126.06 | 118.46 | 116.97 |
(SIPr)Fe(CH2TMS)2 | 2.177 | 2.201 | 2.056 | 2.095 | 2.056 | 2.095 | 122.89 | 126.04 | 118.56 | 116.98 |
(CIIPr)Fe(CH2TMS)2 | — | 2.234 | — | 2.080 | — | 2.080 | — | 117.96 | — | 121.02 |
Complex | Mayer bond order analysis | Charge decomposition analysis (α + β) | ||||
---|---|---|---|---|---|---|
Fe–NHC | α contr. | β contr. | Donation: (NHC → Fe–(CH2TMS)2) | Back donation: (Fe–(CH2TMS)2 → NHC) | Net charge donation to Fe–(CH2TMS)2 | |
(IPr)Fe(CH2TMS)2 | 0.667 | 0.341 | 0.326 | 0.470 e− | 0.068 e− | 0.402 e− |
(SIPr)Fe(CH2TMS)2 | 0.664 | 0.336 | 0.327 | 0.439 e− | 0.082 e− | 0.357 e− |
(CIIPr)Fe(CH2TMS)2 | 0.625 | 0.330 | 0.295 | 0.422 e− | 0.077 e− | 0.345 e− |
Near-infrared MCD studies of distorted tetrahedral (IMes)2FeCl2 compared to a series of L2FeCl2 distorted tetrahedral (L = phosphine or amine) permit the first direct elucidation of 10Dq(Td) and hence, ligand field strength, of NHCs relative to phosphine and diamine ligands. From these studies, 10Dq(Td) for (IMes)2FeCl2 is found to be intermediate in magnitude relative to the corresponding phosphine (largest 10Dq(Td)) and diamine (smallest 10Dq(Td)) complexes. While the observed 10Dq(Td) values initially appear to contradict existing views of NHCs as stronger field ligands than phosphines,35,72 Mayer bond order and charge decomposition analyses indicate that the NHC is a stronger donor ligand than phosphines. The origin of the reduced 10Dq(Td) value with NHC ligation is found to reflect the significant weakening of the Fe–Cl bonds (leading to reduced Fe–Cl bond orders and reduced charge donation from Cl to Fe) in the NHC complexes relative to the comparable amine and phosphine complexes. While strong trans-type influences in metal–NHCs have been previously proposed,72 this study provides the first direct quantitative evaluation of this in terms of the resulting energies of both the occupied and unoccupied d orbitals. These effects are significant in L2FeCl2 complexes as indicated by the large differences in 10Dq(Td), clearly demonstrating that NHCs are not simple analogues of phosphine ligands for iron. Importantly, the differences in the electronic structures of iron–NHC vs. iron–phosphine complexes are likely a significant contributing factor to the differing catalytic performances observed with these ligands. For example, the variations in the d orbital energies (both occupied and unoccupied) could directly affect the reaction barriers for homolytic R–X cleavage or oxidative addition proposed in iron-catalyzed cross-coupling.73–75 Furthermore, the significant effect of iron–NHC coordination on the ligation strength of other ligands to iron, such as nucleophile-derived ligands in cross-coupling, could provide a pathway to modulate the reactivity of coordinated ligands that ultimately form new C–C bonds upon reaction with electrophiles. For example, it is anticipated that the trans-type influence due to the NHC ligation will result in a trans effect where the rate of Fe–Cl substitution/transmetalation will be increased. Similarly, for transmetalated iron species of the form (IMes)2FeRX or (IMes)2FeR2, the rate of Fe–R bond dissociation required for cross-coupled product generation should also be increased due to this effect.
Analogous studies of electronic structure and bonding as a function of NHC backbone structure (e.g. saturated, unsaturated, chlorinated) in both three- and four-coordinate iron(II)–NHC complexes were also performed due to the significant differences reported in catalytic systems as a function of NHC backbone structure. In these high-spin iron(II) systems (S = 2), minimal effects were observed as a function of NHC backbone structure. However, the small differences in back donation observed are likely amplified at lower oxidation states of iron due to the resulting decrease in the energy separation between the occupied iron d orbitals and the unoccupied NHC π* orbitals as the iron oxidation state is reduced below iron(II). While the iron(I)–NHC complex, (IMes)2FeCl, has been recently reported in the literature,76 iron(I)–NHC complexes with varied NHC backbone structures are not yet known. However, preliminary DFT and spectroscopic studies of (IMes)2FeCl are consistent with a significant increase in Fe–Cl → IMes2 backdonation upon reduction to iron(I). Once iron(I) complexes with varied NHC ligands are synthetically accessible, electronic structure studies can provide further insight into the effects of NHC structure on iron–NHC bonding as a function of iron oxidation state.
The continued application of the combined MCD and DFT approach employed herein to additional iron–NHC systems, including N-substituent variations in iron(II)–NHCs and iron(I)–NHCs as a function of coordination number, geometry and supporting ligands, should continue to expand our fundamental understanding of iron–NHC electronic structure and bonding. Ultimately, such studies will continue to provide critical insight into the molecular-level origins of variations in catalytic performance as a function of NHC structure with catalytically relevant supporting ligands.
Footnote |
† Electronic supplementary information (ESI) available: Supplemental MCD and Mössbauer data; MO diagrams, TD-DFT, optimized geometry coordinates and X-ray crystallographic details. CCDC 1023546–1023548. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c4sc02791d |
This journal is © The Royal Society of Chemistry 2015 |