III) and

Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/IC

Electronic and Structural Comparisons between Iron(II/III) and Ruthenium(II/III) Imide Analogs Kelly E. Aldrich,† B. Scott Fales,‡ Amrendra K. Singh,§ Richard J. Staples, Benjamin G. Levine,* John McCracken,* Milton R. Smith, III,* and Aaron L. Odom* Department of Chemistry, Michigan State University, 578 South Shaw Lane, East Lansing, Michigan 48824, United States

Downloaded via NOTTINGHAM TRENT UNIV on August 13, 2019 at 07:26:24 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: To examine structural and electronic differences between iron and ruthenium imido complexes, a series of compounds was prepared with different phosphine basal sets. The starting material for the ruthenium complexes was Ru(NAr/Ar*)(PMe3 )3 (Ru1/Ru1*), where Ar = 2,6(iPr)2C6H3 and Ar* = 2,4,6-(iPr)3C6H2, which were prepared from cis-RuCl2(PMe3)4 and 2 equiv of LiNHAr/Ar*. The starting materials for the iron complexes were the analogous Fe(NAr/Ar*)(PMe3)3 species (Fe1/Fe1*), which were not isolated but could be generated in situ from FeCl2, PMe3, and LiNHAr/Ar*. With both iron and ruthenium, the PMe3 starting materials underwent phosphine replacement with chelating ligands to give new group 8 imido complexes in the +2 oxidation state. Addition of 1,2-bis(diphenylphosphino)ethane (dppe) to M1/M1* gave Ru(NAr/Ar*)(PMe3)(dppe) and Fe(NAr/ Ar*)(PMe3)(dppe). Addition of 1,2-bis(dimethylphosphino)ethane (dmpe) provided Ru(NAr/Ar*)(dmpe)2. A triphos ligand, {P(Me)2CH2}3SitBu (tP3), was also examined. Addition of tP3 to Fe1 provided Fe(NAr)(tP3) (Fe4), but a similar reaction with Ru1 only gave intractable materials. Oxidation of Fe4 with AgSbF6 gave {Fe(NAr)(tP3)}+SbF6− (Fe4a). Oxidation of Ru2 with AgSbF6 gave the unstable cation {Ru(NAr)(PMe3)(dppe)}+, which dimerized in the presence of acetonitrile via C−C bond formation at the aryl group C4 positions, affording {Ru(NAr)(PMe3)(NCMe)(dppe)}2+. This suggested that there was substantial radical character in the imide π system on oxidation and that an aromatic group substituted at the 4-position might provide greater stability. The cations {Fe(NAr*)(PMe3)(dppe)}+ (Fe2a*), {Ru(NAr*)(PMe3)(dppe)}+ (Ru2a*), and Fe4a were examined by EPR spectroscopy, which suggested differences in electronic structure depending on the metal and ligand set. CASPT2 calculations on model systems for Ru2a* and Fe2a* suggested that the large differences in electronic structure are related to the energy gap between the π-antibonding HOMO and the π-bonding HOMO-1. Both the geometry of the phosphines, which is slightly different between the iron and ruthenium analogs, and the metal center seem to contribute to this energetic difference.

■

INTRODUCTION Metal−ligand multiple bonds are critical functional groups for chemists targeting many types of metal complex reactivity.1 Among the group 8 metals are some notable examples that demonstrate the versatility and high degree of reactivity that such complexes can wield. Perhaps best known is Grubbs’ alkene metathesis catalyst series, which prominently features a ruthenium−carbon double bond. Nitrogen is another element which frequently engages in metal−ligand multiple bonding, including with group 8 metals. For example, many ruthenium and iron terminal imido complexes have been reported in recent years.2−9 Group 8 imides are studied for their involvement in carbon−nitrogen bond forming reactions.10−15 Another common reason for studying M−N multiple bonds with group 8 metals stems from interest in nitrogen fixation, the conversion of molecular N 2 to NH3 and related chemistry.16−24 There are already two ubiquitous methods that perform this function, the Haber−Bosch process and biologically preeminent nitrogenase enzyme.16,25−34 In addition, M−N multiple bonds are sought for their ability to make © XXXX American Chemical Society

N−O and N−N bonds for synthetic purposes and for energy storage.24,35−37 Improving the understanding of discrete, homogeneous, unimolecular group 8 metal imido complexes is a fundamental step to better understanding the application-driven chemistry highlighted above. Along these lines, the study of M−N multiple bonds in related moieties, such as nitride, hydrazido, and imide functional groups, has been an active area for many years.1 A small selection of group 8 imide complexes is shown in Chart 1 as an illustration. These complexes highlight some of the commonly employed ligand structures that stabilize terminal Fe and Ru imidos to give isolable molecular complexes. An underlying feature in Chart 1 is the utilization of multidentate ligands, often with a high degree of conjugation. Note also the steric bulk on both the imido groups and the ancillary ligands, which favors low coordination Received: June 5, 2019

A

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

and its Fe analogs were probed by EPR, electrochemical, and NMR techniques and computational analysis. These studies illuminate drastic differences upon switching from a tris(chelating) phosphine to a monodentate/bidentate combination of phosphines, as well as fundamental differences between Fe and Ru imides.

Chart 1. Selected Examples of Group 8 Metal Imide Complexes15,38−43

■

RESULTS AND DISCUSSION Synthesis. The previously reported Ru(NAr)(PMe3)3 (Ru1) is shown in Chart 1 and is denoted as compound F (2,6-diisopropylphenylimide = NAr).42 The synthesis of this complex has been simplified from the previous report (Scheme 1). One can add excess PMe3 to {Ru(COD)Cl2}x60 to generate Scheme 1. Synthetic Route to Ru(NAr)(PMe3)3 (Ru1) and Ru(NAr*)(PMe3)3 (Ru1*)

numbers and terminal imidos (or nitrides) on monomeric metal centers, as opposed to bridging structures.8,9,44,45 Another feature highlighted by the molecules in Chart 1 is the wide variety of formal oxidation states, in combination with molecular charges. Many of these complexes demonstrate both interesting reactivity and electronic structure. For example, in Chart 1, the Fe complexes formally span from a low-spin Fe(II) anionic imido (A)38 to a triplet Fe(IV) cationic imido (B).39 This suggests a wide range of accessible electronic character at the metal center, given the right ligand framework. Very closely related to these iron-based terminal imidos are complexes with nickel, copper, and cobalt, the numbers of which have also expanded in recent years.46−49 Examples of terminal osmium and ruthenium imido complexes in low formal oxidation states are less common. While some imido complexes of ruthenium and osmium exist, these are often in high oxidation states with many strongly donating metal−ligand multiple bonds in a single molecule. Examples of such complexes with osmium include the series Os(O)4−n(NtBu)n (n = 1−4) or Schrock’s trans-Os( NAr)2(PMe3)2.50−52 Similar to these osmium compounds is the report of Ru(NAr) 2 (PMe 3 ) 2 by Wilkinson. 53,54 Ruthenium(VI) bis(imido) compounds containing porphyrin ligands are another class of ruthenium imides well represented in the literature, also in a high oxidation state.55−57 One of the few classes of terminal Ru imido complexes in low oxidations states is Steedman's imido-capped Ru(II) arenes (Chart 1).41,43,53,58 Ruthenium complexes with bridging imido ligands have also been reported.8 In this work, the synthesis of a series of related terminal Ru(II) and Fe(II) imidos is discussed. These systems are fourcoordinate and utilize three phosphines as the tripodal base structurea common motif demonstrated by complexes A and F shown in Chart 1. The properties of these complexes upon one-electron oxidation also are examined, including characterization of a rare example of a Ru(III) imide.59 This complex

cis-RuCl2(PMe3)4, instead of beginning with RuCl2(PPh3)3 as described previously. This provides a higher yield with a more readily isolated product. As shown in Scheme 2, addition of diphenylphosphinoethane (dppe) displaces 2 equiv of PMe3 from Ru1, generating Ru(NAr)(dppe)PMe3 (Ru2). The chelate-containing Ru2 demonstrates greater stability than Ru1 while retaining a similar coordination environment. In a similar reaction, 1 equiv of a smaller chelating ligand, dimethylphosphinoethane Scheme 2. Substitution of PMe3 in Ru(NAr)(PMe3)3 (Ru1) and Ru(NAr*)(PMe3)3 (Ru1*) with the Chelating Phosphine dppe To Form Ru(NAr)(dppe)(PMe3) (Ru2) and Ru(NAr*)(dppe)(PMe3) (Ru2*)a

a

Reaction of dmpe with Ru1/Ru1* gave Ru(NAr/Ar*)(dmpe)2 (Ru3/Ru3*). B

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry (dmpe), was added to Ru1, which resulted in leftover starting material (Ru1) and 5-coordinate, 18-electron Ru(NAr)(dmpe)2 (Ru3). Attempts to prepare a Ru(II) imido with triphos ligands, {P(Me) 2 CH 2 } 3 Si t Bu ( t P 3 ) and {P(Ph)2CH2}3CCH3, did not yield the desired product but rather a combination of decomposition and/or disproportionation products. This was evidenced by the multitude of new resonances that appear in 31P NMR spectra, without any clear or major products as judged by relative integrations. This chemistry was successfully extended to iron, but with notable differences in the relative stabilities of complexes bearing various phosphine ligands. Addition of 6 equiv of PMe3 to a suspension of FeCl2 in THF results in rapid formation of a presumed FeCl2(PMe3)4 intermediate, which can only be observed in situ as a transparent aquamarine solution in THF. The complex is paramagnetic, and the volatile PMe3 ligands appear to be quite labile, as removing the volatiles from the solution results in re-formation of FeCl2. These observations are consistent with those made by Karsch upon the addition of PMe3 to FeCl2(PMe3)2, in 1977.61 By 31P NMR spectroscopy, no resonance for the PMe3 added to the solution was observable, nor was there a new product peak, suggesting a paramagnetic product with bound and unbound phosphine in rapid exchange. A THF solution containing 2.1 equiv of LiNHAr was added dropwise to a solution of the paramagnetic, in situ generated FeCl2(PMe3)4. A new diamagnetic iron complex, formulated as Fe(NAr)(PMe3)3 (Fe1), is observed in situ. This species is comparable with Co(NAr)(PMe3)3 recently reported by Deng and co-workers.46 When exposed to reduced pressure, dark green solutions of Fe1 turn brown, and new complexes are observed, which are likely paramagnetic as judged by the broad resonances observed by NMR spectroscopy. In situ generated Fe1 shows distinct heteroatom resonances at characteristic shifts for the phosphines in the 31P NMR spectrum, at 38 ppm, and for the imido nitrogen in the 14N NMR spectrum, at 312 ppm. This is consistent with the Fe(NAr)(PMe3)3 (Fe1) structural assignment (Scheme 3) analogous to Ru1 (Scheme 2). Further support for the assignment of Fe1 comes from the isolation of complexes Fe2 and Fe4 upon addition of chelating phosphines. Addition of dppe to Fe1 provides Fe(NAr)(dppe)(PMe3) (Fe2), the iron analogue of ruthenium complex Ru2. Complex Fe2 is dark green in solution and is isolable as a crystalline solid. Complex Fe1 also reacts with 1 equiv of triphos {P(Me)2CH2}3SitBu (tP3) to generate Fe(NAr)(tP3) (Fe4), which contrasts with the attempted synthesis of the ruthenium analogue (vide supra). Triphos complex Fe4 is isolable as a dark purple crystalline solid. The low-spin Fe(II) complexes Fe2 and Fe4 are stable when they are stored as solids at reduced temperature. Note in Scheme 3 that, during the synthesis of Fe2, a paramagnetic byproduct also formed, which was isolable as a red powder. This paramagnetic species, Fe5, is presumed to be Fe(dppe)(NHAr)2, on the basis of the properties of the related, and more readily isolated, Fe(dppe)(NHAr*)2 (vide infra). Oxidation of Fe4 with 1 equiv of AgSbF6 provides the iron(III) complex {Fe(NAr)tP3}SbF6 (Fe4a) (eq 1). This reaction generates Ag0 and transforms the starting dark purple solution to a vivid blue solution, from which crystals of Fe4a can be obtained. Complex Fe4a has been characterized by

Scheme 3. Synthesis of Unstable Fe(NAr)(PMe3)3 (Fe1), an Analogue of Ru1, from in Situ Generated FeCl2(PMe3)4a

a

Fe1/Fe1* can be characterized spectroscopically and gives stable Fe(NAr)(dppe)(PMe3)3 (Fe2/Fe2*), an analogue of Ru2, on treatment with dppe. Fe1 provides Fe(NAr)(tP3) (Fe4) on addition of the triphos ligand {P(Me)2CH2}3SitBu (tP3).

single-crystal X-ray diffraction, and the electronic structure has been probed by EPR (vide infra). Attempts at one-electron oxidation of Ru(NAr)(dppe)(PMe3) (Ru2) and Fe(NAr)(dppe)(PMe3) (Fe2) gave quite different results from the simple conversion of Fe4 to Fe4a (eq 1). Oxidation of Ru2 with AgBArF24, BArF24¯ = tetrakis{3,5bis(trifluoromethyl)phenyl}borate, in DME/NCMe (Scheme 4) leads to isolation of dimeric Ru2b, where 2 equivalents of the oxidized complex couple through the para carbons of the imide aromatic groups with coordination of NCMe. Similar reaction pathways have been noted for nickel imide complexes by Stephan and co-workers47 and copper systems reported by Warren and co-workers.49 Betley and co-workers also reported a similar dimerization process for Fe(III) phenylimides.15 On the basis of EPR data of the Ar* derivative Ru2a* (vide infra), radical character is delocalized across the imido ligand, allowing for a radical coupling at the unhindered para carbon of the aromatic imide of 2 equiv of Ru2a, yielding Ru2b. We hypothesize that the radical dimerization observed in Ru2b proceeds through the 19-electron radical cation {Ru(NAr)(dppe)(PMe3)(NCMe)}+, resulting from solvent addition to unstable Ru2a, which on dimerization gives 18electron Ru2b. Running the reaction in the absence of acetonitrile, i.e., in DME, leads to isolation of a complex C

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry Scheme 4. Oxidations of Ru2/Ru2a* and Effect of Solventa

a

Silver(I) oxidation of Ru2 in DME gives {Ru(NAr)(PMe3)(dppe)}+ (Ru2a), which is observable spectroscopically but quickly decomposes. The same oxidation in the presence of acetonitrile led to dimerization through the para position of the aromatic group to give Ru2b. Oxidation of Ru2* gives the relatively stable Ru2a*.

substitution of the para position on the imido group with the iron complexes leads to greater stabilization and easier isolation.

tentatively assigned as {Ru(NAr)(dppe)(PMe3)}SbF6 (Ru2a), which rapidly decomposes, even when it is frozen in 2-MeTHF for examination by EPR spectroscopy. To better stabilize the Ru(III) imido cation, an imide derivative with para carbon substitution, Ru(NAr*)(PMe3)3 (Ru1*,) where Ar* = 2,4,6-triisopropylphenyl, was synthesized in 70% yield following the same procedure as for Ru1 (Scheme 1). The dppe derivative Ru(NAr*)(dppe)(PMe3) (Ru2*), analogous to Ru2, was then prepared in 41% yield (Scheme 2). The more highly substituted Ru2* can be oxidized (Scheme 4) by addition of 1 equiv of AgSbF6 in DME to yield pinkish brown {Ru(NAr*)(dppe)(PMe3)}SbF6 (Ru2a*). As evidenced by the reproducibility of the EPR and Evans method measurements (vide infra) on this compound, Ru2a* is greatly stabilized by the substitution at the para position on the imido in comparison to Ru2a. This, along with the dimerization to form Ru2b discussed previously, suggests significant radical character in the imide group, and the 2,4,6-substitution is likely stabilizing the arene radical character in Ru2a* (Figure 1).

The complexes Fe2*/Fe2a*, Fe4/Fe4a, and Ru2*/Ru2a* were then examined using cyclic voltammetry. The Ru2a* complex could not be successfully analyzed, as it appeared to react rapidly and irreversibly with the electrolyte, {NBu4}+ PF6−. However, the Ru2* complex shows an oxidation at −0.58 V (vs Fc/Fc+) assigned to the Ru(II/III) couple. This potential is a few hundred millivolts more oxidizing than the potentials noted for the Fe2*/Fe2a* and Fe4/Fe4a complexes, which were measured as −0.76 and −0.90 V, respectively. Generally, deeper analysis of the CV data for these complexes proved challenging due to the presumed chemical noninnocence of these complexes upon oxidation and reduction. (See the Supporting Information for more details.) Synthetically, the iron analogs suffer an additional complication not present in the ruthenium syntheses. As shown in Scheme 3, addition of dppe to Fe1 gave a mixture of products, one of which was the diamagnetic Fe(II) imide Fe2 and the other a paramagnetic species formulated as Fe(NHAr)2(PMe3)2 (Fe5). If Ar* was used instead of Ar, the paramagnetic byproduct, Fe(NHAr*)2(PMe3)2 (Fe5*), was more crystalline and was isolated from the crude reaction mixture. Single crystals of Fe5* were used for structural determination by X-ray diffraction. The complex Fe5* can be made more directly by addition of 2 equiv of LiNHAr* to FeCl2(dppe) (Figure 2), giving a product that matches the unit cell and magnetic moment of the side product from the previous route. It is electronically similar to the Fe(II) dppe bis(alkyl) complexes reported by Chirik and others.62,63 The characterization of Fe5* supports the assignment of Fe5 as Fe(NHAr)2(PMe3)2; however, a clean sample of Fe5 has not been forthcoming from the reaction of FeCl2(PMe3)4 and 2 equiv of LiNHAr followed by dppe. The complexes Fe5 and

Figure 1. Some possible ground-state resonance forms for cationic Ru2a* and Ru2a, where the imide moiety has been oxidized to an imide radical and the metal is best described as Ru(II) (R = H, i-Pr). The apparent aryl radical character can be stabilized by adding bulky R groups to the ortho and para positions.

Stability issues were also noted during attempts to isolate {Fe(NAr)dppe(PMe3)}SbF6 (Fe2a). Given the problems demonstrated by the Ru analogue (Ru2a), we employed the same strategy here and prepared the more substituted imido complex Fe(NAr*)(dppe)PMe3 (Fe2*) by the same method used for Fe2 (Scheme 3). Oxidation of Fe2* with AgSbF6 gave {Fe(NAr*)(dppe)PMe3}SbF6 (Fe2a*); the oxidation is shown in eq 2. As was noted for the ruthenium analogs, the D

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

course of 1 week when they are stored as solids at −35 °C under an inert atmosphere. Structural Properties. The solid-state X-ray crystallographic structures of complexes Ru(NAr)(PMe3)3 (Ru1), Ru(NAr)dppe(PMe3) (Ru2), Fe(NAr)dppe(PMe3) (Fe2), Fe(NAr)tP3 (Fe4), Ru(NAr*)(PMe3)3 (Ru1*), and Fe(NAr*)dppe(PMe3) (Fe2*) serve as a reference point to begin comparing the electronic structures of Fe and Ru imido complexes. Table 1 highlights the diagnostic bond lengths and angles of these complexes. As reported previously, the structure of Ru1 has a roughly linear imide with a N1−Ru1−P1 angle (113.7°) much smaller than the N1−Ru1−P2 and N1−Ru1−P3 angles (122.6 and 128.4°). These bond angles highlight a significant distortion from C3v, related to stabilization of the HOMO (Figure 3).

Figure 2. (top) Synthesis of Fe(NHAr) 2(dppe) (Fe5) and Fe(NHAr*) 2 (dppe) (Fe5*). (bottom) Structure of Fe(NHAr*)2(dppe) (Fe5*). The ellipsoids are shown at the 50% probability level, and hydrogens in calculated positions are not shown for clarity.

Fe5* are highly reminiscent of a three-coordinate iron bis(arylamido) phosphine complex prepared by Cummins and co-workers, Fe(NRArF)2(PEt3), where R = C(CD3)2CH3 and ArF = 2,5-F(Me)C6H3.64 Despite its poorer crystallinity, Fe5 was also independently synthesized, similar to Fe5*, by addition of 2 equiv of LiNHAr to 1 equiv of FeCl2dppe in THF at room temperature, albeit in low yield (16%). A crystal structure of Fe5 was obtained from material made by this method. Despite repeated attempts, elemental analysis matching the formulations of Fe5 and Fe5* as the red paramagnetic materials from either synthetic method was not obtained, possibly due to the high sensitivity of the compounds. For instance, it has been noted that even samples of X-ray-quality crystals of Fe5/Fe5* decompose over the

Figure 3. In a previous study, the HOMO of Ru1 was found to be dz2 with σ* character between the imide and ruthenium when the geometry was constrained to be pseudotetrahedral. The distortion relaxes the σ* character, moving the nitrogen lone pair nearer the node of dz2.42

Table 1. Relevant Bond Lengths (Å) and Angles (deg) in the Iron and Ruthenium Imido Complexesa compound bond (Å)/angle (deg)

Ru1

Ru1*

Ru2

Fe2

Fe2*

Fe4

Fe4a(1)b

Fe4a(2)b

M−N1 M−P1 M−P2 M−P3 N1−C1 M−N1−C1 N1−M−P1 N1−M−P2 N1−M−P3 P1−M−P2 P1−M−P3 P2−M−P3

1.811(2) 2.224(1) 2.239(1) 2.254(1) 1.372(4) 174.9(2) 113.69(8) 122.63(8) 128.40(8) 96.47(3) 95.17(3) 93.27(3)

1.817(4) 2.224(1) 2.240(2) 2.253(1) 1.364(6) 179.9(3) 119.3(1) 121.9(1) 124.5(1) 94.70(6) 95.36(5) 93.99(5)

1.808(6) 2.240(2) 2.240(2) 2.275(2) 1.388(9) 166.5(5) 106.1(2) 127.7(2) 136.4(2) 97.38(8) 99.20(8) 82.07(8)

1.657(4) 2.156(2) 2.155(2) 2.175(2) 1.381(5) 171.0(3) 107.2(1) 125.0(1) 130.3(1) 100.77(6) 102.35(6) 86.10(6)

1.653(9) 2.166(4) 2.157(3) 2.162(3) 1.387(13) 172.1(8) 111.7(3) 122.8(3) 132.2(3) 96.4(1) 102.0(1) 85.4(1)

1.667(3) 2.136(1) 2.144(1) 2.148(1) 1.372(5) 178.9(4) 121.2(1) 121.8(1) 124.4(2) 93.97(5) 93.88(5) 93.70(5)

1.643(4) 2.232(2) 2.229(2) 2.195(2) 1.386(6) 172.5(4) 105.6(2) 129.4(2) 130.8(2) 93.93(6) 91.95(6) 93.05(6)

1.653(4) 2.232(2) 2.235(2) 2.192(2) 1.382(6) 167.8(4) 101.5(2) 128.0(2) 133.5(2) 93.23(6) 93.16(6) 94.36(6)

a

See Figure 4 for the labeling schemes and connectivity. Compounds with the same number (e.g., Ru2 and Fe2) have the same ligand sets. The imidos bear a 2,6-(iPr)2C6H3 group when the compound label has no asterisk and 2,4,6-(iPr)3C6H2 if there is an asterisk in the label. bTwo unique Fe4a cations in the asymmetric unit. E

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Figure 4. Crystal structures of Ru1, Fe2, and Fe4 from X-ray diffraction showing the labeling schemes for the parameters in Table 1 and connectivity. Ellipsoids are shown at the 50% probability level. Hydrogens in calculated positions are omitted for clarity.

Figure 5. M(III) cationic complexes examined by EPR spectroscopy. Fe4a has been characterized by single-crystal X-ray crystallography.

angle; thus, the phosphine environments are more similar. Another indication of the reduction of the distortion on going from Ru1 to Ru1* is the reduced range of N1−M−P angles. The range in Ru1* is only 5°, whereas in Ru1 it is 15°. This could suggest a slight difference in the HOMO of the molecules (dz2 σ* interaction) related to substitution of the para position on the phenylimide; however, it could also reflect crystal-packing energies in the solid state. The same trend, but to a smaller extent, i.e., the angular differences are smaller, is noted on comparing the N1−Fe−P angles in Fe2 and Fe2*. Here again, both complexes have the same metal, and the same phosphine basal set (dppe/PMe3), with the only structural difference being in the para substitution on the imide aromatic group. Notably, in Fe4, all of the N1−Fe−P angles are roughly equivalent, and the complex is close to C3v symmetry, likely enforced by the structure of the triphos ligand, tP3. One possibility for the difference in structures between Fe4 and Fe2 is that the loss in flexibility in the phosphine basal set for Fe4 weakens the metal−imido bond. Along these lines, Fe4 shows a slightly longer Fe−N distance, but the difference is not statistically significant in the X-ray diffraction data. Interestingly, upon one-electron oxidation of Fe4 to give Fe4a, the structure is drastically distorted from C3v. The N1−M−P angles for the three phosphines span over 32°, with the smallest angle of 101.46° noted for any of the complexes characterized crystallographically in the class.

This orientation allows for a stabilization by situating the dz2 orbital from the imide nitrogen close to the node of the filled dz2 orbital on ruthenium, which lowers the energy of the filled antibonding orbital.42 In Figure 4 are representative ORTEP diagrams for the M(II) imido complexes to show the labeling schemes. On examination of the structural details given in Table 1, most of the related derivatives synthesized also share this general characteristic of distortion from C3v but with some variance in the severity of the distortion. As a general trend, on comparison of complexes Ru2, Fe2, and Fe2*, which all have the dppe/PMe3 basal set, the ruthenium derivative seems to demonstrate a slightly larger distortion from C3v than the Fe derivatives. Ru2 has N1−M−P angles spanning a range of 30.3°, while the two iron derivatives have ranges of 23.3° (Fe2) and 21.5° (Fe2*). On the basis of previous investigations, these results suggest that the distortion provides less stabilization of the HOMO in Fe2/Fe2* in comparison to Ru2. Intuitively, this observation agrees with trends in bond strength among congeners in the same group. For a more detailed explanation of these results, we turned to computational investigations, which are discussed later. Another feature highlighted by comparison of the crystal structures is the less drastic distortion of the N1−Ru−P bond angles in complex Ru1*. When Ru1 and Ru1* are compared directly, where both derivatives have the tris-PMe3 ligand set, the smallest angle in Ru1 is 7° less than the average N1−Ru−P angle. In Ru1*, this difference is only 3.5° less than the average F

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry EPR Spectroscopy on the M(III) Imido Complexes. EPR spectroscopy is an incredibly useful experimental technique that provides details about the electronic structure of a paramagnetic complex by examining the behavior of the unpaired electron(s). We thought this could provide valuable insight in characterizing the radical cations Fe2a*, Fe4a, and Ru2a*, shown in Figure 5, all of which are S = 1/2 systems. In attempts to synthesize and study these compounds, delocalization of the radical was noted in the Ru2a system (Scheme 4). The Fe complexes were also unstable, but similar radical coupling products were not observed; this perhaps suggests some substantial differences in radical character. Going into these experiments, we suspected that, in the Ru systems, the radical was substantially delocalized across the imide ligand’s π system, while the Fe systems were likely simple Fe(III) complexes with metal-localized paramagnetic centers. To test this hypothesis, oxidized complexes were examined by X-band cw-EPR spectroscopy at several temperatures and with several preparations of each compound. These studies proved challenging due to the instability of the species of interest. Even with the addition of a p-isopropyl substituent on the imido group’s phenyl ring for improved stability (Fe2a* and Ru2a*), these complexes decompose within a few days of storage at −35 °C as solids. This, coupled with their lack of crystallinity, precluded effective purification. Consequently, some variation was noted from spectrum to spectrum among the different samples prepared for Fe2a* and Ru2a*. Even small differences, such as the purity of the silver salt used in the oxidation or the particular batch of PMe3 used to generate the starting material, appear to affect the character of Fe2a* and Ru2a* after oxidation. In contrast, Fe4a, which is crystalline and could be reproducibly purified following the oxidation of Fe4, provided consistent EPR spectra. The EPR spectrum of Fe2a* is shown in Figure 6a (black trace). The red trace in this figure is a spectral simulation assuming that the spectrum arises from a composite of a 98% contribution from a low-spin Fe(III) paramagnetic center characterized by principal g values of 2.49, 2.10, and 1.96 and a 2% contribution from a ligand radical species centered at g = 1.985 with a peak to peak line width of 5.7 mT. The small contribution from a ligand radical was included in the simulation to better account for the broad line shape of the feature at high field and some of the sharper features in the region about 340 mT. Its inclusion in the simulation of Figure 6a improved the normalized RMSD by over 40%. The EPR spectrum highlighted in the inset of Figure 6 (black trace) came from an oxidation of Fe2* to Fe2a*, in which the reaction solution quickly went from the dark green color characteristic of Fe2* and Fe2a* to a dark brownish gray upon standing. This spectrum features three resolved peaks centered about 339.3 mT that can be simulated (red trace) using a g value of 1.982 and a single 14N-hyperfine coupling characterized by an isotropic coupling (Aiso) of 13.3 MHz and a dipolar coupling (Adip) of 80.4 MHz. The 14N hyperfine coupling is commensurate for what might be expected for an imide radical with the unpaired spin mostly confined to the ligand’s π molecular orbitals.65 The g value is lower than that expected for an isolated organic radical, indicating that the imide nitrogen is coupled to the metal, which strengthens our assignment of this spectrum to a ligand radical.66,67 A similar reduction in g value was reported for a ligand radical obtained after oxidation of a ferrous bis(pyridine-2,6-diimine) complex by Wieghardt and co-workers.68

Figure 6. EPR spectra (black) of two different preparations of Fe2a* (a, b). The inset shows the g = 2 region of a spectrum consisting of both high- and low-spin Fe(III) centered paramagnets and a ligand radical that forms upon oxidation of Fe2*. The spectrum shown in (a) is relatively pure low-spin Fe(III), while (b) was obtained on a sample from a different preparation of Fe2a*, which gives a composite spectrum of the low-spin Fe(III) species expected and a ligand radical. Red traces represent simulated spectra using spin Hamiltonian parameters and speciations given in the text. Conditions used for the EPR measurements were as follows: (a) microwave frequency 9.403 GHz, microwave power 3.2 μW, field modulation 0.8 mT at 100 kHz, sample temperature 5 K; (b) microwave frequency 9.389 GHz, microwave power 3.2 μW, field modulation 0.6 mT at 100 kHz, sample temperature 30 K. For the inset the conditions were microwave frequency 9.408 GHz, microwave power 3.2 μW, field modulation 0.5 mT at 100 kHz, and sample temperature 30 K.

Careful examination of the inset to Figure 6 shows that this imide radical spectrum is superimposed on a broad feature that originates from a high-spin Fe(III) species. This high-spin center shows resonances at g = 9.5, 6.3, 4.3, and 1.8 (see Figure S1 in the Supporting Information) and likely originates from a mixture of oxidation products. As noted in Figure 2, the synthesis of Fe2* is accompanied by the formation of at least one other iron species, coproduced in our one-pot−three-step synthesis. Additionally, decomposition of Fe2*/Fe2a* may contribute to contamination. However, the observation of the EPR signals from this decomposed sample provided a glimpse of what impurities in the samples would likely show. The EPR spectrum shown in Figure 6b is from a preparation where the low-spin Fe2a* species is detected along with G

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry contributions from imide radical and high-spin Fe(III) centers. These data were simulated (red trace) using a model that consisted of an 82% contribution from the low-spin Fe2a* species and an 18% contribution from the imide radical. The Fe2a* samples that gave rise to the spectra in Figure 6 were studied at multiple temperatures from 5 to 80 K, and while some broadening was observed for the peaks attributed to the low-spin Fe(III) species as the temperature was raised, the change in speciation was modest (Figure S2). Specifically, for the Fe2a* sample of Figure 6a, the radical contribution went from 1.8% at 5 K to 2.7% at 80 K, and for the Fe2a* sample of Figure 6b, the imide radical contribution went from 18.8% at 30 K to 21.8% at 75 K. For both samples of Fe2a* whose EPR spectra are shown in Figure 6, spectral simulations provided similar g values which support the assignment of Fe2a* as a low-spin Fe(III)centered paramagnetic species characterized by g values of 2.48, 2.10, and 1.96. The rhombic symmetry of the g matrix is commensurate with the distortion from C3v symmetry found for Fe2* by X-ray crystallography. The major difference between the two spectra is the roughly 18% contribution of the three-line imide ligand radical species superimposed on the Fe2a* spectrum of Figure 6b. Because this imide radical shows microwave power saturation properties different from that of the low-spin Fe(III) species (Figure S3) and it was always present along with a substantial amount of high-spin Fe signal, we attribute it to a degradation product or a species that is on the degradation pathway of Fe2*. These studies of the Fe(III) terminal imide species Fe2a* suggest that the primary oxidation product of the complex is a metal-centered paramagnetic center. In contrast to Fe2a*, Fe4a consistently showed a composite EPR spectrum with both metal- and ligand-based radical signals. Figure 7 shows the EPR spectrum of Fe4a in 2MeTHF (black trace). The red trace is a simulation that models the spectrum as a composite of a 71% contribution from a low-spin Fe(III) center, characterized by a nearly axial g

tensor with principal values of 2.36, 2.00, and 1.99, and a 29% contribution from a ligand radical species. The principal g values that characterize the low-spin Fe(III) center have axial symmetry, reflecting the ligand geometry enforced by the triphos ligand, tP3. This simulation made use of the information regarding the nature of the imide ligand radical gained from our analysis of the EPR spectra obtained for Fe2a* and Ru2a*. Specifically, the model used in our simulations considered the ligand radical to have a line shape governed by inhomogeneous broadening due to a largely anisotropic 14N hyperfine interaction. For the simulation shown in Figure 7, the 14N hyperfine coupling was described with Aiso = 8.4 MHz and Adip = 76.5 MHz, similar to the values found for the imide ligand radical of Fe2a*, together with an isotropic g value of 2.014. For the EPR spectrum of Fe4a, the 14 N hyperfine coupling is not resolved, presumably due to larger intrinsic line widths for the ligand radical. In contrast to the decrease in g value that was measured for the imide radical of Fe2a*, the g value for the ligand radical of Fe4a is increased to 2.014 from the free electron value of 2.0023. The sign of this shift in g value depends on the nature of the excited states associated with paramagnetic ground state and is a consequence of the difference in electronic structure of Fe4a vs Fe2a*.67 The EPR spectra of Ru2a* shown in Figure 8 are also a composite of two distinct paramagnetic species and were chosen because they represent extremes in speciation. The spectra are centered at g = 2.00 and are from separate preparations of the oxidized complex. The spectrum shown in Figure 8b features a three-line pattern, similar to that attributed to a ligand-based or imide radical in our analysis of the Fe2a* complex, superimposed on a broader background signal. The red trace of Figure 8b is a simulation where the imide radical is modeled with an isotropic g value of 2.0042 and a 14N hyperfine coupling with Aiso = 15.5 MHz and Adip = 56.6 MHz, which accounts for 70% of the EPR signal amplitude. The remaining 30% of the EPR amplitude stems from an S = 1/2 paramagnetic center with g values of 2.05, 2.04, and 1.96. The 14 N hyperfine coupling used here is in line with what was used to simulate ligand radical contributions for the Fe2a* and Fe4a EPR spectra. The g values used to model the background are not fit well in this simulation because contributions are only clearly resolved at the high- and low-field edges of the spectrum, but they are commensurate with values reported for octahedral Ru(III) complexes isolated in zeolites.69 The spectrum shown in Figure 8a represents the opposite extreme with 90% of the amplitude coming from a Ru(III)-centered S = 1/2 paramagnetic center with g values of 2.06, 2.01, and 1.97. The remaining 10% is assigned to an imide radical characterized by g = 2.007 and an 14N hyperfine coupling of Aiso = 4.0 MHz and Adip = 80.6 MHz. Microwave power saturation studies showed line shape changes commensurate with slower relaxation for the imide radical species and support our interpretation of the spectra as arising from discrete paramagnetic species (Figure S4). For both spectra, the ligandbased radical’s g value is closer to the free electron g value than those resolved for Fe4a and Fe2a*, demonstrating a difference in coupling to the metal ion’s orbital angular momentum. The results of the EPR analyses can be summarized as offering additional experimental support for a large amount of ligand radical character in nominally Ru(III) Ru2a*. With the same ligand set on iron, PMe3 and dppe, the unpaired electron spin is observed to be almost entirely localized on the metal

Figure 7. EPR spectrum of Fe4a (black) in 2-MeTHF and spectral simulation (red). For the simulation, 71% of the EPR signal amplitude was attributed to an Fe(III)-centered paramagnetic center with principal g values of 2.36, 2.00, and 1.99. The remaining 29% originates from a ligand-centered radical with g = 2.014. The spectrum was recorded at 10 K with microwave frequency 9.411 GHz, microwave power 3.2 μW, and modulation amplitude 0.6 mT at 100 kHz. H

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

coordination sphere in some fashion, with differing angles around the metal or perhaps differences in solvent coordination. While some impurities are observed, these did not seem to increase with increasing temperature for these samples, suggesting that the different metal-based and imide-based signals were not simply due to decomposition. As a result, the ratio between the signals is perhaps significant of the relative tendency of the metals to localize the radical in different environments. Computational Analysis. The crystallographic and EPR data above clearly indicate that the central metal has a significant effect on both the geometric and electronic structure of these compounds. Here we apply Møller−Plesset second-order perturbation theory (MP2) and multistate complete active space second-order perturbation theory (CASPT2)70,71 to investigate the observed differences between the iron and ruthenium complexes. For computational convenience, we study truncated versions of Fe(NAr)(PMe3)3 (Fe1) and Ru(NAr)(PMe3)3 (Ru1) with the model compounds Ph−NFe(PH3)3 (Fe6) and Ph−NRu(PH3)3 (Ru6). CASPT2 is an inherently many-body theory, making interpretation of its results in terms of intuitive concepts such as orbitals and orbital energies somewhat ambiguous. Here, we define the HOMO of the neutrally charged species as the singly occupied natural orbital (SONO) of the ground state of the cation of the same species at the same geometry. The negative of the computed vertical ionization potential can be interpreted as the HOMO energy (Koopmans’ theorem). Similar analysis of the excited states of the cation yields orbitals and orbital energies for lower occupied orbitals (HOMO-1 corresponds to the SONO of the first excited state of the cation, and so on). Although one can imagine ambiguities arising in this analysis, in the present work all of the discussed orbitals and orbital energies were completely unambiguous. Computational details are presented in the Experimental Section. Optimized structures are available in the Supporting Information. It was noted above that the experimentally determined structure of Ru(NAr)dppe(PMe3) (Ru2) deviates farther from C3v symmetry than that of Fe2, as indicated by a difference in the range spanned by the three N1−M−P angles (30.3° for Ru2 vs 23.3° for Fe2). A similar trend is noted in the MP2optimized structures of Ru6 and Fe6, which have N1−M−P angles spanning ranges of 9.5 and 3.4°, respectively (Table 2). The smaller ranges in these computed compounds relative to those of Ru2 and Fe2 are attributed to the difference in the ligands. The three phosphine ligands of Ru2 and Fe2 are not identical, and this asymmetric ligand environment likely encourages further deviation from C3v symmetry. The angle ranges computed for Ru6 and Fe6 are more comparable to the

Figure 8. EPR spectra of samples of Ru2a* (black) and their simulated spectra (red). The samples are treated as a composite of two distinct paramagnetic centersone centered on Ru(III) and one centered on the imide ligand. The two spectra show dramatically different proportions of the two different paramagnetic centers. Spectrum (a) is about 90% Ru(III) and 10% ligand radical, whereas spectrum (b) shows about 30% Ru(III) character and 70% ligand radical. These spectra were taken from different batches of Ru2a*, prepared following the same synthetic procedures. The measurement conditions for spectrum (a) were microwave frequency 9.408 GHz, microwave power 12.6 μW, field modulation 0.5 mT at 100 kHz, and sample temperature 30 K. Measurement conditions for spectrum (b) were microwave frequency 9.673 GHz, microwave power 0.2 μW, field modulation 0.5 mT at 10 kHz, and sample temperature = 30 K. The top spectrum was shifted 9.5 mT upfield to align it with the bottom spectrum for display purposes. The red traces in the figure are spectral simulations done with EasySpin, and the Hamiltonian parameters are given in the text.

center. Consequently, the change in metal center, while keeping the ligands constant, led to a dramatic change in observed electronic structure. In most cases, the ratio of metal to ligand radical character was very sensitive to conditions. Alternatively, with the more rigid and roughly C3v symmetric triphos ligand Me3SiC(CH2PMe2)3 (tP3), more ligandcentered radical character was observed for the Fe(III) system. This suggests that changing the ligand set can induce similar behavior between iron and ruthenium metal centers toward their ligands, restoring their electronic structure similarities between the two metals. This more rigid basal framework gave samples that were more reproducible and less sensitive to conditions. In almost all of the samples of these complexes examined, species with electron localization on the metal and species with localization of the radical on the imide ligand are observed. This varied slightly with the preparation, and we assume that the different signals are due to complexes that are different in

Table 2. N1−M−P Angles (deg) As Optimized at the MP2 Level of Theorya compound

a

I

angle (deg)

Ru6

Fe6

N1−M−P1 N1−M−P2 N1−M−P3 range

122.1 122.3 131.6 9.5

121.2 121.6 124.6 3.4

The range spanned by each set of angles is also given. DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

and again the HOMO-1 has significant ligand character. Consequently, in the ruthenium case, small changes in environment or structure may result in an energetic reordering of the HOMO and HOMO-1, thus dramatically changing the character of the radical. It is also worth mentioning that the energy gap of 0.04 eV between the HOMO and HOMO-1 energies is well within the errors of the various approximations employed in this calculation (CASPT2, neglect of solvent environment, ligand truncations, etc.). To investigate the influence of the ligand structure surrounding the central metal on the HOMO and HOMO-1 energies, we created a model, Ru6_mod, by taking the optimized structure of Fe6, replacing the central iron atom with ruthenium, and rigidly stretching the metal−ligand bonds to match the optimized Ru6 bond lengths. Ligand−metal− ligand bond angles and ligand internal coordinates remain frozen at their Fe6-optimized values. Consistent with our previous study, the more C3v-like structure of Ru6_mod relative to Ru6 results in greater antibonding character in the HOMO, thus destabilizing it by 0.29 eV. The bonding lobe of the HOMO remains in Ru6_mod but is diminished in comparison to Ru6. Smaller changes are observed in the energy of the M−N π-bonding HOMO-1, which is stabilized by 0.16 eV relative to optimized Ru6. In total, modifying the ligand structure of Ru6 to more closely match that of Fe6 results in a 0.44 eV increase in the gap between HOMO and HOMO-1, from 0.04 eV in Ru6 to 0.48 eV in Ru6_mod. Thus, both the identity of the central metal and the phosphine ligand structure play important roles in determining the character of the unpaired electron in the cationic species.

ranges seen in the X-ray structures of Ru1 and Fe4 (14.7 and 3.2°, respectively), which each have three identical phosphine ligands. We have previously hypothesized that the deviation of the structures of these species from C3v symmetry can be rationalized by consideration of the HOMO, which is an antibonding combination of the dz2 orbital of the metal atom and the σ-bonding orbital on the imido nitrogen.42 In the more symmetric configuration of Fe6, the HOMO (Figure 9, top left

■

CONCLUSIONS The goal of the current study was to examine structural and electronic differences between ruthenium and iron imido complexes with identical ligand sets. To this end, several new group 8 imides were prepared with phosphine coligands. The previously reported Ru(NAr)(PMe3)3 (Ru1), Ar = 2,6(iPr)2C6H3, was prepared using a simplified procedure and was the starting material for the other ruthenium imido complexes synthesized. It was found that the analogous Fe(NAr)(PMe3)3 (Fe1) can be generated from in situ produced “FeCl2(PMe3)4” and 2 equiv of LiNHAr. The iron imide was spectroscopically characterized, but phosphine lability and volatility precluded isolation. Nevertheless, Fe1 was a very good starting material for several different iron imido complexes through phosphine replacement with chelates. Reaction of M(NAr)(PMe3)3 with dppe gave M(NAr)(dppe)(PMe3) (Ru2 and Fe2), which were both thermally stable and structurally characterized. Oxidation of Fe2 and Ru2 gave the cations {Fe(NAr)(dppe)(PMe3)} + (Fe2a) and {Ru(NAr)(dppe)(PMe3)}+ (Ru2a). Neither of these complexes were highly stable, and it was determined that one cause of the instability of Ru2a was significant delocalization of unpaired spin onto the imido ligand; this favored dimerization of the complex by coupling through the 4-position of the arylimido ligand in acetonitrile to give Ru2b (Scheme 4). Consequently, we prepared {Fe(NAr*)(dppe)(PMe3)}+ (Fe2a*) and {Ru(NAr*)(dppe)(PMe3)}+ (Ru2a*), where Ar* = 2,4,6-(iPr)3C6H2; this substitution of the 4-position of the aromatic ring greatly increased the stability of the oxidized species, allowing isolation and characterization.

Figure 9. Orbital energies (defined as the negative ionization potential, as described in the text) of Fe6, Ru6, and Ru6_mod computed at the CASPT2 level of theory. Insets show the HOMO and HOMO-1 orbitals (SONOs of the cations, as described in the text). A green arrow indicates the bonding lobe of the HOMO of Ru6.

inset) is purely antibonding. However, as the imido ligand moves farther from the C3 axis, as in Ru6, the nitrogen atom sits closer to the node of the metal dz2 orbital, resulting in a HOMO of mixed bonding−antibonding character (Figure 9, top center inset). The partial bonding character is clearly evident in the blue lobe on the right-hand side of this orbital, marked by a green arrow. The fact that Ru6 is more strongly distorted than Fe6 can be attributed to the stronger bonding/ antibonding interactions associated with second-row transition metals in comparison to first-row transition metals in the same group. Finally, we investigated the balance between M(III) and ligand-centered radical character in Ru6 and Fe6. Experimental EPR results above indicate that Fe2a* is predominantly M(III) in character, while a mixture of species of M(III) and ligand radical character is observed for Ru2a*. CASPT2 calculations also suggest that Ru6 would more likely exist in a ligand radical state than Fe6 (Figure 9) and provide a physical explanation for this trend. In Fe6, the largely metal centered (dz2) HOMO is found to be 0.85 eV above the HOMO-1 orbital, which has significant population on the ligand. As a result of this relatively large energy gap, an iron-centered radical should dominate the population under most conditions. In contrast, the HOMO and HOMO-1 of Ru6 are split by only 0.04 eV, J

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Examination of {Ru(NAr*)(dppe)(PMe3)}+ (Ru2a*) by EPR shows that the arylimide ligand based radical character can vary from about 10% to 70% depending on preparation. The data shown in Figure 8 provide spectra obtained at the extremes of speciation for the Ru(III)-based (Figure 8a) and ligand-based (Figure 8b) paramagnets. These results contrasted greatly with the almost exclusively iron-based radical in isostructural Fe2a* (Figure 6a). This large dependence on preparation suggests that the metal-based and ligand-based radical configurations, again perhaps different by angles in the frozen glass or solvent coordination, are quite close in energy. Calculations on the model compounds Fe6 and Ru6 show the HOMO of each complex is a largely σ* orbital between the metal dz2 and the imide nitrogen. In the iron case, the orbital appears to be exclusively antibonding, but some bending in the ruthenium case leads to some slight bonding character. The HOMO-1 of Fe6 and Ru6 are M−N π-bonding orbitals. In the iron case, there is a relatively large energy gap between the HOMO and HOMO-1. For ruthenium, these orbitals are quite close in energy. The higher energy HOMO-1 in Ru6 seems to be due to two contributing factors: (1) small differences in the phosphine angles lead to a larger interaction with the HOMO1 raising its energy and (2) the larger overlap of ruthenium relative to iron leads to a larger destabilization in the antibonding interactions. This was tested by placing ruthenium in the same ligand environment as iron in the calculations (Ru6_mod), which causes the ruthenium complex’s HOMO-1 to drop in energy; however, the decrease in energy due to changing the ligand geometries around the metal was not quite enough to match the HOMO-1 energy of Fe6. Consequently, it appears that the geometry and metal identity both contribute to the energy difference between the HOMO and HOMO-1. The difference in energy between these two orbitals appears to be crucial to understanding the location of the unpaired electron spin density in Ru2a* and Fe2a*. Upon oxidation of Fe2*/Ru2*, an electron is removed from the HOMO. Mixing to place unpaired electron density into the imido π system could be occurring through mixing of the HOMO with HOMO-1. As was pointed out, Fe6 (similar to Fe2a*) has a large energy gap, and one expects that that complex will behave largely as iron(III) with virtually all of the radical character on the metal. In the ruthenium analogue, there is a small HOMO/ HOMO-1 energy gap, allowing the unpaired electron to access the imido π system. While the calculations are consistent with electronic mixing between the metal and imido ligand orbitals assuming the determining role in the electronic structure, EPR spectroscopy suggests that, experimentally, the radical tended to localize either on the metal or in the imide π system. This localization is perhaps due to geometry in the frozen glass or solvent effects. The ratio of metal and ligand radical obtained varied somewhat with preparation but was consistent with expectations from the calculations. From these results, it seems that fairly small alteration of the basal ligand set can have a large effect on the electronic structure of {Fe(NAr)(PR3)3}+ complexes. Changing from (PR3)3 to tP3 dramatically increased the unpaired electron density on the imido ligand. Similarly, changing the metal from iron to ruthenium in {M(NAr*)(dppe)(PMe3)}+ dramatically increased the radical character in the imido π system. Catalysis based on such imide systems should be highly sensitive to metal environment and identity, making the systems highly tunable. Further, the new complexes Fe(NAr)(PMe3)3 and

In additional synthetic experiments, the triphos ligand {P(Me)2CH2}3SitBu (tP3) was added to solutions of Fe1 and Ru1. In the iron case, clean replacement was observed to generate Fe(NAr)(tP3) (Fe4). For the ruthenium reaction, no product could be isolated from the complex reaction mixture. The iron complex was oxidized to isolable {Fe(NAr)(tP3)}+ (Fe4a). The dimerization of Ru2a through the aromatic group of the imido certainly suggests that some radical character is residing in the imido ligand rather than the complex being strictly ruthenium(III) d5. This issue was explored by EPR spectroscopy using the more stable Ru2a* and the iron analogue Fe2a*. Further, the iron imido complex bearing the triphos ligand, Fe4a was explored by EPR spectroscopy as well. It was found that {Fe(NAr*)(dppe)(PMe3)}+ (Fe2a*) is best described as a low-spin iron(III) paramagnetic center characterized by g values of 2.48, 2.10, and 1.96. In contrast, going to the less flexible ligand set in {Fe(NAr)(tP3)}+ (Fe4a) gave rise to a mixture of paramagnetic centers: a metal-based center showing an axial g tensor with g|| = 2.35 and g⊥ = 1.98 that accounted for 71% of the EPR spectral amplitude and a narrow signal with unresolved hyperfine coupling that was attributed to a ligand radical accounting for the remaining 29% of the spectral amplitude. The different species observed perhaps represent different coordination geometries of the Fe4a complex, possibly with different solvation; however, they illustrate that Fe4a is more likely than Fe2a* to exhibit ligand radical character. The speciation of paramagnetic centers is sensitive to the ancillary ligand set, which on the basis of the CASPT2 calculations (vide infra) is perhaps a consequence of the flexibility of the phosphine ligand environment and subsequent effects on the closeness of the tP3 basal set to C3 symmetry. In the calculated structures, the PR3 basal sets have maximum flexibility, each being entirely monodentate for both Ph−NFe(PH3)3 (Fe6) and Ph−NRu(PH3)3 (Ru6). The more flexible basal set of Fe2a*, which can move PMe3 separately from the dppe fragment, mimics the model complex Fe6, suggesting that the HOMO of Fe2a* is best described by the dz2−N(2p) σ* interaction and that there is little mixing in the system. This agrees with the minimal distortion of the N− Fe−P bond angles found for the optimized structure of Fe6, which remains closer to a C3v-symmetric core around iron. In the Fe4a system, however, the experimental behavior much more closely resembles that of the calculated Ru6 model. The basal ligand in Fe4/Fe4a is a trischelate, which imposes a much more rigid framework on the orientation of the phosphine-based orbitals in the complex. This difference in the ancillary ligands appears to be the primary cause of the difference in radical location between Fe2a* and Fe4a. It seems possible that this difference leads to an increase in the energy of the HOMO-1 due to ligand field destabilization, which also decreases the energy difference between the HOMO and HOMO-1 to the extent that the radical is delocalized over the metal and ligand-based orbitals, respectively, and facilitates mixing of the two orbitals. The broken symmetry and radical behavior, which mimic the Ru systems, are supported by the severe distortion of the P−M−N angles observed in the solid-state structure of Fe4a (Table 1 and Figure 3). Regardless of the electronic cause and effect, it seems that subtle manipulations of the ancillary ligands can render a large difference in the electronics of the metal−imide bond. K

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

Synthesis of RuNAr(PMe3)3 (Ru1). Ru1 was prepared as previously reported,42 using cis-RuCl2(PMe3)4 as prepared above. Elemental analysis, 31P/1H/13C NMR spectroscopy, and the structure from X-ray diffraction were published previously.42 UV−vis absorption (THF, 21 °C): 465 nm (3039 cm−1 M−1), 290 (5532 cm−1 M−1), 232 nm (14065 cm−1 M−1). 14N NMR (benzene-d6, 36 MHz, 25 °C): 326.3 ppm. Synthesis of Ru(PMe3)(dppe)(NAr) (Ru2). To a stirred solution of Ru1 (134 mg, 0.398 mmol, 1 equiv) in THF was added a solution of dppe (164 mg, 0.402 mmol, 1 equiv) in 2 mL of THF. The reaction mixture was stirred for 1 h at room temperature. The volatiles were removed in vacuo, and the dark red solid residue was rinsed with cold hexane (3 × 1 mL). The volatiles were once again removed in vacuo, and the solids were dissolved in a minimal amount of toluene. The concentrated toluene solution was stored at −35 °C overnight to yield flaky, red-orange crystals of Ru2 (202 mg, 87%). Mp: 110 °C dec. 1H NMR (500 MHz, benzene-d6): δ 8.08 (t, J = 9.9 Hz, 4H), 7.58−7.51 (m, 4H), 7.13 (s, 3H), 7.12−6.81 (m, 12H), 4.90 (septet, J = 7.0 Hz, 2H), 2.02−1.89 (m, 2H), 1.85−1.71 (m, 2H), 1.41 (d, J = 6.9 Hz, 12H), 0.60 (d, J = 9.8 Hz, 9H). 13C{1H} NMR (126 MHz, benzened6): δ 133.90 (t), 130.94 (t), 128.97(s), 128.14(s), 122.73(s), 122.31 (d), 120.40 (d), 118.51(s), 30.23 (t), 27.77(s), 26.67(s), 23.65(s), 22.20 (d), 21.97 (s). 31P NMR (202 MHz, benzene-d6): δ 99.31 (d, J = 21.3 Hz), 23.65 (t, J = 21.2 Hz). 14N NMR (36 MHz, THF): δ 337.6 (s). UV−vis absorption (THF, 21 °C): 472 nm (3610 cm−1 M−1), 314 (6576 cm−1 M−1), 272 nm (16709 cm−1 M−1). Anal. Calcd for C41H50NP3Ru: C, 65.59; H, 6.71; N, 1.87. Found: C, 65.07; H, 6.80; N, 1.73. Synthesis of {Ru(dppe)(PMe 3 )(NAr)(NCCH 3 )} 2 {BAr F 24 } 2 (Ru2b). A solution of Ru2 was prepared (60 mg, 0.080 mmol) in 6 mL of a 1/1 (volume/volume) mixture of MeCN and DME. This redorange solution was stirred at room temperature. A separate solution of Ag(BArF24) (30 mg, 0.087 mmol) in 2 mL of MeCN/DME was added dropwise to the solution of Ru2. After addition, the reaction mixture was stirred for 24 h at room temperature, over which time the reaction solution turned bright purple. The reaction mixture was filtered using Celite as a filtering agent to remove Ag0, and the filtrate was concentrated in vacuo. The concentrated filtrate was layered with hexane and stored at −35 °C for 3 days to give dark purple crystals of Ru2b (62.1 mg, 75.6%). Mp: 140 °C dec. 1H NMR (500 MHz, CD2Cl2): δ 7.92 (s, 8H), 7.64 (t, J = 8.2 Hz, 8H), 7.38 (t, J = 7.4 Hz, 8H), 7.33 (dd, J = 8.2, 5.2 Hz, 12H), 7.27−7.20 (m, 12H), 6.93 (d, J = 7.4 Hz, 4H), 6.91−6.86 (m, 2H), 4.55 (septet, J = 6.8 Hz, 4H), 2.42−2.25 (m, 4H), 2.09 (m, 2H), 2.07−1.95 (m, 4H), 1.18 (d, J = 6.9 Hz, 24H), 0.67 (d, J = 9.9 Hz, 18H). 19F NMR (470 MHz, CD2Cl2): δ − 62.88 (s). 31P NMR (202 MHz, CD2Cl2): δ 99.24 (d, J = 22.5 Hz), 24.43 (t, J = 22.5 Hz). (Note: 13C and 14N could not be obtained due to compound instability in solvents in which it was soluble enough to see an NMR signal.) Synthesis of Ru(NAr)(dmpe)2 (Ru3). A 20 mL scintillation vial was charged with Ru1 (106 mg, 0.2 mmol, 1 equiv), 5 mL of THF, and a magnetic stir bar. To this red-orange solution was added a solution of dmpe (65 mg, 0.4 mmol, 2 equiv) in 2 mL of THF, dropwise, at room temperature. The solution was stirred for 4 h, over which time the solution became brownish yellow. The volatiles were removed in vacuo, and the residue was rinsed with several small aliquots of n-hexane. The solids were then dissolved in a minimum amount of THF and layered with n-hexane. The layered solution was stored at −35 °C overnight to yield platelike green-brown crystals of Ru3 (52 mg, 45%). 1H NMR (500 MHz, benzene-d6): δ 7.27 (d, J = 7.3 Hz, 2H), 6.51 (t, J = 7.3 Hz, 1H), 2.87 (septet, J = 7.0 Hz, 2H), 1.51−1.38 (m, 4H), 1.34 (dd, J = 10.6, 6.9 Hz, 12H), 1.29−1.19 (m, 4H), 1.17 (d, J = 8.2 Hz, 6H), 1.10 (t, J = 3.2 Hz, 6H), 0.99 (t, J = 2.2 Hz, 6H), 0.82 (d, J = 6.3 Hz, 6H). 13C{1H} NMR (126 MHz, THFd8) δ 119.84, 109.28, 108.19, 104.28, 33.33 (m), 30.44, 30.10 (m), 27.78, 25.73, 24.81, 23.73−22.89 (m), 22.39 (d, J = 12.2 Hz), 21.06− 20.21 (m), 18.97, 13.92. 31P NMR (202 MHz, benzene-d6): δ 42.88 (t, J = 14.7 Hz), 30.76 (t, J = 14.8 Hz). 14N NMR (36 MHz, THF): δ 577.0.

Ru(NAr)(PMe3)3 can likely act as starting materials to a host of group 8 imido complexes.

■

EXPERIMENTAL SECTION

General Considerations. All manipulations were carried out under an inert N2 atmosphere, either in an MBraun glovebox or under standard Schlenk techniques. The solvents acetonitrile, toluene, dimethoxyethane (DME), pentane, and diethyl ether were sparged with nitrogen and passed over an activated alumina column prior to use. The solvents benzene, tetrahydrofuran, and n-hexane were dried over sodium-benzophenone ketal radical, refluxed, and distilled under nitrogen prior to use. All deuterated NMR solvents were purchased from Cambridge Isotope Laboratories. Benzene-d6 was dried over CaH2 and distilled under N2. The solvents CDCl3 and CD2Cl2 were dried over P2O5 and distilled under N2. Tetrahydrofuran-d8 and 2methyltetrahydrofuran used in the EPR experiments were dried over Na and distilled under nitrogen. All solvents were stored over 3 Å molecular sieves in a glovebox after purification. The triphos ligand (tP3) was prepared according to literature procedures.72 Trimethylphosphine was purchased from Strem Chemicals, Inc., and used as received. Anhydrous FeCl2 was purchased from Sigma-Aldrich, Inc., and used as received. 2,6Diisopropylaniline was purchased from Sigma-Aldrich, Inc., distilled under N2 from CaH2, and stored in the glovebox after purification. 2,4,6-Triisopropylaniline was prepared following literature procedures73,74 and dried by azeotropic removal of water in a Dean−Stark apparatus using benzene. FeCl2(dppe) was prepared as described in the literature.75 AgSbF6 was purchased form Sigma-Aldrich, Inc., and used as received. LiNHAr and LiNHAr* were prepared by addition of 1 equiv of 2.5 M nBuLi in hexanes to a cold (−78 °C) solution of the respective amine in hexane; after the mixture was stirred for 2 h and warmed to room temperature, the salts were collected by filtration, washed with hexane, and used without further purification. For additional experimental and instrumental details, see the Supporting Information. EPR Spectroscopy. EPR measurements were made on a Bruker E-680X spectrometer at X-band using a 4122 SHQE-W1 resonator. Cryogenic sample temperatures were achieved using an Oxford ESR900 cryostat together with an ITC-503 temperature controller. Low microwave powers were used to ensure that portions of the EPR spectra near g = 2 were not saturated. EPR data were simulated using EasySpin 5.2.23 running in the MATLAB 2018b environment and the MATLAB function “fminsearch” for fitting.76 Electronic Structure Calculations. Neutral Ru6 and Fe6 were optimized at the MP2 level of theory. Vertical ionization potentials are computed as the difference between the MP2 energies of the neutral and the CASPT2 energies of the cations at the neutral-optimized structures. An active space of seven electrons in four orbitals and a state average over four states was used for CASPT2 calculations of Ru6 and Fe6. All CASPT2 and MP2 calculations were performed using the cc-pVTZ basis for Fe, the cc-pVTZ-PP basis and effective core potentials for Ru, and the cc-pVDZ basis for all other atoms.77−80 The multistate variant of CASPT281 was used for all calculations. All calculations were performed with the MolPro software package.70,82−85 Orbital pictures were created with VMD.86 Improved Synthesis of cis-RuCl2(PMe3)4. A 35 mL pressure tube was charged with Ru(COD)Cl2 (1.00 g, 1 equiv), a stir bar, toluene (10 mL), and PMe3 (1.60 g, 6 equiv). The tube was sealed inside the glovebox and transferred to a 110 °C oil bath. The solution was stirred for 12 h. Over the reaction time, the solution changed from opaque brown to transparent yellow. The pressure tube was removed from heat and transferred to the glovebox. The reaction solution was concentrated to about 2 mL in vacuo to yield large blocky yellow crystals of cis-RuCl2(PMe3)4. The remaining reaction solution was decanted from the crystals and chilled to yield additional product. NMR spectroscopy of the material match published spectra.42 Yield: 1.4 g (83%). L

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

The mixture was filtered using Celite as a filtering agent, concentrated to ∼3 mL total volume, and layered with 4 mL of hexane. This layered solution was stored at −35 °C overnight to yield small, blue X-rayquality crystals of Fe4a (43 mg, 76%). Mp: 121 °C dec. UV−vis absorption (THF, 21 °C): 624 nm (3704 cm−1 M−1), 353 (5463.6 cm−1 M−1 ), 289 nm (13031 cm−1 M −1 ). Anal. Calcd for C25H50F6FeNP3SbSi: C, 38.63; H, 6.48; N, 1.80. Found: C, 38.60; H, 7.03; N, 1.73. μeff (THF-d8, 25 °C): 2.17 μB. E1/2 (THF, 21 °C, 0 V = Fc/Fc+): Fe2+/3+ −0.89 V (rev). Synthesis of Fe(dppe)(NHAr)2 (Fe5). Method A. A clean sample of X-ray-quality single crystals could not be obtained from the reaction mixture to synthesize Fe2. However, the presence of a red, paramagnetic impurity was noted in this reaction. Method B. A 20 mL scintillation vial was charged with FeCl2(dppe) (150 mg, 0.29 mmol, 1 equiv), 4 mL of THF, and a stir bar. This mixture was chilled at −78 °C. Separately, a solution of LiNHAr (105 mg, 0.58 mmol, 2 equiv) was prepared in 2 mL of room-temperature THF. The chilled suspension of FeCl2(dppe) was stirred, and the LiNHAr solution was added dropwise. Upon complete addition, the solution was opaque and red. The mixture was stirred for 4 h and warmed to room temperature. The volatiles were removed in vacuo, resulting in a dark red residue. This residue was extracted with diethyl ether and filtered using Celite as a filtering agent. The filtrate was concentrated in vacuo, and the concentrated solution was stored at −35 °C to yield large, red crystals of Fe5 (37 mg, 16%). μeff (benzene-d6, 25 °C): 5.19 μB. Synthesis of Ru(NAr*)(PMe3)3 (Ru1*). In an Erlenmeyer flask was placed a stir bar, 200 mg of cis-RuCl2(PMe3)4 (0.410 mmol, 1 equiv), and 20 mL of THF to give a pale yellow solution. This solution was chilled to −78 °C, and to this cold solution was added a room-temperature solution of LiNHAr* (196 mg, 0.875 mmol, 2.1 equiv) in 3 mL of THF dropwise. Upon complete addition, the solution had turned light orange. The reaction mixture was stirred overnight at room temperature to give a viscous dark red solution. The volatiles were removed in vacuo, and the residue was extracted with hexane until the filtrate was colorless. This extract was filtered using Celite as a filtering agent and concentrated in vacuo. The concentrated solution was stored in the freezer for 2−3 days at −35 °C to yield blocky, red-orange X-ray-quality crystals of Ru1* (162 mg, 70%). Note: due to the NH2Ar* generated upon imido production and the high solubility of Ru1* in aliphatic solvents, an analytically pure compound was obtained by repeated recrystallization from hexamethyldisiloxane. This resulted in a substantially reduced yield of approximately 12%. CHN analysis was taken of these single crystals. For the purposes of utilizing the complex in further reactions, specifically in the subsequent reaction to make Ru2*, samples of such high purity were not necessary, and a small amount of the 2,4,6triisopropylaniline was tolerable in samples of Ru1*. Mp: 112.7−114 °C. 1H NMR (500 MHz, benzene-d6): δ 7.04 (s, 2H), 4.42 (septet, J = 7.0 Hz, 2H), 3.03−2.45 (septet, J = 6.9 Hz, 1H), 1.43 (d, J = 7.0 Hz, 12H), 1.33 (d, J = 7.0 Hz, 6H), 1.29 (m, 26H). 13C{1H} NMR (126 MHz, benzene-d6): δ 139.50 (d), 138.95 (s), 138.32 (s), 132.03 (s), 120.46 (s), 119.62 (s), 34.89 (s), 34.14 (s), 27.97 (s), 26.59 (s), 26.32−25.40 (m), 24.50 (s), 24.20 (s), 23.22 (s), 22.33 (s). 31P NMR (202 MHz, benzene-d6): δ 19.47 (s). 14N NMR (36.5 MHz, benzened6): 328.0 (s). UV−vis absorption (THF, 21 °C): 459 nm (4808 cm−1 M−1), 338 (5642 cm−1 M−1), 270 nm (14016 cm−1 M−1). Anal. Calcd for C24H50NP3Ru: C, 52.73; H, 9.22; N, 2.56. Found: C, 52.79; H, 8.70; N, 2.57. Synthesis of Ru(NAr*)(PMe3)(dppe) (Ru2*). This compound was prepared similarly to Ru2 starting with a reaction solution of Ru1* (0.30 mmol, 1 equiv) and adding dppe (120 mg, 0.30 mmol, 1 equiv). The crude reaction mixture was stirred for ∼1 h at room temperature, and then volatiles were removed in vacuo. The sticky red residue was washed with hexamethyldisiloxane (3 × 2 mL) and once again dried in vacuo. The remaining solids were extracted with Et2O and filtered using Celite as a filtering agent until the extracts were colorless. The Et2O solution was then concentrated and stored at −35 °C to yield flaky, red crystals of Ru2* (98 mg, 41%). Unfortunately, these crystals demonstrate severe full-molecule disorder by X-ray

Synthesis of Fe(NAr)(PMe3)3 (Fe1). A 20 mL scintillation vial was charged with FeCl2 (50 mg, 0.394 mmol, 1 equiv), a magnetic stir bar, and 8 mL of THF. To this off-white suspension was added trimethylphosphine (0.25 mL, 2.37 mmol, 6 equiv) at room temperature. The mixture was stirred for 1 h, over which time the FeCl2 dissolved, and the solution changed color from pink to pale aquamarine. After this color change, the solution was chilled (−78 °C), while a separate solution of LiNHAr (152 mg, 0.827 mmol, 2.1 equiv) in 2 mL of THF was prepared. The chilled iron-containing mixture was stirred, and to it was added the LiNHAr solution dropwise. Upon addition, the solution turned orange. The solution was warmed to room temperature, and stirring was continued for 18 h, at which point the solution had turned dark green. Attempts to isolate Fe1 led to decomposition, but the complex is stable in the reaction solution for a few days with a slight excess of PMe3 present. 31 P NMR (127 MHz, THF, 20 °C): δ 38.37 (s). 14N NMR (36 MHz, THF, 20 °C): δ 312.1 (s). Synthesis of Fe(NAr)(PMe3)(dppe) (Fe2). An in situ generated solution of Fe1 (30 mg scale FeCl2) was stirred at room temperature. To this mixture was added dppe (84 mg, 0.21 mmol, 1 equiv) as a solution in 2 mL of THF. The resulting mixture was stirred for 1 h at room temperature. The volatiles were then removed in vacuo, and the resulting dark brown residue was rinsed with hexane (4 mL, discarded). The residue was then extracted with diethyl ether to give a bright green solution. This ether solution was filtered using Celite as a filtering agent, concentrated in vacuo, and stored at −35 °C overnight. This yielded flaky, dark green X-ray-quality crystals of Fe2 (64 mg, 39%). Mp: 134−135 °C. 1H NMR (500 MHz, benzened6): δ 8.05 (t, J = 8.7 Hz, 4H), 7.55 (t, J = 8.2 Hz, 4H), 7.34 (t, J = 7.6 Hz, 1H), 7.13 (t, J = 7.4 Hz, 4H), 7.10−7.00 (m, 9H), 6.97 (t, 1H), 4.78 (septet, J = 6.7 Hz, 2H), 1.89 (m, 4H), 1.33 (d, J = 6.8 Hz, 12H), 0.65 (d, J = 8.2 Hz, 9H). 13C{1H} NMR (126 MHz, benzene-d6): δ 156.53 (s), 141.58 (s), 140.18 (s), 133.84 (d), 130.81 (s), 128.72 (s), 120.94 (s), 120.46 (s), 35.13 (s), 27.00 (s), 23.90 (s), 23.56 (s), 22.32 (s), 21.73 (d). 31P NMR (202 MHz, benzene-d6): δ 115.16 (d, J = 6.5 Hz), 35.11 (t, J = 6.4 Hz). 14N NMR (36 MHz, THF): δ 325.6 (s). UV−vis absorption (THF, 21 °C): 591 nm (2009 cm−1 M−1), 388 (4471 cm−1 M−1), 298 nm (13460 cm−1 M−1). Anal. Calcd for C41H50NP3Fe: C, 69.79; H, 7.14; N, 1.99. Found: C, 69.40; H, 6.77; N, 1.77. Synthesis of Fe(NAr)tP3 (Fe4). An in situ solution of Fe1 (prepared using 100 mg of FeCl2, 0.79 mmol, 1 equiv) was stirred at room temperature. To the mixture was added tP3 (247 mg, 0.79 mmol, 1 equiv) as a solution in 2 mL of THF. The reaction solution rapidly changed color from dark green to dark purple upon addition. The resulting solution was stirred for 1 h at room temperature, and the volatiles were removed in vacuo. The resulting black residue was extracted with hexane and filtered using Celite as a filtering agent until the extracts were colorless. The filtrate was then concentrated in vacuo to ∼3 mL and stored at −35 °C overnight to yield blocky, dark purple X-ray-quality crystals of Fe4 (142 mg, 34%). Mp: 196 °C dec. 1 H NMR (500 MHz, benzene-d6): δ 7.15 (dd, J = 12.4, 4.7 Hz, 1H), 6.97 (d, J = 7.5 Hz, 2H), δ 4.31 (septet, J = 6.9 Hz, 2H), 1.66 (m, 18H), 1.38 (d, J = 7.0 Hz, 12H), 0.67 (s, 9H), 0.27−0.24 (m, 6H). 13 C{1H} NMR (126 MHz, benzene-d6): δ 159.73, 139.51, 122.35, 119.98, 28.26−27.65 (m), 27.41, 26.43, 23.44, 16.76 (q, J = 5.9 Hz), 10.74. 31P NMR (202 MHz, benzene-d6): δ 47.38. 14N NMR (36 MHz, THF): δ 315.2 (s). UV−vis absorption (THF, 21 °C): 632 nm (2809 cm−1 M−1), 549 (3327 cm−1 M−1), 383 nm (7015 cm−1 M−1). Anal. Calcd for C25H50FeNP3Si: C, 55.45; H, 9.31; N, 2.59. Found: C, 55.45; H, 9.54; N, 2.61. E1/2 (THF, 21 °C, 0 V = Fc/Fc+): Fe2+/3+ −0.91 V (rev). Synthesis of {Fe(NAr)tP3}SbF6 (Fe4a). A 20 mL scintillation vial was charged with Fe4 (40 mg, 0.0727 mmol, 1 equiv), a magnetic stir bar, and 4 mL of DME. This solution was stirred at room temperature. Separately, a solution of AgSbF6 (25 mg, 0.0727 mmol, 1 equiv) was prepared in 2 mL of DME, which was added dropwise to the stirred solution of 6. The resulting mixture was stirred for 12 h at room temperature, during which time the reaction solution changed from dark purple to bright blue and solid Ag0 precipitated. M

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

could not be separated from the reaction mixture, but the complex is stable for a few days in the reaction solution in the presence of excess PMe3. 31P NMR (127 MHz, THF): δ 43.06 (s). 14N NMR (36 MHz, THF): δ 320.6 (s). Synthesis of Fe(NAr*)(PMe3)(dppe) (Fe2*). This compound was prepared similarly to Fe2. A reaction mixture of Fe1* (50 mg, 1 equiv) and dppe (145 mg, 0.95 equiv) was used. After the reaction volatiles were removed in vacuo, the crude residue was extracted with pentane and filtered using Celite as a filtering agent. The filtrate was concentrated in vacuo and stored at −35 °C to yield dark green crystals of Fe2* (200 mg, 66%). These crystals were not of X-ray quality but were suitably pure for NMR spectroscopy and elemental analysis. X-ray-quality crystals were grown from a very dilute solution in n-hexane at −35 °C. Mp: 148.9−150.2 °C. 1H NMR (500 MHz, benzene-d6): δ 8.02 (t, J = 8.6 Hz, 4H), 7.52 (t, J = 8.1 Hz, 4H), 7.09 (d, J = 7.3 Hz, 3H), 7.06−6.96 (m, 8H), 6.93 (t, J = 7.2 Hz, 2H), 4.76 (septet, J = 6.9 Hz, 2H), 2.81−2.68 (septet, J = 6.8 Hz, 1H), 1.84 (m, J = 36.7, 21.9, 13.8 Hz, 4H), 1.33 (d, J = 6.9 Hz, 12H), 1.27 (d, J = 6.8 Hz, 6H), 0.62 (d, J = 8.3 Hz, 9H). 13C{1H} NMR (126 MHz, benzene-d6): δ 156.53 (s), 141.58 (s), 140.18 (s), 133.79 (s), 130.81 (s), 128.72 (s), 120.94 (s), 120.46 (s), 35.13 (s), 27.97 (s), 27.00 (s), 24.49 (s), 23.90 (s), 23.56 (s), 22.32 (s), 21.73 (d). 31P NMR (202 MHz, benzene-d6): δ 115.35 (d, J = 5.2 Hz), 34.93 (t, J = 5.2 Hz). 14N NMR (36 MHz, benzene-d6): δ 324.8. UV−vis absorption (THF, 21 °C): 589 nm (1754 cm−1 M−1), 404 (4114 cm−1 M−1), 301 nm (11596 cm−1 M−1). Anal. Calcd for C44H56FeNP3: C, 70.68; H, 7.55; N, 1.87. Found C, 71.48; H, 7.57; N, 1.75. Cyclic voltammetry (THF, 21 °C, 0 V = Fc/Fc+): E1/2(Fe2+/3+) = −0.75 V (rev). Synthesis of Fe(dppe)(NHAr*)2 (Fe5*). Method A. Fe5* was recovered as a byproduct from the synthesis of Fe2* (vide supra). This was accomplished by extracting the remaining solid residue, after pentane extraction, with diethyl ether, resulting in a red solution. This solution was filtered using Celite as a filtering agent, concentrated in vacuo to ∼1 mL, and stored at −35 °C for 4 days to yield large, red Xray-quality crystals of Fe5* (26 mg, 7%). Method B. A 20 mL scintillation vial was charged with FeCl2(dppe) (100 mg, 0.19 mmol, 1 equiv), 4 mL of THF, and a stir bar. This mixture was chilled at −78 °C. Separately, a solution of LiNHAr* (69 mg, 0.38 mmol, 2 equiv) was prepared in 2 mL of room-temperature THF. The chilled suspension of FeCl2(dppe) was stirred, and the LiNHAr* solution was added dropwise. Upon complete addition, the solution was opaque with a pink color. The mixture was stirred for 4 h and warmed to room temperature. The volatiles were removed in vacuo, resulting in a dark red residue. This residue was extracted with diethyl ether and filtered using Celite as a filtering agent. The filtrate was concentrated in vacuo, and the concentrated solution was stored at −35 °C to yield large, red crystals of Fe5* (47 mg, 28%). Mp: 144.4−146.6 °C. μeff (benzene-d6, 25 °C): 5.24 μB. Synthesis of {Fe(NAr*)(PMe3)(dppe)}SbF6 (Fe2a*). A 20 mL scintillation vial was charged with Fe2* (50 mg, 0.067 mmol, 1 equiv), a magnetic stir bar, and 4 mL of DME. This solution was stirred at room temperature. Separately, a solution of AgSbF6 (23 mg, 0.067 mmol, 1 equiv) was prepared in 2 mL of DME. The silver solution was added dropwise to the solution of Fe2*. The resulting mixture was stirred for 12 h at room temperature, during which time the reaction solution remained dark green and solid Ag0 precipitated. The mixture was filtered using Celite as a filtering agent, and the solvent was removed in vacuo. This yielded an oily green residue, which was rinsed with 5 mL of pentane and dried in vacuo to obtain a powdery green solid of Fe2a* (32 mg, 49%). Mp: 110 °C dec. UV− vis absorption (THF, 21 °C): 584 nm (1245 cm−1 M−1), 387 (2566 cm−1 M−1), 293 nm (5233 cm−1 M−1). μeff (THF-d8, 25 °C): 1.97 μB. Cyclic voltammetry: (THF, 21 °C, 0 V = Fc/Fc+): E1/2(Fe2+/3+) = −0.77 V (rev).

diffraction, and an adequate solution for the data could not be found. Mp: 188.2−189.6 °C. 1H NMR (500 MHz, benzene-d6): δ 8.18 (t, J = 7.9 Hz, 4H), 7.64 (m, 4H), 7.23 (s, 2H), 7.14 (dt, J = 19.7, 7.3 Hz, 8H), 7.09−6.96 (m, 4H), 5.01 (septet, J = 6.7 Hz, 2H), 2.97 (septet, J = 6.8 Hz, 1H), 2.04 (m, 2H), 1.87 (m, 2H), 1.53 (d, J = 6.5 Hz, 12H), 1.43 (d, J = 6.8 Hz, 6H), 0.70 (d, J = 9.6 Hz, 9H). 13C{1H} NMR (126 MHz, benzene-d6): δ 133.96 (d), 132.74 (t), 130.95 (s), 128.91 (s), 128.67−128.21 (m), 128.08 (s), 127.94 (s), 120.47 (s), 120.16 (d), 34.98 (s), 34.13 (s), 30.29 (t), 27.98 (s), 26.79 (s), 24.50 (s), 24.28 (s), 24.12 (s), 23.73 (s), 22.50−21.88 (m). 31P NMR (202 MHz, benzene-d6): δ 99.84 (d, J = 21.1 Hz, 2H), 23.02 (t, J = 21.2 Hz, 1H). 14N NMR (36 MHz, THF): δ 337.2 (s). UV−vis absorption (THF, 21 °C): 465 nm (3053 cm−1 M−1), 330 (8035 cm−1 M−1), 280 nm (10045 cm−1 M−1). Anal. Calcd for C44H56NP3Ru: C, 66.65; H, 7.12; N, 1.77. Found: C, 66.46; H, 7.04; N, 1.73. Cyclic voltammetry (THF, 21 °C, 0 V = Fc/Fc+): E1/2(Ru2+/3+) = −0.58 V (rev). Synthesis of {Ru(NAr*)(PMe3)(dppe)}SbF6 (Ru2a*). A 20 mL scintillation vial was charged with a stir bar, Ru2* (50 mg, 0.063 mmol, 1 equiv), and 4 mL of DME. This red-orange solution was stirred at room temperature. Separately, a solution of AgSbF6 (22 mg, 0.063 mmol, 1 equiv) was prepared in 2 mL of DME. The AgSbF6 was then added dropwise to the solution of Ru2* with vigorous stirring. The solution gradually changed from bright red to pinkishbrown, and a precipitate formed on the sides of the vial. The reaction mixture was stirred at room temperature for 24 h and then filtered using Celite as a filtering agent to remove Ag0. The filtrate was concentrated in vacuo to give a brown oil. The oily residue was washed with hexane until the filtrate was colorless. The residue was once again dried in vacuo and was then dissolved in a minimum amount of DME. The DME solution was layered with hexane and stored at −35 °C overnight to yield a fine pinkish brown powder (34 mg, 53%). The mother liquor was decanted, and the powder was dried in vacuo. The pink-brown powder was assigned as complex Ru2a* as identified by EPR spectroscopy. Mp: 147 °C (color change 80 °C). μeff (THF-d8, 25 °C): 1.47 μB. UV−vis absorption (THF, 21 °C): 612 nm (912 cm−1 M−1), 475 (1297 cm−1 M−1), 353 nm (2937 cm−1 M−1), 280 nm (6575 cm−1 M−1). Note: satisfactory elemental analysis was not obtained for Ru2a* after several attempts, presumably due to its high sensitivity. Sample masses were noted to change rapidly when sample holders when taken out of the inert atmosphere glovebox and into air, despite our best attempts. Synthesis of Ru(NAr*)(dmpe)2 (Ru3*). This complex was synthesized by following the same procedure as for Ru3. A 20 mL scintillation vial was charged with Ru1* (150 mg, 0.27 mmol, 1 equiv), 5 mL of THF, and a magnetic stir bar. To this red-orange solution was added a solution of dmpe (85 mg, 0.54 mmol, 2 equiv) in 2 mL of THF, dropwise, at room temperature. The solution was stirred for 4 h, over which time the solution became brownish yellow. The volatiles were removed in vacuo, and the residue was rinsed with several small aliquots of n-hexane. The solids were then dissolved in a minimum amount of THF and layered with n-hexane. The layered solution was stored at −35 °C overnight to yield platelike greenbrown crystals of Ru3* (42 mg, 25%). 1H NMR (500 MHz, benzened6): δ 7.19 (s, 1H), 7.05 (s, 1H), 3.19 (septet, J = 7.1 Hz, 1H), 2.91 (septet, J = 7.1 Hz, 2H), 1.61 (d, J = 6.9 Hz, 6H), 1.38 (d, J = 6.8 Hz, 12H), 1.35−1.31 (m, 2H), 1.21 (m, 12H), 1.03 (s, 6H), 0.98−0.90 (m, 3H), 0.86 (d, J = 5.5 Hz, 6H), 0.84 (t, J = 1.6 Hz, 3H). 13C{1H} NMR (126 MHz, benzene-d6): δ 125.81, 120.42, 118.55, 108.76 (t, J = 8.0 Hz), 34.26, 33.26 (d, J = 17.1 Hz), 30.28, 27.99, 26.56, 25.79 (d, J = 7.5 Hz), 23.92 (dd, J = 13.9, 7.5 Hz), 22.94 (d, J = 8.6 Hz), 22.36, 21.07 (t, J = 7.6 Hz), 19.27, 13.03 (t, J = 13.5 Hz). 31P NMR (202 MHz, benzene-d6): δ 42.22 (t, J = 14.5 Hz), 30.24 (t, J = 14.7 Hz). 14 N NMR (36 MHz, benzene-d6): δ 595.3. Anal. Calcd for C27H55RuNP4: C, 52.42; H, 8.96; N, 2.26. Found: C, 53.24; H, 9.11; N, 2.27. HRMS (ESI+): calcd, 618.2558; found, 618.2336 (m/ z). Synthesis of Fe(NAr*)(PMe3)3 (Fe1*). This compound was prepared similarly to compound Fe1 using FeCl2 (50 mg, 1 equiv), PMe3 (180 mg, 6 equiv), and LiNHAr* (185 mg, 2.1 equiv). Similarly, it was found to decompose during attempted isolation and N

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

■

Reactivity in Bond Activation and Bond Formation Reactions. Comments Inorg. Chem. 2016, 36, 92−122. (3) Ezhova, M. B.; James, B. R. Catalytic oxidations using ruthenium porphyrins. Catal. Met. Complexes 2003, 26, 1. (4) Park, J. Y.; Kim, Y.; Bae, D. Y.; Rhee, Y. H.; Park, J. Ruthenium Bisammine Complex and Its Reaction with Aryl Azides. Organometallics 2017, 36, 3471−3476. (5) Wang, L.; Hu, L.; Zhang, H.; Chen, H.; Deng, L. ThreeCoordinate Iron(IV) Bisimido Complexes with Aminocarbene Ligation: Synthesis, Structure, and Reactivity. J. Am. Chem. Soc. 2015, 137, 14196−14207. (6) Wilding, M. J. T.; Iovan, D. A.; Betley, T. A. High-Spin Iron Imido Complexes Competent for C-H Bond Amination. J. Am. Chem. Soc. 2017, 139, 12043−12049. (7) Zhang, L.-L.; Li, L.-H.; Wang, Y.-Q.; Yang, Y.-F.; Liu, X.-Y.; Liang, Y.-M. Ruthenium-Catalyzed Direct C-H Amidation of Arenes: A Mechanistic Study. Organometallics 2014, 33, 1905−1908. (8) Kee, T. P.; Park, L. Y.; Robbins, J.; Schrock, R. R. Synthesis of the ruthenium imido complexes, [Ru(η-C6H6)(N-2,6-R2C6H3)]2 (R = Pri or Me), and the crystal structure of [Ru(η-C6H6)(N-2,6Pri2C6H3)]2. J. Chem. Soc., Chem. Commun. 1991, 0, 121−122. (9) Iovan, D. A.; Betley, T. A. Characterization of Iron-Imido Species Relevant for N-Group Transfer Chemistry. J. Am. Chem. Soc. 2016, 138, 1983−1993. (10) Hennessy, E. T.; Betley, T. A. Complex N-Heterocycle Synthesis via Iron-Catalyzed, Direct C-H Bond Amination. Science 2013, 340, 591−595. (11) Park, Y.; Kim, Y.; Chang, S. Transition Metal-Catalyzed C-H Amination: Scope, Mechanism, and Applications. Chem. Rev. 2017, 117, 9247−9301. (12) Shang, R.; Ilies, L.; Nakamura, E. Iron-Catalyzed C-H Bond Activation. Chem. Rev. 2017, 117, 9086−9139. (13) Wilding, M. J. T.; Iovan, D. A.; Wrobel, A. T.; Lukens, J. T.; MacMillan, S. N.; Lancaster, K. M.; Betley, T. A. Direct Comparison of C-H Bond Amination Efficacy through Manipulation of NitrogenValence Centered Redox: Imido versus Iminyl. J. Am. Chem. Soc. 2017, 139, 14757−14766. (14) Xiong, T.; Zhang, Q. New amination strategies based on nitrogen-centered radical chemistry. Chem. Soc. Rev. 2016, 45, 3069− 3087. (15) King, E. R.; Hennessy, E. T.; Betley, T. A. Catalytic CH Bond Amination from High-Spin Iron Imido Complexes. J. Am. Chem. Soc. 2011, 133, 4917−4923. (16) Anderson, J. S.; Cutsail, G. E.; Rittle, J.; Connor, B. A.; Gunderson, W. A.; Zhang, L.; Hoffman, B. M.; Peters, J. C. Characterization of an Fe≡N−NH2 Intermediate Relevant to Catalytic N2 Reduction to NH3. J. Am. Chem. Soc. 2015, 137, 7803−7809. (17) Buscagan, T. M.; Oyala, P. H.; Peters, J. C. N2 to NH3 Conversion by a triphos-Iron Catalyst and Enhanced Turnover under Photolysis. Angew. Chem., Int. Ed. 2017, 56, 6921−6926. (18) Chalkley, M. J.; Del Castillo, T. J.; Matson, B. D.; Peters, J. C. Fe-Mediated Nitrogen Fixation with a Metallocene Mediator: Exploring pKa Effects and Demonstrating Electrocatalysis. J. Am. Chem. Soc. 2018, 140, 6122−6129. (19) Matson, B. D.; Peters, J. C. Fe-Mediated HER vs N2RR: Exploring Factors That Contribute to Selectivity in (P3Fe)-Fe-E(N2) (E = B, Si, C) Catalyst Model Systems. ACS Catal. 2018, 8, 1448− 1455. (20) Rittle, J.; Peters, J. C. An Fe-N2 Complex That Generates Hydrazine and Ammonia via Fe = NNH2: Demonstrating a Hybrid Distal-to-Alternating Pathway for N2 Reduction. J. Am. Chem. Soc. 2016, 138, 4243−4248. (21) Thompson, N. B.; Green, M. T.; Peters, J. C. Nitrogen Fixation via a Terminal Fe(IV) Nitride. J. Am. Chem. Soc. 2017, 139, 15312− 15315. (22) Figg, T. M.; Holland, P. L.; Cundari, T. R. Cooperativity Between Low-Valent Iron and Potassium Promoters in Dinitrogen Fixation. Inorg. Chem. 2012, 51, 7546−7550.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.9b01672. Synthetic and characterization details, cyclic voltammograms, details on EPR studies, details on the computational methods, and NMR spectra of diamagnetic compounds (PDF) Accession Codes

CCDC 1856951−1856961 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing [email protected], or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail *E-mail *E-mail *E-mail

for for for for

B.G.L.: [email protected]. J.M.: [email protected]. M.R.S.: [email protected]. A.L.O.: [email protected].

ORCID

B. Scott Fales: 0000-0002-0627-2782 Amrendra K. Singh: 0000-0001-7263-7914 Richard J. Staples: 0000-0003-2760-769X Benjamin G. Levine: 0000-0002-0356-0738 Milton R. Smith, III: 0000-0002-8036-4503 Aaron L. Odom: 0000-0001-8530-4561 Present Addresses

† Dr. Kelly E. Aldrich, Los Alamos National Laboratory, Chemistry Division, Los Alamos, NM 87544. ‡ Dr. B. Scott Fales, Department of Chemistry and the PULSE Institute, Stanford University, Stanford, CA 94305. § Dr. Amrendra K. Singh, Chemistry Discipline, School of Basic Science, Indian Institute of Technology-Indore, Simrol, Indore 453552, M.P., India.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS A.L.O. thanks Prof. Jonas Peters and Prof. Theodore Betley for helpful discussion. A.L.O., K.E.A., and A.K.S. thank the NSF under grant CHE1562140 for funding. A.L.O. and K.E.A. thank the donors of the American Chemical Society Petroleum Research Fund for support (57670-ND3). M.R.S., A.L.O., and A.K.S. thank MSU for funding. M.R.S. thanks the U.S. Department of Energy Catalysis Science Program for support (DE-SC001664). B.S.F. and B.G.L. acknowledge support from the National Science Foundation under Grant No. CHE1565634. K.E.A. thanks the American Association of University Women for support from the 2018−2019 American Dissertation Fellowship.

■

REFERENCES

(1) Nugent, W. A.; Mayer, J. M.: Metal-Ligand Multiple Bonds: The Chemistry Of Transition Metal Complexes Containing Oxo, Nitrido, Imido, Alkylidene, Or Alkylidyne Ligands; Wiley: 1988. (2) Bellow, J. A.; Yousif, M.; Groysman, S. Discrete Complexes of 3d Metals with Monodentate Bulky Alkoxide Ligands and Their O

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry (23) Smith, J. M.; Lachicotte, R. J.; Pittard, K. A.; Cundari, T. R.; Lukat-Rodgers, G.; Rodgers, K. R.; Holland, P. L. Stepwise reduction of dinitrogen bond order by a low-coordinate iron complex. J. Am. Chem. Soc. 2001, 123, 9222−9223. (24) Habibzadeh, F.; Miller, S. L.; Hamann, T. W.; Smith, M. R. Homogeneous electrocatalytic oxidation of ammonia to N-2 under mild conditions. Proc. Natl. Acad. Sci. U. S. A. 2019, 116, 2849−2853. (25) Modak, J. M. Haber process for ammonia synthesis. Resonance 2002, 7, 69−77. (26) Anderson, J. S.; Rittle, J.; Peters, J. C. Catalytic conversion of nitrogen to ammonia by an iron model complex. Nature 2013, 501, 84−87. (27) Hoffman, B. M.; Dean, D. R.; Seefeldt, L. C. Climbing Nitrogenase: Toward a Mechanism of Enzymatic Nitrogen Fixation. Acc. Chem. Res. 2009, 42, 609−619. (28) Eady, R. R. Structure-Function Relationships of Alternative Nitrogenases. Chem. Rev. 1996, 96, 3013−3030. (29) Barney, B. M.; Lee, H.-I.; Dos Santos, P. C.; Hoffman, B. M.; Dean, D. R.; Seefeldt, L. C. Breaking the N2 triple bond: insights into the nitrogenase mechanism. Dalton Trans. 2006, 2277−2284. (30) Burgess, B. K.; Lowe, D. J. Mechanism of Molybdenum Nitrogenase. Chem. Rev. 1996, 96, 2983−3012. (31) Schrock, R. R. Catalytic Reduction of Dinitrogen to Ammonia at a Single Molybdenum Center. Acc. Chem. Res. 2005, 38, 955−962. (32) Hohenberger, J.; Ray, K.; Meyer, K. The biology and chemistry of high-valent iron−oxo and iron−nitrido complexes. Nat. Commun. 2012, 3, 720. (33) Coric, I.; Holland, P. L. Insight into the Iron−Molybdenum Cofactor of Nitrogenase from Synthetic Iron Complexes with Sulfur, Carbon, and Hydride Ligands. J. Am. Chem. Soc. 2016, 138, 7200− 7211. (34) Kandemir, T.; Schuster, M. E.; Senyshyn, A.; Behrens, M.; Schlögl, R. The Haber−Bosch Process Revisited: On the Real Structure and Stability of “Ammonia Iron” under Working Conditions. Angew. Chem., Int. Ed. 2013, 52, 12723−12726. (35) Coia, G. M.; White, P. S.; Meyer, T. J.; Wink, D. A.; Keefer, L. K.; Davis, W. M. Preparation of osmium hydrazido complexes by interception of an osmium(iv) imido intermediate. J. Am. Chem. Soc. 1994, 116, 3649−3650. (36) Thompson, M. S.; Meyer, T. J. Oxidation of coordinated ammonia to nitrate. J. Am. Chem. Soc. 1981, 103, 5577−5579. (37) Little, D. J.; Smith, M. R., III; Hamann, T. W. Electrolysis of liquid ammonia for hydrogen generation. Energy Environ. Sci. 2015, 8, 2775−2781. (38) Brown, S. D.; Peters, J. C. Ground-State Singlet L3Fe-(μ-N)FeL3 and L3Fe(NR) Complexes Featuring Pseudotetrahedral Fe(II) Centers. J. Am. Chem. Soc. 2005, 127, 1913−1923. (39) Bucinsky, L.; Breza, M.; Lee, W.-T.; Hickey, A. K.; Dickie, D. A.; Nieto, I.; DeGayner, J. A.; Harris, T. D.; Meyer, K.; Krzystek, J.; Ozarowski, A.; Nehrkorn, J.; Schnegg, A.; Holldack, K.; Herber, R. H.; Telser, J.; Smith, J. M. Spectroscopic and Computational Studies of Spin States of Iron(IV) Nitrido and Imido Complexes. Inorg. Chem. 2017, 56, 4751−4768. (40) Cowley, R. E.; DeYonker, N. J.; Eckert, N. A.; Cundari, T. R.; DeBeer, S.; Bill, E.; Ottenwaelder, X.; Flaschenriem, C.; Holland, P. L. Three-Coordinate Terminal Imidoiron(III) Complexes: Structure, Spectroscopy, and Mechanism of Formation. Inorg. Chem. 2010, 49, 6172−6187. (41) Burred, A. K.; Steedman, A. J. (η6-p-Cymene)Ru-μ-N(2,4,6-tritert-butylphenyl): A monomeric ruthenium(II) complex containing a terminal imido ligand. J. Chem. Soc., Chem. Commun. 1995, 2109− 2110. (42) Singh, A. K.; Levine, B. G.; Staples, R. J.; Odom, A. L. A 4coordinate Ru(II) imido: unusual geometry, synthesis, and reactivity. Chem. Commun. 2013, 49, 10799−10801. (43) Burrell, A. K.; Steedman, A. J. Synthesis, Structure, and Reactivity of Ruthenium(II) Terminal Imido Complexes. Organometallics 1997, 16, 1203−1208.

(44) Zart, M. K.; Powell, D.; Borovik, A. S. Synthesis and structures of dimeric iron(III)-Oxo and -imido complexes containing intramolecular hydrogen bonds. Inorg. Chim. Acta 2007, 360, 2397−2402. (45) Song, J. S.; Han, S. H.; Nguyen, S. T.; Geoffroy, G. L.; Rheingold, A. L. Reactivity of triiron and triruthenium μ3-phenylimido clusters with alkynes, allene, and 1,3-cyclohexadiene. Organometallics 1990, 9, 2386−2395. (46) Liu, Y.; Du, J.; Deng, L. Synthesis, Structure, and Reactivity of Low-Spin Cobalt(II) Imido Complexes [(Me3P)3Co(NAr)]. Inorg. Chem. 2017, 56, 8278−8286. (47) Bai, G.; Stephan, D. W. Formation of C-C and C-N Bonds in Ni(II) Ketimide Complexes via Transient Ni(III) Aryl Imides. Angew. Chem., Int. Ed. 2007, 46, 1856−1859. (48) Del Castillo, T. J.; Thompson, N. B.; Suess, D. L. M.; Ung, G.; Peters, J. C. Evaluating Molecular Cobalt Complexes for the Conversion of N2 to NH3. Inorg. Chem. 2015, 54, 9256−9262. (49) Bakhoda, A.; Jiang, Q.; Bertke, J. A.; Cundari, T. R.; Warren, T. H. Elusive Terminal Copper Arylnitrene Intermediates. Angew. Chem., Int. Ed. 2017, 56, 6426−6430. (50) Schofield, M. H.; Kee, T. P.; Anhaus, J. T.; Schrock, R. R.; Johnson, K. H.; Davis, W. M. Osmium imido complexes: synthesis, reactivity, and SCF-Xα-SW electronic structure. Inorg. Chem. 1991, 30, 3595−3604. (51) Muniz, K. Imido-osmium(VIII) compounds in organic synthesis: aminohydroxylation and diamination reactions. Chem. Soc. Rev. 2004, 33, 166−174. (52) Anhaus, J. T.; Kee, T. P.; Schofield, M. H.; Schrock, R. R. Planar ″20-electron″ osmium imido complexes. A linear imido ligand does not necessarily donate its lone pair of electrons to the metal. J. Am. Chem. Soc. 1990, 112, 1642−1643. (53) Danopoulos, A. A.; Wilkinson, G.; Hussain-Bates, B.; Hursthouse, M. B. Synthesis and X-ray crystal structure of transbis(2,6-diisopropylphenylimido) bis-(trimethylphosphino)ruthenium(IV): The first structural determination of a terminal imido ruthenium compound. Polyhedron 1992, 11, 2961−2964. (54) Aldrich, K. E.; Odom, A. L. A photochemical route to a square planar, ruthenium(IV)- bis(imide). Chem. Commun. 2019, 55, 4403− 4406. (55) Leung, S. K.-Y.; Tsui, W.-M.; Huang, J.-S.; Che, C.-M.; Liang, J.-L.; Zhu, N. Imido Transfer from Bis(imido)ruthenium(VI) Porphyrins to Hydrocarbons: Effect of Imido Substituents, CH Bond Dissociation Energies, and RuVI/V Reduction Potentials. J. Am. Chem. Soc. 2005, 127, 16629−16640. (56) Zhou, X.-G.; Yu, X.-Q.; Huang, J.-S.; Che, C.-M. Asymmetric amidation of saturated C-H bonds catalysed by chiral ruthenium and manganese porphyrins. Chem. Commun. 1999, 2377−2378. (57) Huang, J.-S.; Che, C.-M.; Poon, C.-K. Synthesis and spectroscopy of tert-butylimido complexes of osmium(VI) and ruthenium(VI) porphyrins. J. Chem. Soc., Chem. Commun. 1992, 161−163. (58) Danopoulos, A. A.; Wilkinson, G.; Sweet, T. K. N.; Hursthouse, M. B. Reactions of imido complexes of iridium, rhodium and ruthenium. J. Chem. Soc., Dalton Trans. 1996, 3771−3778. (59) Berry, J. F. Terminal nitrido and imido complexes of the late transition metals. Comments Inorg. Chem. 2009, 30, 28−66. (60) Oakdale, J. S.; Sit, R. K.; Fokin, V. V. Ruthenium-Catalyzed Cycloadditions of 1-Haloalkynes with Nitrile Oxides and Organic Azides: Synthesis of 4-Haloisoxazoles and 5-Halotriazoles. Chem. Eur. J. 2014, 20, 11101−11110. (61) Karsch, H. H. Funktionelle Trimethylphosphinderivate, IV. Hydrido(phosphin)eisen(II)-Komplexe. Chem. Ber. 1977, 110, 2222− 2235. (62) Hoyt, J. M.; Shevlin, M.; Margulieux, G. W.; Krska, S. W.; Tudge, M. T.; Chirik, P. J. Synthesis and Hydrogenation Activity of Iron Dialkyl Complexes with Chiral Bidentate Phosphines. Organometallics 2014, 33, 5781−5790. (63) Hermes, A. R.; Girolami, G. S. Fourteen-electron organoiron species: synthesis and characterization of high-spin iron(II) alkyls and P

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry the crystal structure of the benzyl derivative Fe(CH2C6H4CH3)2(dippe). Organometallics 1987, 6, 763−768. (64) Stokes, S. L.; Davis, W. M.; Odom, A. L.; Cummins, C. C. Lowcoordinate iron complexes stabilized by N-(tert-hydrocarbyl)anilide ligation: Adduct formation, chemical oxidation, and nitric oxide complexation. Organometallics 1996, 15, 4521−4530. (65) Maire, P.; Konigsmann, M.; Sreekanth, A.; Harmer, J.; Schweiger, A.; Grutzmacher, H. A tetracoordinated rhodium aminyl radical complex. J. Am. Chem. Soc. 2006, 128, 6578−6580. (66) Kaim, W. Mononuclear and binuclear tricarbonyl and tetracarbonyl complexes of chromium(0), molybdenum(0), and tungsten(0) with the 2,2’-bipyrimidine radical-anion. Inorg. Chem. 1984, 23, 3365−3368. (67) Kaim, W. The transition-metal coordination chemistry of anion radicals. Coord. Chem. Rev. 1987, 76, 187−235. (68) de Bruin, B.; Bill, E.; Bothe, E.; Weyhermuller, T.; Wieghardt, K. Molecular and electronic structures of bis(pyridine-2,6-diimine)metal complexes ML2 (PF6)n (n = 0, 1, 2, 3; M = Mn, Fe, Co, Ni, Cu, Zn). Inorg. Chem. 2000, 39, 2936−2947. (69) Oliver, S. W.; Smith, T. D.; Pilbrow, J. R.; Harvey, T. G.; Matheson, T. W.; Pratt, K. C. An electron-spin-resonance study of zeolite-supported ruthenium hydrodenitrogenation catalyst. Inorg. Chim. Acta 1986, 117, L9−L10. (70) Celani, P.; Werner, H. J. Multireference perturbation theory for large restricted and selected active space reference wave functions. J. Chem. Phys. 2000, 112, 5546−5557. (71) Andersson, K.; Malmqvist, P. A.; Roos, B. O. 2nd-order perturbation-theory with a complete active space self-consistent field reference function. J. Chem. Phys. 1992, 96, 1218−1226. (72) Karsch, H. H.; Appelt, A. Functional trimethylphosphane derivatives. 18. Methyl(phosphinomethyl)silanes and methyl(phosphinomethyl)stannanes. Z. Naturforsch., B: J. Chem. Sci. 1983, 38, 1399−1405. (73) Thoreson, K. A.; Follett, A. D.; McNeill, K. Synthesis and Characterization of Pentaphosphino Zero-Valent Iron Complexes and Their Corresponding Iron(II)-Chloride and -Hydride Complexes. Inorg. Chem. 2010, 49, 3942−3949. (74) Liu, J.-Y.; Zheng, Y.; Li, Y.-G.; Pan, L.; Li, Y.-S.; Hu, N.-H. Fe(II) and Co(II) pyridinebisimine complexes bearing different substituents on ortho- and para-position of imines: synthesis, characterization and behavior of ethylene polymerization. J. Organomet. Chem. 2005, 690, 1233−1239. (75) Evans, D. J.; Hitchcock, P. B.; Leigh, G. J.; Nicholson, B. K.; Niedwieski, A. C.; Nunes, F. S.; Soares, J. F. The synthesis of triangulo-trimetal complexes containing both iron(II) and vanadium(II). Inorg. Chim. Acta 2001, 319, 147−158. (76) Stoll, S.; Schweiger, A. EasySpin, a comprehensive software package for spectral simulation and analysis in EPR. J. Magn. Reson. 2006, 178, 42−55. (77) Dunning, T. H. Gaussian-basis sets for use in correlated molecular calculations. 1. The atoms boron through neon and hydrogen. J. Chem. Phys. 1989, 90, 1007−1023. (78) Woon, D. E.; Dunning, T. H. Gaussian-basis sets for use in correlated molecular calculations. 3. The atoms aluminum through argon. J. Chem. Phys. 1993, 98, 1358−1371. (79) Balabanov, N. B.; Peterson, K. A. Systematically convergent basis sets for transition metals. I. All-electron correlation consistent basis sets for the 3d elements Sc-Zn. J. Chem. Phys. 2005, 123, 064107. (80) Peterson, K. A.; Figgen, D.; Dolg, M.; Stoll, H. Energyconsistent relativistic pseudopotentials and correlation consistent basis sets for the 4d elements Y-Pd. J. Chem. Phys. 2007, 126, 124101. (81) Finley, J.; Malmqvist, P. A.; Roos, B. O.; Serrano-Andres, L. The multi-state CASPT2 method. Chem. Phys. Lett. 1998, 288, 299− 306. (82) Knowles, P. J.; Werner, H. J. An efficient 2nd-order MC SCF method for long configuration expansions. Chem. Phys. Lett. 1985, 115, 259−267.

(83) Werner, H.-J.; Knowles, P. J. A second order multiconfiguration SCF procedure with optimum convergence. J. Chem. Phys. 1985, 82, 5053−5063. (84) Werner, H. J. Third-order multireference perturbation theory The CASPT3 method. Mol. Phys. 1996, 89, 645−661. (85) Werner, H. J.; Knowles, P. J.; Knizia, G.; Manby, F. R.; Schutz, M. Molpro: a general-purpose quantum chemistry program package. Wires Comput. Mol. Sci. 2012, 2, 242−253. (86) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual molecular dynamics. J. Mol. Graphics 1996, 14, 33−38.

Q

DOI: 10.1021/acs.inorgchem.9b01672 Inorg. Chem. XXXX, XXX, XXX−XXX