SEARCH

SEARCH BY CITATION

Keywords:

  • Cooperative catalysis;
  • Pincer complexes;
  • N ligands;
  • Homogeneous catalysis­;
  • C–H activation

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information

In this review, the coordination chemistry of electron-rich metal complexes with the simple aliphatic, anionic diphosphanylamido ligand {N(CH2CH2PR2)2} is covered and compared with other commonly used, anionic PEP (E = C, N) pincer ligands. The strong π-basicity of this ligand enables both the stabilization of electronically and coordinatively highly unsaturated complexes and their use as cooperating ligands in bifunctional stoichiometric bond activation reactions and catalysis. Versatile ligand backbone dehydrogenation gives access to related enamido and dienamido ligands {(R2PCHCH)N(CH2CH2PR2)} and {N(CHCHPR2)2}, respectively. This oxidative functionalization enables fine-tuning of the ligand donor properties and thereby of the structural features, electronic structure, and reactivity of the respective complexes, which is discussed for several examples.


1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information

“Pincer ligands” are tridentate ligands that coordinate to a metal center in a meridional fashion.1 For many pincer ligand frameworks, bulky substituents are easily introduced to sterically protect vacant coordination sites and allow for reactivity control by rational ligand design. Consequently, pincer complexes have found widespread use for several applications, such as catalysis, sensing, or metal-centered bond activation.1,2 Electronic effects that arise from pincer modifications, particularly at the central ligand framework in trans-position with respect to the substrate binding site, offer further synthetic strategies for catalysis and bond activation reactions. As an example, variation of electron-donating/withdrawing substituents X on the aryl backbone of “archetypal” PCP pincer ligand {p-X-C6H2-2,6-(CH2PR2)2} was shown to have a strong influence on the thermodynamics of the oxidative addition of E–H (E = H, C) to a d8 metal center, attributable to M–CPCP π-MO interactions.3 Alternatively, rigid, anionic pincer ligands with a central amido unit derived from diarylamines (Figure 1, {LAR}) or from pyridine-based chelates upon benzylic deprotonation (Figure 1, {LBR}) have recently attracted considerable attention. The latter were extensively utilized for novel dehydrogenative coupling reactions of polar functional groups as metal–ligand cooperative catalysts based on reversible amide/pyridine de-/aromatization.1,46

thumbnail image

Figure 1. Amido pincer ligands structurally related to the aliphatic PNP ligands discussed here.

Download figure to PowerPoint

In comparison, despite seminal work from Fryzuk's group using a disilylamido PNP ligand (Figure 1, {LCR }) in combination with d6 and d8 metal ions, aliphatic amido pincer ligands were neglected for some time.7 Amido ligands with C–H bonds in the α-position are generally prone to β-hydride elimination giving rise to imine formation.8 Hence, the expected lower thermal stabilities and more complex reactivity patterns from ligand chemical non-innocence might have impeded the use of alkylamido ligands to some extent. However, chelating amido ligands exhibit higher barriers for imine extrusion as compared to terminal amides. Particularly, the great success of “bifunctional catalysis” for the hydrogenation of polar multiple bonds sparked great interest in flexible, chelating ligands that can facilitate H2 heterolysis by the stabilization of both amido and amino intermediates, which occur during ionic hydrogenation.9

This review covers the recent work with aliphatic, anionic PNP ligands, particularly ethylene bridged ligand N(CH2CH2PR2)2 ({L1R}), and related ligands derived by simple ligand backbone functionalization (Figure 2). These oxidative ligand modifications are discussed in combination with their impact on donor properties and electronic structure and reactivity of the respective complexes. After a concise introduction into the bonding and reactivity of amido ligands coordinated to electron-rich transition metal centers (section 1.1), the properties of PNP amido, enamido, and dienamido ligands are compared with those of other common pincer ligands (section 2). Finally, their use for the stabilization of unusual, electronically unsaturated complexes (section 3) and for metal–ligand cooperative bond activation and catalysis (section 4) are discussed.

1.1. Electron-Rich Transition Metal Amides: Bonding, Electronic Structure, and Reactivity

Covalently bound π-donating ligands, such as amido ligand (NR2), are frequently used to stabilize complexes ofelectron-poor (early) transition metals in high oxidation states.10 In contrast, until more recently electron-rich (late) transition metal complexes in low oxidation states (d6–d10) with these ligands were comparatively rare.11 However, the participation of late transition metal amido complexes in catalytic transformations, such as C–N cross coupling reactions,12 has stimulated several studies on M–NR2 bonding. Pearson's hard and soft acid and base (HSAB) theory predicts weak bonding of hard, π-donating amido ligands with electron-rich metal centers.13 Nevertheless, Bercaw and co-workers demonstrated a linear relationship of M–R (M = RuII, PtII; R = alkyl, aryl, acetylide, hydride, alkoxide, and amide) bond dissociation energies (BDEs) with the corresponding H–R BDEs.14 On the other hand, comparison of metal amido hydrogenolysis along the transition series (LnM–NR2 + H2 [lrarr2] LnM–H + HNR2) provides some qualitative information with respect to relative M–Namido bond energetics:15 While d6 and d8 anilido complexes [Cp*Ru(NHPh)(PMe3)2] or [PtH(NHPh)(PEt3)2] readily react with H2 to give the corresponding hydride complex and free amine, [Cp*2Zr(H)2] exhibits reverse reactivity and releases H2 with NH2Ph.14,16 This reactivity is in line with calculated relative BDEs {BDEM–H – BDEM–N: 151 kJ mol–1 [Cp*Ru(PMe3)2]; 121 kJ mol–1 [PtH(PEt3)2]; 4 kJ mol–1 (Cp*2ZrH)}. These results can be rationalized on the basis of M–N bonding: High anionic charge density on the nitrogen atom is expected as a result of polar metal–amido σ-bonding.17 If vacant metal orbitals for N[RIGHTWARDS ARROW]M π-donation are available, the charge will be effectively delocalized, strengthening the M–N bond as in the case of d0 complex [Cp*2Zr(NHPh)(H)].18 For metal centers without empty d orbitals of suitable symmetry and energy, for example, octahedral d6 or square-planar d8 complexes, no net M–N π-bond results from the M–N π-interaction.19

thumbnail image

Figure 2. Neutral and anionic ligands derived from parent HN(CH2CH2PR2)2.

Download figure to PowerPoint

This lack of charge delocalization results in high basicity and nucleophilicity of the amido ligand, exemplified by C–N coupling with C-electrophiles.20,21 Proton transfer to the amido ligand upon H2 heterolysis was first demonstrated by Fryzuk and co-workers for d6 and d8 complexes with disilylamido pincer ligand {LCR}.22 This reaction was later recognized as key step for “bifunctional” (also denoted “metal–ligand cooperative” or “Noyori–Morris type” catalysis or initially as “N–H effect”) hydrogenation catalysis and was therefore extensively studied.9 Other than H2, weakly acidic sp, sp2, and sp3 C–H bonds were also shown to add across M–Namido bonds.21a,23

However, cooperativity of amido ligands is not restricted to acid–base chemistry. A mainly N-centered HOMO suggests potential ligand redox non-innocence for amido radical complexes or, in terms of a simple Lewis formalism, a description as aminyl (N-radical) rather than amido (metalloradical) complex (Figure 3).24 The electronic structure of amido radical complexes is of relevance for catalysis, for example, oxidative dehydrogenation of amines, which is particularly well promoted by metals of the iron triad.25 The oxidation of alcohols to aldehydes catalyzed by the oxidoreductase enzyme galactose oxidase provides a related example in nature.26 In the active site, the copper phenoxide moiety undergoes both metal- and ligand-centered one-electron redox steps to accomplish the overall two-electron redox reaction.

thumbnail image

Figure 3. Canonic Lewis structures for open-shell amido complexes.

Download figure to PowerPoint

para-Aryl C–C-coupling was reported in early studies as a consequence of platinum(II) anilide oxidation, pointing towards ligand-centered radical reactivity.27 EPR spectroscopic characterization revealed some metal-to-ligand spin transfer for radical complex [Cp*Mn(CO)2{N(H)C6H4-4-NMe2}] (Cp* = η5-C5Me5).28 Wieghardt and co-workers isolated the first examples that could be unequivocally assigned to persistent anilino radical complexes (Figure 4, A).29 In recent years, several other aminyl complexeswere spectroscopically characterized or isolated (Figure 4).21d,3034 For example, spectroscopic and quantum chemical examination of radical complex [NiCl{LAiPr}]+ (Figure 4, B) indicates considerable spin-density delocalization into the diarlyamido PNP pincer ligand and, in turn, nickel(II) as the most appropriate assignment for the metal oxidation state.31 Besides the relevance of transient aminyl complexes as possible intermediates in oxidation catalysis, more persistent aminyl complexes might allow for the utilization of amido non-innocence in metal–ligand cooperative catalysis.35 As a rare example, Grützmacher and co-workers presented the use of an iridium amido complex as catalyst for the oxidation of alcohols with benzoquinone (Scheme 1). The authors proposed a mechanism with an intermediate aminyl radical complex as the crucial intermediate, undergoing C–H hydrogen abstraction from the substrate ligand.36

thumbnail image

Figure 4. Representative examples for radical complexes with predominant aminyl character (py = 2-pyridyl).

Download figure to PowerPoint

thumbnail image

Scheme 1. Oxidative dehydrogenation of alcohols catalyzed by an iridium amido complex.

Download figure to PowerPoint

2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information

2.1. Synthesis of Amine Ligands HL1R

Nucleophilic substitution of N,N-bis(2-chloroethyl)amine37 provides versatile access to amine ligands HN(CH2CH2PR2)2 (HL1R, Figure 2) with a wide variety of aromatic and aliphatic substituents at the phosphane moiety. The synthesis of diphenylphosphanyl derivative HL1Ph was reported by reaction of HN(CH2CH2X)2 (X = Cl, Br) with KPPh2 or, more conveniently, from hydrochloride [H2N(CH2CH2Cl)2]Cl with diphenylphosphane and excess KOtBu (Scheme 2).38,39 This route also affords aryl phosphanes HL1R [R = mesityl, p-tolyl, C6H3-3,5-(CF3)2].40 The primary phosphane ligand HL1H is similarly obtained in high yield from [H2N(CH2CH2Cl)2]Cl and PH3 under basic conditions (KOH/DMSO).41 Initial attempts to synthesize dialkylphosphanyl ligands HL1Me and HL1Et from HN(CH2CH2Cl)2 and NaPMe2 in liquid ammonia or from [H2N(CH2CH2Cl)2]Cl and LiPEt2 in THF, respectively, suffered from moderate to poor yields.42 Synthesis of ligands with bulkier dialkylphosphanyl substituents failed by this route. The low yield is attributed to the formation of aziridines (CH2CH2N)CH2CH2PR2 as side products.43 This base-promoted nucleophilic cyclization can be avoided by silyl-protection of the amine. Deprotection of the product is easily achieved by heating to reflux in H2O or treatment with KF or [nBu4N]F in MeOH.43,44,45 This pathway affords a wide range of dialkylphosphanyl ligands HL1R (R = Me, Et, iPr, tBu, 1-adamantyl, cyclohexyl) with good yields (Scheme 2).43,46,47 Particular mention should be made of the possibility to introduce very bulky groups, such as PtBu2 or PAd2, which was not reported for the popular diarylamido pincer ligands (Figure 1, {HLA}).

thumbnail image

Scheme 2. Synthesis of PNP amine ligands HL1R (1-Ad = 1-adamantyl, Cy = cyclohexyl, Mes = mesityl, Tol = p-tolyl).

Download figure to PowerPoint

Chiral, nonracemic derivatives were also reported.Bis{2-[(2S,5S)-2,5-dimethyl-phospholanoethyl]}amine(HL1S,S–Me2Phospholane) can be synthesized by reaction of (2S,5S)-2,5-dimethyl-1-phenylphospholyllithium with Me3SiN(CH2CH2Cl)2 (Scheme 3).45 Another pathway towards chiral phospholane ligands is provided by deprotonation of the N-protected primary phosphane Me3SiN(CH2CH4PH2)2 with nBuLi and subsequent reaction with chiral, cyclic 2S,4S-2,4-pentanediol sulfate or 2S,5S-2,5-hexanediol sulfate (Scheme 3).46a Axially chiral binaphthyl ligand (HL1Binaph) is obtained by P-lithiation of enantiomerically pure 4-chloro-4,5-dihydro-3H-4-phosphacyclohepta[2,1-c;1′,2′-e]phosphepine,48 subsequent reaction with Me3SiN(CH2CH2Cl)2, and N-deprotection (Scheme 3).46a

thumbnail image

Scheme 3. Synthesis of chiral PNP ligands HL1R.

Download figure to PowerPoint

2.2. Square-Planar 16 VE Amine and Amido Complexes: Synthesis, Stability, and N–H Acidity

Reaction of [IrCl(COE)2]2 (COE = cyclooctene) with HL1iPr in 2-propanol affords iridium(III) complex [Ir(H)2Cl{HL1iPr}].46a Use of THF as solvent provides access to a wide range of stable, square-planar d8 amine complexes, [IrL{HL1iPr}]+ (L = CO, COE, C2H4, C3H6, PMe3) (Scheme 4).49 However, exchange of chloride with a weakly coordinating counteranion is crucial. In nonprotic solvents, the chloride anion of [Ir(COE){HL1iPr}]Cl is hydrogen-bonded to the N–H proton, which is evidenced by a strong downfield shift in the 1H NMR spectrum, accompanied by decomposition towards an equimolar mixture of amide [Ir(COE){L1iPr}] and iridium(III) complex [Ir(Cl)2H{HL1iPr}] and minor amounts of cyclooctenyl complex [IrCl(C8H13)H{HL1iPr}] (Scheme 5). Kinetic examinations indicate that the amido complex results from HCl elimination, which oxidatively adds to the initial cyclooctene complex. Intramolecular, vinylic C–H activation and subsequent trapping by chloride is kinetically competitive with this reaction.50 In comparison, [Ir(COE){HL1tBu}]Cl and [IrCl(C8H13)H{HL1tBu}] form a strongly solvent-dependent equilibrium.51 The increased steric bulk favors the vinyl hydride over the olefin isomer, slows down C–H oxidative addition/reductive elimination, and renders chloride trapping of iridium(III) reversible.

thumbnail image

Scheme 4. Synthesis of iridium(I) PNP amine and amido complexes.

Download figure to PowerPoint

thumbnail image

Scheme 5. Decomposition of [Ir(COE){HL1iPr}]Cl in nonprotic solvents.

Download figure to PowerPoint

Deprotonation of [IrL{HL1iPr}] provides access to iridium(I) amido complexes [IrL{L1iPr}] (L = CO, COE, C2H4, PMe3), which are thermally relatively robust towards β-H elimination (Scheme 4).49,52 Likewise, palladium(II) PNP amido complexes [PdX{L1iPr}] (X = Cl, Me, Ph) and [PdL{L1iPr}]PF6 (L = PMe3, CNtBu) were synthesized.53 The large hypsochromic shift (Δν = 68 cm–1, Table 1) of the [Ir(CO){L1iPr}] C–O stretching vibration and the large [Ir(olefin){L1iPr}] olefin 13C NMR spectroscopic high-field shifts (ΔδC2H4 = 22.1 ppm; ΔδCOE = 20.8 ppm) upon deprotonation of the respective amine complexes indicate a very electron-rich Ir{L1R} fragment. In turn, the ligand trans to the nitrogen atom has a profound effect on the amine pKa value, which varies over more than 7 orders of magnitude (in dmso) within the series L = CO [14.9(3)], COE [16.0(4)], PMe3 [22.0(1)].52 In comparison, the value for the olefin complex is close to that of the square-planar cis-diamine-diolefin–RhI complex [Rh{H2LE}]+ (pKadmso = 15.7) but considerably less acidic than the corresponding IrI complex [Ir{H2LE}]+ (pKadmso = 10.5).36,54 However, the pKa trend within the [IrL{HL1iPr}]+ series can be rationalized with a N[RIGHTWARDS ARROW]Ir[RIGHTWARDS ARROW]L push-pull-interaction that stabilizes the N–Ir π* orbital (HOMO) if trans-ligand L is a π-acceptor ligand, such as CO or olefins (Figure 5). The same bonding situation was found for iridium(I) olefin amido complex [Ir(COD){κ2-bpa}] {COD = cyclooctadiene; bpa = N(CH2C5H4N)2}.33b High nucleophilicity of the amido group towards C-nucleophiles is also observed. [Ir(PMe3){L1iPr}] and [PdCl{L1iPr}] are selectively attacked by MeOTf at the nitrogen atom (Scheme 6), and an iridium(III) methyl complex was not observed.52,53

Table 1. Comparison of M(CO){PEP} (E = N, C) CO stretching vibrations. Complexes with ligands derived from HL1R are shown in bold italic typeface.
ComplexνCO [cm–1]Ref.
  • [a]

    Sample preparation: Nujol mull.

  • [b]

    Sample preparation: Neat film.

  • [c]

    Pentane solution.

  • [d]

    THF solution.

  • [e]

    Cyclooctane solution.

  • [f]

    Hexane solution.

  • [g]

    KBr pellet.

  • [h]

    CH2Cl2 solution.

[Ir(CO){HL1iPr}]PF61976[a]49
[Ir(CO){C5H3N(CH2PtBu2)2}]PF61962[b]61
[Ir(CO){C6H3(OPtBu2)2}]1949[c]62
[Ir(CO)Cl(PiPr3)2]1939[a]63
[Ir(CO)(L4tBu)]1937[a]69
[Ir(CO){NC5H3(CHPtBu2)(CH2PtBu2)}]1932[b]64
[Ir(CO){LCiPr}]1930[d]7c
[Ir(CO){LAiPr}]1930[d]65
[Ir(CO){C6H3(CH2PiPr2)2}]1928[e]3
[Ir(CO){CpFeC5H2(CH2PiPr2)2}]1926[f]66
[Ir(CO){C7H4(CHPtBu2)(CH2PtBu2)}]1919[g]67
[Ir(CO)(L1iPr)]1908[a]49
[Ir(CO){C7H5(CH2PtBu2)2}]1905[g]68
[RuCl(CO)(L4tBu)]1916[a]69
[RuCl(CO){CpFeC5H2(CH2PtBu2)2}]1914[h]70
[RuCl(CO){C6H3(CH2PtBu2)2}]1909[a]71
[RuCl(CO)(L3tBu)]1896[a]69
[RuCl(CO)(L1tBu)]1888[a]69
[RuCl(CO){CH(CH2CH2PtBu2)2}]1887[a]72
thumbnail image

Figure 5. Molecular structure of [Ir(CO){L1iPr}] from X-ray diffraction (left), DFT-computed HOMO (center), and qualitative representation of the N–Ir–CO orbital π-interaction (right).

Download figure to PowerPoint

thumbnail image

Scheme 6. Reactions of amido and enamido complexes with MeOTf.

Download figure to PowerPoint

2.3. Functionalization of the Pincer Backbone: Fine-Tuning N[RIGHTWARDS ARROW]M Donation

2.3.1. Amido, Enamido, and Dienamido Pincer Ligands: Comparison of Donor Properties

Dehydrogenation of the chelate backbone bridges towards enamido and dienamido ligands {L3R} and {L4R} (Figure 2) provides another approach to fine-tune electronic properties and reactivity. These enamido ligands are related to chelating (2-phosphanyl)vinylalcoholates, which are used in the Shell Higher Olefin Process (SHOP).55 In analogy, Braunstein and co-workers prepared the enamido nickel(II) complexes [Ni(Ph)L{N(Ph)CPhCHPPh2}] (L = PMe3, PMe2Ph, PMePh2) by P–Ph oxidative addition of an α-iminophosphorus ylide to [Ni(COD)2] (Scheme 7).56 However, deprotonation of coordinated or free imine ligands at the C–H-acidic α-position represents a more general approach to phosphanylenamido ligands.57,58 For example, Morris recently reported the synthesis of dienamido PNNP iron(II) complexes by deprotonation of the corresponding imines (Scheme 7).59 Both catalyze transfer hydrogenation of ketones with comparable rates, suggesting that ligand backbone proton transfer reactions might be involved in a metal–ligand cooperative hydrogenation mechanism. Deprotonation of bis(2-picolyl)amido ligands coordinated to rhodium(I) and iridium(I)affords “dearomatized” vinylenediamido complexes [M(COD){RNC(H)C5H4N}] (M = Rh, Ir; COD = cyclooctadiene; R = CH2C5H4N) (Scheme 7).33b,60

thumbnail image

Scheme 7. Syntheses of chelating enamido complexes.

Download figure to PowerPoint

Synthetic access to free ligands {L3R} and {L4R} has not been reported yet. Their metal complexes are obtained by starting from coordinated amine, imine, or amido precursors and following several routes, which are discussed in the subsequent sections. However, comparison of parent amido complexes carrying {L1R} with related enamido and dienamido complexes formally derived from it ({L3R} and {L4R}) and with complexes of other popular PEP (E = N, C) pincer ligands is informative at this point to understand reactivity trends. CO stretching vibrations (νCO) can serve as a probe in trying to scale the electronic effects of backbone dehydrogenation. Table 1 shows representative examples of square-planar [IrI(CO){PEP}] and five-coordinate [RuIICl(CO){PEP}] complexes (E = N, C), all carrying bulky dialkylphosphane (PiPr2 or PtBu2) substituents but different pincer backbones.73 Amine and amido complexes [Ir(CO){HL1iPr}]+ (νCO = 1976 cm–1) and [Ir(CO){L1iPr}] (νCO = 1908 cm–1) represent the upper and lower ends of a wide range of CO stretching vibrations found for cationic and neutral [IrI(CO){PEP}]+/0 pincer complexes. On the other hand, fully dehydrogenated pincer complex [Ir(CO){L4tBu}] (Scheme 11) exhibits an intermediate νCO (1937 cm–1).69 This observation suggests reduced N[RIGHTWARDS ARROW]Ir electron donation upon oxidative ligand backbone functionalization, owing to stabilization of the N-lone-pair by C(p)–N(p) overlap, that is, the vinyl substituents acting as π-acceptors. These qualitative considerations are further reflected within the series [RuCl(CO){L1tBu}] (νCO = 1888 cm–1), [RuCl(CO){L3tBu}] (νCO = 1896 cm–1), and [RuCl(CO){L4tBu}] (νCO = 1916 cm–1), which can be prepared by reaction of CO with the respective four-coordinate complexes.69 As for the square-planar IrI platform, the five-coordinate RuII complexes span a range of about 30 cm–1, indicating comparable electronic effects for the different metals and coordination geometries upon backbone dehydrogenation. [RuCl(CO){L1tBu}] and [RuCl(CO){CH(CH2CH2PtBu2)2}] exhibit almost identical νCO. Hence, weaker E[RIGHTWARDS ARROW]Ru (E = N, C) σ-donation of the amido relative to the alkyl ligand is offset by N[RIGHTWARDS ARROW]Ru π-donation. Other frequently used anionic pincer ligands, such as the aryl pincer {C6H3(CH2PiPr2)2}, diarylamido pincer {LA}, disilylamido pincer {LC}, or pyridine-based “dearomatized” pincer {LB}, fall within a relatively narrow range between {L3} and {L4} with respect to their donor properties, as judged by their CO stretching vibrations.

2.3.2. 16 VE Amido Versus Enamido Complexes: Consequences for Structure and Reactivity

The molecular structure and reactivity of [RuH(PMe3){L1iPr}], as compared with [RuH(PMe3){L3iPr}], provides further information. The enamido complex is quantitatively obtained from [RuCl2(PMe3){HL1iPr}] with excess KOtBu (Scheme 8).74 Decomposition of the tentative amido intermediate [RuCl(PMe3){L1iPr}] by β-hydride migration and backbone deprotonation of the resulting imine provides a reasonable mechanism.75 [RuH(PMe3){L3iPr}] adds two equivalents of H2 to give amine complex [Ru(H)2(PMe3){HL1iPr}] (Scheme 15). This reaction is fully reversible, but elimination of the first equivalent is considerably faster (see below), rendering amido complex [RuH(PMe3){L1iPr}] easily isolable.

thumbnail image

Scheme 8. Proposed mechanism of synthesis of enamido complex [Ru(H)PMe3{L3iPr}].

Download figure to PowerPoint

The molecular structures of both [RuH(PMe3){L1iPr}] (X-ray) and [RuH(PMe3){L3iPr}] (DFT) can be described by starting from trigonal-bipyramidal (TBPY) geometry, with N, H, and PMe3 in equatorial positions, but with different modes of distortion from ideal TBPY.74 For [RuH(PMe3){L1iPr}], the strongly compressed H–Ru–PMe3 angle [76.0(7)°] and widened H–Ru–N [124.8(7)°] and N–Ru–PMe3 [159.12(3)°] angles indicate a “Y-shaped” distortion from TBPY coordination. This geometry is typically found for five-coordinate d6 complexes with a strong π-donor ligand to avoid repulsive filled–filled N(p)–M(d) π-interactions and instead maintain stabilizing N[RIGHTWARDS ARROW]M π-bonding with the coordinatively unsaturated metal center (Figure 6).76,77 In contrast, [RuH(PMe3){L3iPr}] (H–Ru–PMe3: 84.6°; H–Ru–N: 102.5°; N–Ru–PMe3: 172.6°) exhibits stronger resemblance of “T-shaped” distortion from TBPY [i.e. square-pyramidal (SQPY) geometry], suggesting again attenuated π-donation by the enamido nitrogen atom. This interpretation is further supported by comparison of the Ru–N Wiberg bond indices for [RuH(PMe3){L1iPr}] (0.50) and [RuH(PMe3){L3iPr}] (0.37), indicating a lower degree of covalent bonding for the latter.

thumbnail image

Figure 6. Qualitative valence d-orbital splitting in a TBPY (D3h) ligand field (center) with d6 occupation (dmath image LUMO omitted) and lifted degeneracy of the e′-orbitals upon Y-shaped (left) and T-shaped (right) distortion, respectively. The respective LUMOs indicate the suitability in the Y-TBPY case for π-bonding with a π-donor ligand (axial ligands omitted for clarity).

Download figure to PowerPoint

Hence, the ruthenium center in the enamido complex exhibits a higher degree of electronic unsaturation, whichis reflected in its reactivity with nucleophiles. [RuH(PMe3){L3iPr}] shows Lewis acid reactivity, typical for SQPY RuII complexes, exemplified by the formation of octahedral [RuH(PMe3)2{L3iPr}] with PMe3 (Scheme 9).75 In contrast, [RuH(PMe3){L1iPr}] does not form an isolable six-coordinate amido complex. Upon addition of PMe3, broadening of all signals in the 31P NMR spectrum indicates rapid ligand exchange. Furthermore, slow ligand disproportionation towards amine [Ru(H)2(PMe3){HL1iPr}] and enamide [RuH(PMe3)2{L3iPr}] is observed, which is attributed to the mechanism in Scheme 9. The instability of proposed imine intermediate [Ru(H)2(PMe3){L2iPr}] towards H2 elimination was confirmed by DFT computations (Scheme 15). The reactivity towards electrophiles was also examined. With MeOTf, [Ru(H)(PMe3){L1iPr}] forms the expected tertiary amine complex [Ru(H)OTf(PMe3){MeL1iPr}] selectively. [Ru(H)(PMe3){L3iPr}] exhibits reactivity with both nucleophilic positions on the ligand backbone, resulting in a mixture of enamine [Ru(H)OTf(PMe3){MeL3iPr}] and the products from backbone C-methylation to a minor extent (Scheme 6).

3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information

3.1. Oxidation of 16 VE Dialkylamido Complexes: Formation of Unstable Radicals

The d8 amido complexes [IrL{L1iPr}] (L = CO, COE, PMe3) and [PdCl{L1iPr}] in THF exhibit irreversible oxidation waves in the cyclic voltammograms (THF) at scan speeds up to 1 V s–1, suggesting rapid chemical degradation of initially formed radical cations on the experimental timescale.52 As for the pKa values of the conjugate acids (see above), the oxidation potentials of the series [IrL{L1iPr}] are highly dependent on the ligand in the trans-position to the nitrogen atom (L = CO: –0.39 V, COE: –0.49 V, PMe3: –1.09 V vs. Fc/Fc+). Chemical oxidation of [IrPMe3{L1iPr}] and [PdPh{L1iPr}] with AgPF6 gives amine complexes [IrPMe3{HL1iPr}]+ or [PdPh{HL1iPr}]+ in around 90 % or 80 % yield, respectively, and a minor amount of the corresponding imines [IrPMe3{L2iPr}]PF6 and [PdPh{L2iPr}]PF6.52,53 Deuterium labeling in case of Pd supports initial formation of radical cation [PdPh{L1iPr}]+·, which undergoes competitive hydrogen abstraction from the solvent (THF) and ligand backbone disproportionation (Scheme 10). Hence, the reactivity indicates “redox non-innocent” behavior of the dialkylamido ligand. Accordingly, DFT computations (B3LYP/6-311+G**) for [IrCO{L1iPr}]+ indicate about 63 % of the Mullikan spin density to be located on the nitrogen atom, and the SOMO of [IrCO{L1iPr}]+ (Figure 7) represents the N(p)–Ir(dxz) π*-orbital resulting from removal of an electron from the HOMO of parent [IrCO{L1iPr}] (Figure 5).

thumbnail image

Scheme 9. Reactions of enamide [RuH(PMe3){L3iPr}] (top) and amide [RuH(PMe3){L1iPr}] (bottom) with PMe3 and proposed mechanism for the latter.

Download figure to PowerPoint

thumbnail image

Scheme 10. Chemical oxidation of amido complex [PdPh{L1iPr}] and proposed mechanism via aminyl radical complex [PdPh{L1iPr}]+·.

Download figure to PowerPoint

thumbnail image

Figure 7. DFT-computed molecular structure of [Ir(CO)(L1iPr)]+· (left) and spin density (right).

Download figure to PowerPoint

3.2. Ligand Modification: Stabilization of Square-Planar 15 VE Complexes

The decomposition mechanism of PNP dialkylamido iridium and palladium radical complexes suggests a simple synthetic strategy for the isolation of persistent open-shell complexes.78 Replacement of the backbone ethylene with vinylene bridges (Figure 2), prevents putative aminoyl ligand disproportionation towards amine and imine complexes. In fact, in contrast to transient [IrPMe3{L1iPr}]+· (see above), disilylamido complex [IrCl{LCtBu}] was isolated and characterized by EPR spectroscopy.79 Furthermore, this modification should also stabilize a 15 VE radical complex electronically by delocalization of the N-lone-pair. Hence, the SOMO of enamido and dienamido radical complexes should exhibit less N(p) character as compared with an analogous dialkylamido radical complex, providing further stabilization with respect to possible ligand-centered radical decomposition reactions.80 Oxidation of amine [IrH(C8H13)Cl{HL1tBu}] with benzoquinone gives dienamido complex [IrCl{L4tBu}] in good yield (Scheme 11).81 As for [IrCl{LCtBu}], the large anisotropy of the g values is in agreement with predominant d7 metalloradical character. DFT computations confirmed 67 % spin density to be located at the Ir metal. Isolable, monomeric d7 complexes of iridium are relatively rare compared with cobalt (and to some extend rhodium).79,82

thumbnail image

Scheme 11. Synthesis of iridium(I), iridium(II), and iridium(III) dienamido complexes.

Download figure to PowerPoint

3.3. Fine-Tuning of N[RIGHTWARDS ARROW]M Donation: Low-Spin Versus Intermediate-Spin Square-Planar 14 VE Complexes

Ligand field considerations for square-planar complexes (D4h) with strong-field ligands predict splitting of the d-orbital manifold into three nonbonding orbitals (dxy, dxz and dyz), weakly antibonding dmath image orbital (stabilized by valence s-orbital mixing), and a high-lying, strongly antibonding dmath image orbital. Hence, for d6 ions, an intermediate-spin (S = 1) electronic configuration is expected (Figure 8A).83 Iron(II) porphyrinato and bis(dithiolato) complexes provide well-examined examples.84,85 In contrast, four-coordinate 14 VE complexes of the platinum metals are usually found in a sawhorse (cis-divacant octahedral) conformation with electronic low-spin (S = 0) configuration and stabilizing C–H agostic interactions at the vacant coordination sites.86 DFT computations also suggested an intermediate-spin electronic ground state for square-planar alkyl PCP pincer complex [RuCl{HC(CH2NHPtBu2)2}].87 This species was postulated as a transient intermediate in the equilibrium formed by the closed-shell α- and β-hydrogen elimination products (Scheme 12). The high reaction rate of hydrido alkylidene and olefin isomer interconversion was attributed to two-state reactivity88 via triplet intermediate [RuCl{HC(CH2NHPtBu2)2}].

thumbnail image

Figure 8. Valence d-orbital splitting in square-planar coordination for a d6 ion with pure σ-donor ligands (A) and one weak (B) or strong (C) single-faced, perpendicular π-donor, respectively.

Download figure to PowerPoint

thumbnail image

Scheme 12. Equilibrium of [Ru(H)Cl{PtBu2NHCH=CHCH2NHPtBu2}] and [Ru(H)Cl{C(CH2NHPtBu2)2}] attributed to two-state reactivity via triplet intermediate [RuCl{CH(CH2NHPtBu2)2}].

Download figure to PowerPoint

The introduction of one perpendicular single-faced π-donor, such as an amido ligand, further perturbs the valence orbital splitting by lifting the dxz/dyz degeneracy. If the π-donor is located along the x-axis, the dxz orbital will be considerably destabilized, subject to the extent of N[RIGHTWARDS ARROW]M π-donation (Figure 8, B/C). Hence, two electronic configurations can arise: An intermediate-spin (S = 1) state having a singly occupied N–M π*-orbital [(xy,yz,z2)5(xz)1(x2y2)0; Figure 8, B] or a low-spin configuration (S = 0) in which this orbital is vacant [(xy,yz,z2)6(xz)0(x2y2)0, Figure 8, C]. Therefore, N[RIGHTWARDS ARROW]M π-bonding should directly influence the singlet–triplet gap, eventually favoring a low-spin over intermediate-spin ground state if the spin pairing energy is overcompensated.

3.3.1. [IrCl(L4tBu)]+ and [RuCl(L1tBu)]: Square-Planar d6 Low-Spin Complexes

Electrochemical characterization of [IrCl{L4tBu}] indicates reversible oxidation at E1/2 = +0.02 V (vs. FeCp2/FeCp2+).81 Accordingly, diamagnetic d6 complex [IrCl{L4tBu}]PF6 can be synthesized upon chemical oxidation with AgPF6, as a unique example for a square-planar IrIII complex (Scheme 11). DFT calculations confirm that the singlet state is more stable by more than 4 kcal mol–1 (ZORA-B3LYP/TZVP), which is attributable to strong N[RIGHTWARDS ARROW]Ir π-donation. This complex is isoelectronic with Shaw's “classic” carbene complex [IrCl{=C(CH2CH2PtBu)2}].89 Square-planar RuII complex [RuCl{L1tBu}] is obtained by HCl elimination from [RuCl2{HL1tBu}] (Scheme 13).90 Like [IrCl{L4tBu}]+, [RuCl{L1tBu}] is diamagnetic, as evidenced by sharp NMR spectroscopic signals. DFT computations suggested a small singlet–triplet gap. The short Ru–N bond derived by single-crystal X-ray diffraction [1.890(2) Å] is in agreement with the DFT-optimized structure for the singlet state (1.90 Å) and considerably shorter than that computed for the triplet state (1.99 Å). This structural feature is particularly indicative for the spin state, since the N–M π*-orbital is vacant in the singlet state (Figure 8, C) but singly occupied in the triplet state (B). Furthermore, the NMR spectra (1H and 31P) exhibit some unusual chemical shifts with strong temperature dependence. Population of an energetically low-lying triplet state was proposed as explanation, and the thermodynamic parameters were estimated by fitting of the NMR spectroscopic data (ΔH = 10.6 ± 0.3 kJ mol–1; ΔS = 4.2 ± 0.8 J mol–1 K–1).90

3.3.2. [RuCl{LCtBu}], [RuCl{L3tBu}], and [RuCl{L4tBu}]: In Favor of Intermediate Spin

In contrast, Caulton's disilylamido complex [RuCl{LCtBu}] exhibits strongly paramagnetically shifted signals in the 1H NMR spectrum, and no 31P NMR signal was detected.91 At 298 K, the magnetic susceptibility is in agreement with two unpaired electrons (χMT = 1.03), and the triplet state was calculated to be favored by 10 kcal mol–1.92 However, at low temperatures, χMT drops to near zero, indicating a nonmagnetic ground state.91 These results were rationalized with a spin triplet ground state which is further split into the nonmagnetic | S = 1, MS = 0 [rang] microstate placed below | S = 1, MS = ±1 [rang] owing to zero-field-splitting (ZFS).93 The magnetic data fitting gave a large axial ZFS parameter (D = +273 cm–1). The intermediate-spin ground state, also observed for [RuX{LCtBu}] (X = F, OTf), [OsI{LCtBu}], and hydrolysis product [Ru(OSiMe2CH2PtBu2)2], are highly unusual for second- and third-row transition metal complexes.92,94 The magnetic properties are reminiscent of those of RuII/RuII paddlewheel complexes [Ru2(O2CR)4] (R = CH3, Ph) which exhibit two unpaired electrons but are essentially nonmagnetic at low temperatures as a result of a σ2π4δ2δ*2π*2 electronic triplet configuration of the Ru24+ core and a large ZFS.95

thumbnail image

Scheme 13. Synthesis of square-planar RuII amido, enamido, and dienamido complexes.

Download figure to PowerPoint

Direct comparison of [RuCl{L1tBu}] and [RuCl{LCtBu}] suggests that the different electronic structure might be attributed to stronger N[RIGHTWARDS ARROW]M π-donation by the dialkyl with respect to the disilylamido ligand. In terms of donor properties, enamido and dienamido ligands scale closer to disilylamides (see above). Therefore, the corresponding square-planar RuII complexes [RuCl{L3tBu}] and [RuCl{L4tBu}] were prepared to probe this qualitative picture (Scheme 13).69 Both complexes exhibit two unpaired electrons at room temperature, as judged by SQUID magnetometry and NMR spectroscopic data. In analogy to [RuCl{LCtBu}], χMT drops to near zero at low temperatures. Fitting of the magnetic data to a model with S = 1 gave large axial ZFS parameters at around +380 cm–1 ([RuCl{L3tBu}]) and +230 cm–1 ([RuCl{L4tBu}]), as found for [RuCl{LCtBu}] (D = +273 cm–1). To exclude a gradual spin transition as an alternative explanation for the magnetic data, the molecular structure of [RuCl{L4tBu}] was derived by single-crystal X-ray diffraction over a wide temperature range (30–200 K). Both the Ru–N distance [1.994(2) Å] and the N–Ru–Cl angle (180.0°) are invariant. DFT computations confirmed that both structural parameters should be sensitive indicators for the spin multiplicity (ΔRu–N = 0.07 Å; ΔN–Ru–Cl = 15.2°).

Hence, the bulky ligands {L1tBu}, {L3tBu}, and {L4tBu} enable the isolation of square-planar 15 and 14 VE IrII, IrIII, and RuII complexes. In case of RuII, the ligand variation allows for control of the electronic structure, attributed to modified N[RIGHTWARDS ARROW]M π-donation. As pointed out by Caulton and co-workers, other factors such as pincer ligand conformational variations, certainly affect M–N π-bonding, as well, and therefore the generally small singlet–triplet gap for this class of compounds.92 As for [IrCl{L4tBu}]+, the supposedly weakest π-donor within the series {L1R}/{L3R}/{L4R} is able to cause spin pairing, which is probably an effect of the charge and generally stronger d-orbital splitting compared with that of RuII compounds.

3.4. M(PNP) Complexes as Platforms for Late Transition Metal Nitrides

The coordinatively and electronically highly unsaturated 14 VE complexes should be suitable platforms for the stabilization of multiply bonded ligands. Accordingly, diamagnetic RuIV nitrides [RuN{LCtBu}] and [RuN{L1tBu}] are isolated in high yield from the corresponding chloride complexes by salt metathesis with azides and N2 elimination (Scheme 14).96,97 Terminal nitrido complexes with dn-electron counts n > 2 are exceedingly rare.98 At variance with more common RuVI nitrides,99 [RuN{LCtBu}] exhibits a nucleophilic nitrido ligand, exemplified by methylation with MeOTf and lack of reactivity with phosphanes.100 However, while [RuN{LCtBu}] does not react with CO, selective nitride–CO coupling was observed for [RuN{L1tBu}] (Scheme 14).97 Nitride hydrogenolysis with H2 was also examined and will be discussed in section 4.3. The mechanism of N–CO coupling was not reported, but [RuN{LCtBu}] does not react with the π-acid CO,101 possibly because of the weaker nucleophilicity of its nitride ligand. As a consequence of the mutual trans-configuration of two strong π-donors (nitride and amide), both nitrido complexes exhibit molecular structures with considerable deviation from planarity by N–Ru–N angle bending. Weaker Ru≡N bonding in case of [RuN{L1tBu}] is indicated by the comparatively low stretching vibration (νRu≡N = 976 cm–1; [RuN{LCtBu}]: 1030 cm–1).102

thumbnail image

Scheme 14. Synthesis of square-planar ruthenium(IV) nitrido complexes and their reactivity with CO.

Download figure to PowerPoint

4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information

4.1. Cooperative H2 Activation

The twofold H2 elimination/addition equilibrium of amine, amido, and enamido complexes [Ru(H)2PMe3{HL1iPr}], [Ru(H)PMe3{L1iPr}], and [Ru(H)PMe3{L3iPr}] was investigated by kinetic examinations and DFT computations for the PMe2 truncated model (Scheme 15).75 H2 elimination from amine complex [Ru(H)2PMe3{HL1iPr}] (Scheme 15, left branch) proceeds by proton transfer from the amine group to the syn-coplanar hydride ligand to give a dihydrogen complex (omitted in Scheme 15) with a moderate barrier (ΔG = 18.3 kcal mol–1 for the PMe2 model), which loses H2 in almost barrierless fashion. The reverse, H2 heterolysis by amido complexes, was well examined experimentally and computationally for several related systems in the context of hydrogenase model reactivity and metal–ligand cooperative hydrogenation of ketones.9,103

thumbnail image

Scheme 15. Center: H2 addition/elimination equilibrium of [Ru(H)2PMe3{HL1iPr}], [Ru(H)PMe3{L1iPr}], and [Ru(H)PMe3{L3iPr}] and proposed intermediate [Ru(H)2PMe3{L2iPr}] (H2 complex intermediates omitted). Top: Computed energies for the PMe2 truncated model (kcal mol–1). Bottom: Analogies with catalysts of the Noyori (left) and Milstein (right) type.

Download figure to PowerPoint

On the right branch of the equilibrium reaction, that is, elimination of H2 from the ligand backbone, considerably higher barriers were obtained from computations and H/D exchange kinetics (Scheme 15).75 The putative imine intermediate [Ru(H)2PMe3{L2iPr}] was not detected spectroscopically. Accordingly, the calculations predict that it is unstable towards both H2 elimination and migrative imine insertion. Dihydrogen complex [Ru(H)2(H)PMe3{L3iPr}] (omitted in Scheme 15) was located as an intermediate on the H2-elimination path, losing H2 in almost barrierless fashion. The computations indicate that C–H proton transfer from the ligand backbone to the hydride ligand is kinetically feasible without dissociation of one of the pincer “arms”. However, predissociation might considerably lower the barrier for amide β-H elimination, which could account for the more rapid disproportionation of [Ru(H)PMe3{L1iPr}] in the presence of PMe3 (see above). The equilibrium LnM(H){L2R} [lrarr2] LnM{L3R} + H2 provides an aliphatic analogue of Milstein's pyridine- and acridine-based PNP pincer catalysts, which operate upon reversible ligand backbone aromatization/dearomatization (Scheme 15).6

Several studies for related “Noyori–Morris type” metal–ligand cooperative catalysts suggest that Brønsted acids catalyze the H2 heterolysis step.104 For [Ru(H)2PMe3{HL1iPr}], proton exchange rates with H2O were studied by 1H-EXSY NMR spectroscopy.105 The examinations indicate stereoselective H+/H exchange of water with the hydride ligand that is syn-coplanar with the N–H proton. This observation is in line with a water-catalyzed proton shuttle mechanism for H2 heterolysis upon hydrogen bonding with the secondary amine (Scheme 16). Accordingly, tertiary amine complex [Ru(H)2PMe3{MeL1iPr}] shows considerably slower and unselective H+/H exchange of both hydride ligands with H2O. DFT computations confirm the proposed mechanism, indicating that water catalysis reduces the kinetic barrier for H2 heterolysis by around 8 kcal mol–1 for the PMe2 truncated model (Scheme 16). Similarly, computations for proton transfer (“dearomatization”) from the ligand backbone to a hydride ligand in [Ir(H)2Ph{NC5H3(CH2PMe2)2}] indicate water catalysis, as well.106

thumbnail image

Scheme 16. DFT-computed water-catalyzed H2 heterolysis.

Download figure to PowerPoint

4.2. Outer-Sphere Hydrogenation and Dehydrocoupling of Organic Substrates

Early studies reported on olefin hydrogenation with [RhCl{HL1Ph}] and [IrCl{HL1Ph}] as precatalysts.107,108 However, the observation of the “N–H effect” by Noyori and co-workers sparked interest in these ligands for metal–ligand cooperative (“bifunctional”) catalysis.9 As a rare example, iridium(III) complex [Ir(H)2Cl{HL1iPr}] catalyzes both the hydrogenation of a wide range of aldehydes and ketones and the transfer hydrogenation of ketones andimines at mild reaction conditions upon activation with base (Scheme 17).46b,109 Chemoselective direct hydrogenation of the carbonyl group is observed for benzylideneacetone, while transfer hydrogenation results in the saturated alcohol.

thumbnail image

Scheme 17. Transfer hydrogenation and direct hydrogenation of carbonyl and carboxyl groups catalyzed by HL1R complexes.

Download figure to PowerPoint

Amine and amido complexes [Ir(H)3{HL1iPr}] and [Ir(H)2{L1iPr}] catalyze the reaction with comparable rates without base activation. Therefore, a metal–ligand cooperative “Noyori–Morris type” mechanism, with concerted H+/H transfer to the substrate was proposed for hydrogenation and transfer hydrogenation (Scheme 18), and DFT computations confirmed this suggestion.110 Ruthenium and osmium catalysts with ligand HL1R (R = iPr, Ph) were also reported as excellent bifunctional catalysts for transfer hydrogenation of ketones111 and for direct hydrogenation of esters to alcohols (Scheme 17).40 α-Hydroxy esters and β-boc-amino esters are hydrogenated with retention of the configuration. Despite the availability of chiral PNP ligands, asymmetric hydrogenation has not been reported yet.

thumbnail image

Scheme 18. Proposed catalytic cycles for transfer hydrogenation (inner cycle) and direct hydrogenation (outer cycle) of ketones with [Ir(H)3{HL1iPr}].

Download figure to PowerPoint

Microscopic reversibility suggests that these catalysts could be suitable for acceptorless dehydrocoupling and reactions using the “borrowing hydrogen” methodology, as well.112 Beller and co-workers reported the highly efficient (TON > 40000; TOFmax > 14000 h–1) release of H2 and acetone from neat 2-propanol at very mild conditions (90 °C) with [RuH2(CO)(PPh3)3]/HL1iPr (4.0 ppm loading) in the absence of base. Interestingly, the addition of base (e.g. NaOiPr) results in considerably smaller TOF and the dehydrogenation of the PNP-ligand backbone in the presence of base (see above) provides a possible explanation.113 Ruthenium and osmium complexes with ligand HL1iPr were also used for dehydrocoupling of primary alcohols to symmetric esters and for amine alkylation with alcohols, and they exhibited high turnover numbers (Scheme 19).111 The latter reaction was reported for [Ir(H)2Cl{HL1iPr}]/base as precatalyst, as well.114

thumbnail image

Scheme 19. Dehydrogenative coupling reactions with ruthenium and osmium PNP complexes.

Download figure to PowerPoint

4.3. De-/Hydrogenation of Inorganic Substrates: CO2, Azide, and Borane-Amines

The d6 complexes [Re(H)2NO{HL1iPr}] and [Ir(H)3{HL1iPr}] rapidly insert CO2 into a metal–hydride bond to form the formate complexes [Re(O2CH)NO(H){HL1iPr}] and [Ir(O2CH)(H)2{HL1iPr}], respectively.115,116 The latter catalyzes the hydrogenation of CO2 to formate in basic, aqueous solution with high activity (Scheme 20). In analogy to “Noyori–Morris type” ketone hydrogenation, an outer-sphere hydrogen transfer mechanism to the substrate was proposed and was supported by DFT computations.

thumbnail image

Scheme 20. Hydrogenation of CO2 in aqueous solution catalyzed by [Ir(O2CH)(H)2{HL1iPr}].

Download figure to PowerPoint

Nitrido complexes could serve as model compounds to examine ammonia formation en route to a highly desirable homogeneous catalyst for N2 hydrogenation at ambient conditions. However, only a few examples were reported that show reactivity towards H2.117 Otherwise, terminal nitrides are surprisingly inert towards H2,118 including [RuN{LCtBu}].100 In contrast, nitrido complex [RuN{L1tBu}] reacts with H2 at relatively mild conditions (1 bar, 50 °C).97 Unprecedented formation of ammonia and polyhydride [RuH4{HL1tBu}] is observed in high yield (Scheme 21). Preliminary mechanistic examination is in agreement with initial, rate-determining heterolytic hydrogen splitting across the Ru–amido bond followed by stepwise proton transfer reactions to the nitride via imido, amido, and ammine intermediates. Hence, the role of the PNP pincer ligand is twofold: weakening Ru–N bondingto the trans-nitrido ligand (see above) and acceleration of the reaction by metal–ligand cooperative H2 activation. Recycling of the ruthenium product by protonolysis followed by reaction with azide and base with high yields was also demonstrated, giving an overall synthetic cycle for catalytic hydrogenation of azide to NH3 (Scheme 21).

thumbnail image

Scheme 21. Synthetic cycle for the hydrogenation of azide to ammonia.

Download figure to PowerPoint

In recent years, the dehydrogenation of borane–amine adducts has attracted considerable interest for chemical hydrogen storage and the synthesis of novel inorganic polymers and B–N ceramics.119,120 Both [Ru(H)PMe3{L1iPr}] and [Ru(H)2PMe3{HL1iPr}] catalyze the release of H2 (1 equiv.) from H3N–BH3 (AB) solutions with high activities (TOF ≈ 20 s–1) and turnover numbers (TON > 3000) without base at room temperature (Scheme 22).74 According to 11B-MQ-MAS spectra of the AB dehydrocoupling product, linear polyaminoboranes H3N–[H2B–H2N]n–BH3 are formed.121 Dehydrocoupling of MeH2N–BH3 towards linear, high-molecular-mass poly-N-methylaminoborane MeH2N–[H2B–MeHN]n–BH3 and of Me2HN–BH3 towards aminoborane dimer [Me2N–BH2]2 were reported, as well.122,123 The related catalyst [RuCl2(H2NCH2CH2PtBu2)2] (with 10 equiv. KOtBu) was also used for dehydrogenation of borane–amine adducts.124,125 On the basis of DFT computations, Fagnou proposed for this catalyst a metal–ligand cooperative (“Noyori–Morris type”) mechanism with concerted H+/H transfer from the substrate to a catalyst amido species and subsequent, turnover-limiting H2 elimination from the amine complex. In fact, for H3N–BH3 dehydrocoupling catalyzed by [Ru(H)PMe3{L1iPr}], high kinetic isotope effects upon deuteriation of the substrate both at the N- and at the B-termini were reported, as expected for such a mechanism.74 However, the experimental rate law is first-order in substrate, indicating a different rate-determining step for this catalyst. Iridium(III) complexes [Ir(Cl)2H{HL1tBu}] and [Ir(C8H13)HCl{HL1tBu}] were also reported to be efficient catalysts for the hydrolysis of ammonia borane in iPrOH/H2O solution, mediating the rapid release of almost 3 equiv. H2 (ca. 9 wt.-%).126

thumbnail image

Scheme 22. Dehydrocoupling of borane–amine adducts catalyzed by [Ru(H)PMe3{HL1iPr}] and [Ru(H)2PMe3{HL1iPr}].

Download figure to PowerPoint

Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information

The present review emphasizes the remarkable versatility of the chelating amido, imine, enamido, and dienamido ligands derived from ethylene-bridged amine HN(CH2CH2PR3)2 by deprotonation or oxidative backbone functionalization. The parent amine, arguably one of the simplest PEP (E = C, N) pincer ligand platforms, is synthetically easily accessible with a wide range of available substituents including very bulky and some chiral examples. Most importantly, backbone dehydrogenation allows for electronic fine-tuning of the ligand donor properties, giving rise to a relatively wide range as compared with many other popular pincer ligand platforms.

The strong π-basicity of the dialkylamide {L1R} accounts for the main properties, such as ligand centered nucleophilicity. Hence, these ligands are ideal building blocks for metal–ligand cooperative (Noyori–Morris type) stoichiometric bond activation reactions and catalysis, exemplified by several efficient hydrogenation and dehydrocoupling reactions of organic and inorganic substrates. In addition to the bifunctional reactivity of the amido moiety, the easily obtained enamido ligand {L3R} represents an aliphatic analogue of another ligand class, Milstein's “dearomatized” PNP ligands, which is successfully applied in hydrogenation and dehydrocoupling catalysis. In addition to nitrogen-centered basic reactivity, the π-basicity also affects the electronic structure of coordinatively and electronically unsaturated complexes. Strong N(p)[RIGHTWARDS ARROW]M(d) π-donation by the dialkylamido ligand results in considerable nitrogen contribution to the HOMO of square-planar 16 VE complexes. Consequently, 1-electron oxidation of such iridium(I) and palladium(II) amides gives unstable radical complexes with predominant aminyl character. In turn, d6 dialkylamides in this geometry are strongly stabilized by N[RIGHTWARDS ARROW]M π-donation. The d-orbital needed for π-bonding is vacant in the electronic low-spin configuration found for [RuCl{L1tBu}] and [IrCl{L4tBu }]+, which can therefore be described as 16 VE complexes. In case of ruthenium(II), the use of weaker π-donating ligands, {LC}, {L3}, or {L4}, enables the control of the ground state spin multiplicity. Hence, it will be an interesting goal for future developments to assess whether the relatively unusual and to some extent adjustable electronic structure of these compounds will translate into unusual reactivity.

Biographical Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information
Thumbnail image of

Sven Schneider studied chemistry at the Technische Universität Darmstadt (Diploma 1999 with Prof. H.-J. Klein) and received a Ph.D. from Humboldt Universität Berlin (2003) working on group 6 amido complexes under the supervision of Prof. A. C. Filippou. He subsequently joined the group of Prof. T. J. Marks at Northwestern University for postdoctoral studies dedicated to single-source precursors for copper chalcogenide materials. In 2006, he started his independent career at the Technische Universität München as an Emmy-Noether group leader, and in 2010 he received his current position as W2 (associate) professor for inorganic chemistry at the Universität Erlangen-Nürnberg. His research interests include organometallic chemistry and homogeneous catalysis, particularly as directed towards small molecule activation, (de-)hydrogenation, and inorganic materials synthesis.

Biographical Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information
Thumbnail image of

Jenni Meiners studied chemistry at Humboldt Universität Berlin and obtained her Diploma (M.Sc.) in 2007 in the group of Prof. P. W. Roesky (Freie Universität Berlin). She joined the group of Sven Schneider at the Technische Universität München in 2008. Currently she is finishing her Ph.D. at the Universität Erlangen-Nürnberg, working on the coordination chemistry of iridium amido complexes and C–H functionalization reactions.

Biographical Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. 16 Valence Electron (16 VE) Amido, Enamido, and Dienamido Complexes: Probing Properties and Reactivity
  5. 3. 15 and 14 VE Complexes: Control of Electronic Structure by Pincer Variation
  6. 4. Metal–Ligand Cooperativity: Hydrogenation and Dehydrogenation Catalysis
  7. Conclusions
  8. Acknowledgements
  9. Biographical Information
  10. Biographical Information
  11. Biographical Information
Thumbnail image of

Bjorn Askevold studied chemistry at the Technische Universität München in 2008 and obtained an M.Sc. under the supervision of Prof. J. Eppinger. In 2009 he joined the group of Sven Schneider at TU München as a Ph.D. student to work on low-coordinated iridium and ruthenium PNP complexes. He is currently finishing his Ph.D. Thesis at the Universität Erlangen-Nürnberg.