Scandium Terminal Imido Chemistry - Accounts of Chemical Research

Publication Date (Web): January 30, 2018. Copyright © 2018 ..... of CuI to Nimide diminishes the electron density donation from the imido ligand to S...
1 downloads 4 Views 2MB Size
Article Cite This: Acc. Chem. Res. 2018, 51, 557−566

pubs.acs.org/accounts

Scandium Terminal Imido Chemistry Erli Lu,† Jiaxiang Chu,‡ and Yaofeng Chen* State Key Laboratory of Organometallic Chemistry, Center for Excellence in Molecular Synthesis, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 345 Lingling Road, Shanghai 200032, P. R. China CONSPECTUS: Research into transition metal complexes bearing multiply bonded main-group ligands has developed into a thriving and fruitful field over the past half century. These complexes, featuring terminal ME/ME (M = transition metal; E = main-group element) multiple bonds, exhibit unique structural properties as well as rich reactivity, which render them attractive targets for inorganic/organometallic chemists as well as indispensable tools for organic/ catalytic chemists. This fact has been highlighted by their widespread applications in organic synthesis, for example, as olefin metathesis catalysts. In the ongoing renaissance of transition metal−ligand multiple-bonding chemistry, there have been reports of ME/ME interactions for the majority of the metallic elements of the periodic table, even some actinide metals. In stark contrast, the largest subgroup of the periodic table, rare-earth metals (Ln = Sc, Y, and lanthanides), have been excluded from this upsurge. Indeed, the synthesis of terminal LnE/LnE multiple-bonding species lagged behind that of the transition metal and actinide congeners for decades. Although these species had been pursued since the discovery of a rare-earth metal bridging imide in 1991, such a terminal (nonpincer/bridging hapticities) LnE/LnE bond species was not obtained until 2010. The scarcity is mainly attributed to the energy mismatch between the frontier orbitals of the metal and the ligand atoms. This renders the putative terminal LnE/LnE bonds extremely reactive, thus resulting in the formation of aggregates and/or reaction with the ligand/environment, quenching the multiple-bond character. In 2010, the stalemate was broken by the isolation and structural characterization of the first rare-earth metal terminal imidea scandium terminal imideby our group. The double-bond character of the ScN bond was unequivocally confirmed by singlecrystal X-ray diffraction. Theoretical investigations revealed the presence of two p−d π bonds between the scandium ion and the nitrogen atom of the imido ligand and showed that the dianionic [NR]2− imido ligand acts as a 2σ,4π electron donor. Subsequent studies of the scandium terminal imides revealed highly versatile and intriguing reactivity of the ScN bond. This included cycloaddition toward various unsaturated bonds, C−H/Si−H/B−H bond activations and catalytic hydrosilylation, dehydrofluorination of fluoro-substituted benzenes/alkanes, CO2 and H2 activations, activation of elemental selenium, coordination with other transition metal halides, etc. Since our initial success in 2010, and with contributions from us and across the community, this young, vibrant research field has rapidly flourished into one of the most active frontiers of rare-earth metal chemistry. The prospect of extending LnN chemistry to other rare-earth metals and/or different metal oxidation states, as well as exploiting their stoichiometric and catalytic reactivities, continues to attract research effort. Herein we present an account of our investigations into scandium terminal imido chemistry as a timely summary, in the hope that our studies will be of interest to this readership.

1. INTRODUCTION The synthesis and reactivity of transition metal−ligand multiple bonds (ME/ME; M = transition metal, E = main-group element) is one of the most vibrant areas of modern inorganic/ organometallic chemistry. Historically, from a fundamental perspective, novel ME/ME bonds challenged, tested, and aided the development of chemical bonding theory. From a practical perspective, ME/ME bonds have exhibited versatile stoichiometric and catalytic reactivitird, for example, in olefin metathesis and hydroelementations of CC/CC bonds, for which the Nobel Prize in Chemistry was awarded in 2005.1−3 In the transition metal complexes bearing multiply bonded main-group ligands, valence-shell d orbitals of the metal ion and p orbitals of the ligand E atom energetically match, resulting in © 2018 American Chemical Society

bonding interactions that are favorable. In stark contrast, even thought they constitute the largest subgroup of the periodic table, examples of multiple bonds to rare-earth metals (Ln = scandium, yttrium, and lanthanides) remain scarce.4 Attempts to make LnE bonds can be traced to the early 1990s,5 where the imido ligand ([NR]2−, R = alkyl, silyl, aryl) was selected as the top choice for stabilizing such interactions. In comparison with oxo (O2−), the imido ligand provides the necessary steric protection and electronic tunability by virtue of the R substituent. In comparison with the alkylidene ligand ([CR2]2−), where carbon is a softer donor atom than nitrogen, the hard nitrogen donor atom in the imido ligand is more Received: December 3, 2017 Published: January 30, 2018 557

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research

amide 4 or AlMe3-masked scandium imide 5 at room temperature in the presence of pyridine or AlMe3, respectively (Scheme 2).11 Scandium alkyl amide 3b, which bears a

favorable for bonding with hard Lewis acids such as rare-earth metal ions. Despite this, until 2010, all attempts to synthesize LnN terminal imides failed, leading to either dimers/ oligomers with bridging imide(s) or C−H bond activation by reaction of putative LnN species with the surroundings (solvent or the ligand scaffold).6,7 Though a stable LnN terminal imide remained unknown, these results consolidated the versatile and tempting reactivity prospect of this functionality, and it was believed that capturing the elusive target would require a well-designed ancillary ligand. The breakthrough in this field came in 2010, when we reported the isolation and structural characterization of the first rare-earth metal terminal imide: a scandium terminal imide.8 In the following 7 years, with contributions from our group and others, the field has flourished into one of the most vibrant frontiers of rare-earth metal chemistry. In this Account, we provide our personal perspective on this burgeoning field. Our scope focuses on scandium terminal imides, and the latest development of yttrium and lanthanide terminal imides will be introduced briefly. For readers to have a clear image of the development of scandium imido chemistry, bridging and masked scandium imides will be introduced. A detailed comparison with other transition metal imides will not be provided because of space limitations.

Scheme 2. Methane Elimination of Scandium Methyl Amide 3a in the Presence of Pyridine or AlMe3 and Alkane Elimination of Scandium Alkyl Amide 3b in Benzene

2. BRIDGING AND MASKED SCANDIUM IMIDES The first scandium imide was obtained unexpectedly, rather than being planned.9 In 2003, Hessen and co-workers reported a dimeric μ2-bridging imide 2, which was made from a reaction between scandium but-2-ene-1,4-diyl complex 1 and benzonitrile via insertion of the CN bond into the Sc−C bond (Scheme 1). Formation of the bridging imide 2 is a good

sterically bulky alkyl group, eliminates an alkane to give scandium phenyl amide 6 in benzene at 70 °C. Isotopic labeling studies demonstrated that 4 and 6 were formed via C−H activation of pyridine or benzene by a transient terminal imide. A scandium imide trimer 8 was reported by Piers and coworkers in 2009.12 This complex was obtained from the reaction of β-diketiminato-ligand-supported scandium dichloride 7 with LiHBEt3 (Scheme 3). The reaction likely progresses

Scheme 1. Formation of Scandium Bridging Imide 2 by C N Insertion into the Sc−C Bond and Subsequent Rearrangement

Scheme 3. Scandium Imide Trimer 8 Formed via Nucleophilic Migration and C−N Bond Cleavage

footnote for the inherent tendency of the postulated LnN species to dimerize, even for ScIII, which has the smallest ionic radius among all the rare-earth elements. Inspired by successes of a PNP-pincer monoanionic ligand ([N(2-P(iPr)2-4-Me-C6H3)2]−) in TiIVE/E (E = C, N) chemistry,10 Mindiola and co-workers adopted this ligand for the pursuit of a scandium terminal imide. However, scandium methyl amide 3a eliminates methane to afford scandium pyridyl

via a scandium hydride chloride (or borohydride chloride), which undergoes hydride transfer to the imine carbon of the βdiketiminato ligand to provide a diamide intermediate. Fragmentation of this intermediate via C−N bond cleavage gives an organic imine and a scandium imide, and the latter assembles into 8 in the presence of LiHBEt3. 558

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research

the ScNimide linkage in 10 was confirmed by the very short ScNimide bond (1.881(5) Å). This contrasts with the Sc− Namide single bond in 9 (2.047(3) Å). Furthermore, the Sc− Nimide−Cipso angle (169.6(5)°) is nearly linear, while the Sc− Namide−Cipso angle in 9 is 153.7(3)° (Figure 1). The ScNimide

3. SCANDIUM TERMINAL IMIDES: SYNTHESIS AND STRUCTURE As indicated by the formation of the scandium bridging imides, selection of the ancillary ligand is of crucial importance for pursuing rare-earth metal terminal imides. An ideal ancillary ligand is one with (a) steric bulk to prevent aggregation, (b) appropriate rigidity to fulfill coordination requirements and suppress ligand redistribution, and (c) chemical robustness to preclude ligand cannibalization. From a charge balance perspective, a monoanionic ancillary ligand (L−) is preferable for a ScIII complex, with the remaining charge balanced by the imido dianion. Dianionic or polyanionic ancillary ligands are undesirable, as they would likely bring in extra alkali metal cations (M2 = Li, Na, K), which would facilitate the formation of a Sc−(μ-Nimide)−M2 bridging structure and compromise the desired terminal ScN bond. According to the hard−soft acid−base (HSAB) principle, a hard Lewis acid, such as ScIII, prefers hard Lewis base donor(s), such as nitrogen and oxygen. On the basis of these considerations, we designed a monoanionic tridentate β-diketiminato ligand (L1−; Chart 1).13 A bulky 2,6-diisopropylphenyl (Dipp) substituent and a

Figure 1. Molecular structures of 9 and 10. Solvent molecules in the lattice, H atoms (except H4 in 9), and isopropyl substituents have been omitted for clarity.

bond length in 10 was the shortest Sc−N distance reported at that time. Density functional theory (DFT) calculations suggested that the Wiberg bond order of the Sc−Nimide bond in 10 is 1.32, which is much larger than that of the Sc−Namide bond in 9 (0.66). Molecular orbital analysis showed that two p orbitals of the Nimide atom form two π bonds with two d orbitals of the ScIII ion in 10 (Figure 2), whereas in 9, only one p orbital

Chart 1. Monoanionic Tridentate β-Diketiminato Ligand L1− and Tetradentate β-Diketiminato Ligand L2−

pendant amino arm were introduced for steric and electronic considerations, respectively. We adopted this ligand for the pursuit of a scandium terminal imide. Scandium methyl amide 9 supported by L1− (Scheme 4) was synthesized in three steps starting with ScCl3(THF)3 (THF = Scheme 4. Synthesis of Scandium Terminal Imide 10 Figure 2. Selected frontier molecular orbitals involved in the Sc Nimide interaction in 10. Reproduced with permission from ref 8. Copyright 2010 Royal Society of Chemistry.

of the Namide atom, which is vertical to the phenyl plane, overlaps with a d orbital of the ScIII ion. The imido ligand in 10 acts as a 2σ,4π electron donor. In the synthesis of 10, the external Lewis base DMAP was essential. Thus, we designed a tetradentate β-diketiminato ligand (L2−) (Chart 1) by introducing into L1− an additional amine donor, which is a functional analogue of DMAP. By the use of L2−, pentacoordinate scandium alkyl amide 11 was synthesized, in which the arm-end amine donor was coordination-free.14 11 was converted into the desired scandium terminal imide 12 upon SiMe4 elimination under mild conditions (Scheme 5).14 In 12, the arm-end amine donor coordinates to ScIII to provide a pentacoordinate scandium center similar to that found in 10. The ScNimide bond and Sc−Nimide−Cipso angle in 12 (1.8591(18) Å and 167.90(17)°) are comparable to those in 10 (1.881(5) Å and 169.6(5)°). After our initial successes, others reported more scandium terminal imides supported by a range of functionalized cyclopentadienyl or non-cyclopentadienyl ligands. All of these

tetrahydrofuran) and LiL1.8 In the presence of 1 equiv of the Lewis base 4-(N,N-dimethylamino)pyridine (DMAP) under mild conditions (50 °C, 48 h), 9 was converted into the desired scandium terminal imide 10 with concomitant elimination of methane (Scheme 4).8 10 was isolated as red crystals in 56% yield and was structurally characterized by single-crystal X-ray diffraction as the first rare-earth metal terminal imide with a bona fide LnN double bond. The double-bond character of 559

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research

similar dark-red color as 10 and 12, which may reflect the inherent electronic structure of the ScNimide linkage. As early as 2004, Piers and co-workers found that scandium methyl amide 18 supported by a prototypical β-diketiminato ligand (Nacnac) was converted into cyclometalated product 19 with concomitant CH4 elimination at 90 °C (Scheme 8).17 A

Scheme 5. Synthesis of Scandium Terminal Imide 12 without External Lewis Base

Scheme 8. Thermolysis of Scandium Methyl Amide 18 with and without DMAP efforts adopted alkane elimination from a scandium alkyl amide in the presence of DMAP. In 2012, Cui and co-workers reported scandium imide 14 supported by a phosphazene− cyclopentadienyl ligand ([C5H4P(Ph)2N(2,6-iPr2C6H3)]−) and DMAP (Scheme 6).15 Though 14 itself was not structurally Scheme 6. Synthesis of Scandium Imide 14 and Its Decomposition

deuterium labeling study suggested that 19 was formed via σbond metathesis between the C−H bond of the iPr group and the Sc−CH3 bond. In 2013, the authors reinvestigated the thermolysis of 18, this time in the presence of DMAP.18 Instead of the cyclometalated product 19, a new product 20 was formed. Multinuclear NMR spectroscopy, elemental analysis, and reactivity studies suggested that 20 is a tetracoordinate scandium terminal imide. Upon interrogation via comprehensive kinetic studies and isotopic labeling experiments, the formation of 20 was unveiled as proceeding via initial metalation of 18 to give 19 followed by rapid DMAP-promoted alkane elimination. In 2014, we reported that the DMAP ligand in 10 could be abstracted by a Lewis acidic borane to generate an unsaturated scandium terminal imido intermediate ([L1ScNDipp]), which upon trapping with THF provided THF-coordinated scandium terminal imide 21 (Scheme 9).19 The ScN bond length in 21 (1.852(4) Å) is slightly shorter than that in 10 (1.881(5) Å). 21 could not be directly prepared by thermolysis of scandium methyl amide 9 in THF. Compared with DMAPcoordinated 10, 21 offers some advantages: it is more reactive

characterized by X-ray crystallography, spectroscopic evidence supported the presence of the terminal ScN linkage. 14 is unstable at room temperature in solution: it conducts C−H activation of the ligand’s phenyl ring, yielding phenyl amide 15, which was characterized by single-crystal X-ray diffraction. In a follow-up work, Cui and co-workers switched to a nonCp ligand. In the presence of 2 equiv of DMAP, scandium alkyl amide 16 was rapidly converted into scandium terminal imide 17 at room temperature (Scheme 7).16 The kinetic barrier for this conversion is remarkably low; even at 0 °C the reaction was complete within 12 h. 17 was characterized by single-crystal Xray diffraction, and the Sc−Nimide bond length here (1.853(3) Å) is comparable to those in 10 (1.881(5) Å) and 12 (1.8591(18) Å). It is noteworthy that despite possessing a significantly different ancillary ligand framework, 17 has a

Scheme 9. Synthesis of THF-Coordinated Scandium Terminal Imide 21

Scheme 7. Synthesis of Scandium Terminal Imide 17

560

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research

imides and CO2.28 In both cases, two molecules of CO2 are inserted into the MN double bond. Upon reaction of 12 with 1 equiv of benzonitrile, an unexpected scandium amidinate, 23, was isolated and characterized as the major product. Within this complex, the methyl group of the β-diketiminate backbone is observed to have been deprotonated.27 The formation of 23 was postulated to proceed via initial [2 + 2] cycloaddition between the ScN and CN bonds followed by proton abstraction by the dianionic amidinate ligand in the intermediate. The reaction between 12 and phenyl isocyanate provided scandium N,O-bound ureate 24, which is apparently formed via [2 + 2] cycloaddition between the ScN and C O bonds followed by an isomerization to minimize steric repulsion.27 Treatment of 12 with propene oxide gave scandium allylic alkoxide amide 25.27 The formation of 25 involves ring-opening and hydrogen-transfer steps, but the detailed mechanism is unclear. A similar reaction in zirconium terminal imido chemistry was postulated to proceed via a zwitterionic intermediate.29 The reaction between 12 and a conjugated unsaturated substrate, methyl methacrylate (MMA), was also investigated. The peculiar scandium enolate 26 containing two eight-membered rings was isolated as the major product.27 The initial step is a [4 + 2] cycloaddition between 12 and MMA to form a six-membered enolate intermediate, which then participates in a Michael addition toward another molecule of MMA to provide the final product 26. This Michael addition step mimicked propagation in early transition-metal-catalyzed MMA polymerization. A related reaction of a zirconium terminal imide with methyl methacrylate was reported, which gave an analogous [2 + 4] cycloaddition product that did not undergo further Michael addition with a second methyl methacrylate.30 Cycloadditions of 12 with unsaturated substrates clearly demonstrated the nitrogen nucleophilicity and scandium Lewis acidity of the Sc N double bond. Furthermore, these cycloadditions were followed by other intriguing reactions, including Michael addition, hydrogen transfer, and isomerization; the unique products that resulted would have been difficult to synthesize otherwise. Though the reactivity of the scandium terminal imide toward polar unsaturated substrates (e.g., CO2, nitrile, and MMA) was comprehensively mapped, it was found to be inert toward alkenes and internal alkynes.31 This inertness can be attributed to coordinative saturation around the scandium ion and weak affinity of the hard Lewis acidic ScIII toward soft Lewis basic donors, e.g., alkenes and internal alkynes. Thus, we postulated that the reactivity of the ScN functionality could be further enhanced by applying a Lewis acid to abstract the Lewis basic donor (DMAP in 10), leading to coordinative unsaturation, thereby forming a highly reactive scandium imide species. A

than 10 because THF is more labile than DMAP, and the THF byproduct can be readily removed under vacuum.

4. SCANDIUM TERMINAL IMIDES: REACTIVITY The reactivity of the MN bond heavily depends upon the nature of the metal ion.20,21 While imido groups act as spectator ligands in molybdenum and tungsten complexes because of the inertness of these MN bonds,22 titanium and zirconium terminal imides bear highly reactive MN bonds.23−26 In this section, we will cover the stoichiometric reactivity of the scandium terminal imides toward organic/inorganic substrates and their potential applications in catalysis. This section is organized according to the different types of substrates being explored. Reactions toward Unsaturated Chemical Bonds

Cycloaddition toward unsaturated chemical bonds is a benchmark for the reactivity of ME bonds. Initially, when scandium terminal imide 10 was subjected to unsaturated substrates, reactions were observed to have taken place, but the isolation of product(s) was hampered by simultaneous formation of free DMAP as a byproduct. Thus, the DMAPfree scandium terminal imide 12 was employed instead. When a solution of 12 was exposed to 1 atm CO2 at ambient temperature, scandium dicarboxylate 22 was obtained (Scheme 10),27 which is comparable to reactions between nickel(II) Scheme 10. Investigations of the Reactivity of Scandium Terminal Imide 12 toward Polar Unsaturated Substrates

Scheme 11. [2 + 2] Cycloaddition between ScN and CC Bonds

561

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research strong Lewis acid, B(C6F5)3, was initially used, but the reactions gave complicated mixtures. We then turned to other boranes and found that a three-component reaction between 10, 9borabicyclononane (9-BBN), and 1-phenylpropyne clearly produced azascandacyclobutene complex 27 along with a DMAP→9-BBN adduct (Scheme 11).19 27 is a product of [2 + 2] cycloaddition between the ScN and CC bonds and represents the first azametallacyclobutene derivative of a rareearth metal.

Scheme 13. Coordinatively Unsaturated Scandium Imide Intermediate-Induced C−H Activation of Terminal Alkenes and Dehydrofluorination of Fluorobenzenes and Fluoroalkanes

Reactions toward Saturated Chemical Bonds

Even before the isolation of scandium terminal imides, it had been demonstrated that transient ScN species can activate C−H bonds.11,32 Inherently, C−H activation could compete with cycloaddition as a reaction pathway for the ScN bond. A good example of this was observed when imide 12 was treated with cyclopentadiene: instead of [2 + 2] or [2 + 4] cycloadditions with the CC bonds, C−H activation occurred to provide scandium amido cyclopentadienyl complex 28 (Scheme 12).27 Similarly, 12 reacted with (trimethylsilyl)This intramolecular C−H bond activation also indicated that coordination of DMAP is important for the stabilization of scandium terminal imide 10. Mindiola and co-workers reported that a transient scandium imide could activate and functionalize pyridine, with isonitriles at the para positions, to afford mono- and bis(imino)substituted pyridines.34 In one case, the reaction is catalytic. A proposed cycle for the formation of mono(imino)substituted pyridines (Scheme 14) includes C−H activation of pyridine by the imide, insertion of isonitrile into the Sc−C bond of the pyridyl complex, and regeneration of the ScN bond via 1,3-hydrogen migrations.

Scheme 12. C−H Activation of Cyclopentadiene or Diazoalkane Conducted by 12

Scheme 14. Proposed Cycle for C−H Activation and Functionalization of Pyridine by a Scandium Imide To Give Mono(imino)-Substituted Pyridines

diazomethane to give C−H activation product 29 rather than a cycloaddition product.33 Interestingly, 29 was structurally characterized as an η1-N binding scandium nitrilimine, which was the first example of a transition-metal-substituted nitrilimine. Steric congestion around the scandium ion and the strong Sc−Nnitrilimine interaction are responsible for this unusual bonding mode. 29 exhibited versatile reactivity toward a variety of unsaturated organic substrates, including aldehyde, ketone, nitrile, and allene derivatives.33 Treating terminal alkenes (3,3-dimethyl-1-butene and 1pentene) with coordinatively unsaturated, transient [L1Sc NDipp], generated via borane-facilitated DMAP abstraction from 10, provided scandium alkenyl amides 30 and 31 (Scheme 13).19 A combination of kinetic studies and DFT calculations revealed that the reactions are direct vinyl C−H bond activations rather than two-step reactions involving cycloadditions. Transient [L1ScNDipp] also activates the C−H bond of fluorobenzenes to generate scandium amido aryl intermediates, which subsequently undergo β-fluoride elimination to give scandium amido fluoride 32 and benzynes (Scheme 13). Beyond fluorobenzenes, fluoroalkanes can also be dehydrofluorinated by [L1ScNDipp] at room temperature to yield 32 and alkenes with high selectivity (Scheme 13). In the absence of a substrate, [L1ScNDipp] performed intramolecular C−H bond activation to generate alkyl amide 33.19 562

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research We found that scandium imide 10 was able to activate the Si−H bond in phenylsilane to produce scandium silylanilido hydride 34 (Scheme 15).35 The reaction is reversible, and the

Scheme 16. Reactions of Scandium Terminal Imides with Boranes

Scheme 15. Scandium Imide for Si−H Bond Activation and Catalytic Hydrosilylation of Imines

equilibrium constant (Ke = ([34][DMAP])/([10][PhSiH3])) was determined to be 3.04 at room temperature. 34 was not isolable because of the facile reversibility of the process, but it was trapped by an insertion reaction with diphenylcarbodiimide to give a silylanilido amidinate, which was isolated and structurally characterized. Furthermore, 10 was developed into an effective catalyst for the hydrosilylation of N-propyl-1phenylmethanimine using PhSiH3.35 With 5 mol % 10, the reaction was complete within 12 h at room temperature or 2 h at 50 °C. The monoaminosilane was obtained in nearly quantitative yield, whereas the diaminosilane was not observed. Compared with rare-earth-metal-catalyzed olefin hydrosilylation, examples of rare-earth-metal-catalyzed imine hydrosilylation are much scarcer because of the difficulty of silanolysis of the Ln−N bond.36,37 The scope and mechanism of this scandium terminal imide-catalyzed imine hydrosilylation is under investigation. ScN species could also activate the B−H bond. The reaction of 10 with 3 equiv of 9-BBN yielded scandium amido borohydride 35 and the DMAP→BBN adduct (Scheme 16).19 It should be noted that although B−H···M agostic-type interactions are occasionally observed in transition metal imido chemistry, only in very few cases is the B−H bond cleaved.38 The related reaction between 12 and 9-BBN was also studied (Scheme 16).39 In this case, C−H bond borylation on the ancillary ligand (L2−) occurred, which produced borohydride 36. We proposed that C−H activation of the methyl group (−CH2NMe2) of L2− by the ScN functionality generated a scandium alkyl amido intermediate, which subsequently took part in a nucleophilic attack on the boron atom of 9-BBN to give the final product 36. When catecholborane (CatBH) was used, the reaction provided scandium catecholate 37 and iminoborane. In catecholborane, the bond dissociation energy of the B(H)−O bond (890 kJ/ mol) is much higher than that of the B(O)−H bond (416 kJ/ mol).40 The observed cleavage of the stronger B−O bond instead of the weaker B−H bond is due to the strong oxophilicity of ScIII. Unlike 10 and 12, THF-coordinated scandium terminal imide 21 reacted readily with H2 to give scandium terminal hydride 38 (Scheme 17).41 This can be ascribed to the lower

Scheme 17. Reversible H2 Activation by Scandium Imide

energy for dissociation of THF from the scandium center compared with that of DMAP or an amino group. The reaction proceeded via a 1,2-addition mechanism, which is distinct from the well-established σ-bond metathesis mechanism for rareearth-metal-mediated H2 activation. DFT studies revealed that the energy barriers for the replacement of THF by H2 and the 1,2-addition of the H−H bond across the ScN bond are both very low, with a total activation barrier of 7.5 kcal/mol. Scandium amido hydride 38 can eliminate H2 to regenerate the ScN bond; thus, the 1,2-addition of the H−H bond across the ScN bond is reversible. The reversibility was demonstrated by treatment of 38 with D2, which gave 38-D, as shown in Scheme 17.41 In the presence DMAP, the known imide 10 was formed. DFT calculations demonstrated that DMAP facilitates the H2 elimination process. Reactions toward Inorganic Substrates

In contrast to organic substrates, the reactivity of metal imides with inorganic substrates has been far less thoroughly studied. Nonetheless, the scandium terminal imides exhibited interesting reactivity toward chalcogen and transition metal halides. The reaction of scandium terminal imide 10 or 12 with 563

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research 2

3

elemental selenium at ambient temperature gave Csp /sp −H selenation product 39 or 40 (Scheme 18).14 There are two

(2.926(8) Å). DFT calculations showed that the Wiberg bond order of the Sc−Nimide bond in 41 (1.12) is smaller than that in 10 (1.33), indicating that coordination of CuI to Nimide diminishes the electron density donation from the imido ligand to ScIII. DFT calculations also showed overlaps between the empty 3d orbitals of ScIII and the filled 3d orbitals of CuI, with a Wiberg bond order of 0.20 for the Cu−Sc interaction. On the other hand, reactions of 12 with [M(COD)Cl]2 (M = Rh, Ir; COD = 1,5-cyclooctadiene) gave heterobimetallic complexes 42 and 43.43 Here the electrophilic ScIII center in 12 abstracted a chloride, and the [M(COD)]+ fragment coordinated to the phenyl ring of the imido ligand. The Sc−Nimide bond lengths in 42 (1.928(4) Å) and 43 (1.937(4) Å) are longer than that in 12 (1.8591(18) Å), while the C−Nimide bond lengths are shorter (1.300(5) Å in 42 and 1.286(6) Å in 43 vs 1.360(3) Å in 12). Accordingly, the weakening of the Sc−Nimide bond and the strengthening of the C−Nimide bond in going from 12 to 42/43 were also corroborated by the Wiberg bond order analysis (Sc− Nimide: 1.35 in 12 vs 0.95/0.94 in 42/43; C−Nimide: 1.22 in 12 vs 1.58/1.60 in 42/43).

Scheme 18. Reactions of Scandium Terminal Imides toward Elemental Selenium

5. SUMMARY AND PROSPECTS After the report of a rare-earth metal bridging imide in 1991,5 it took nearly two decades before the first rare-earth metal terminal imide (the scandium terminal imide) was isolated and structurally characterized.8 In stark contrast to the covalently bound transition-metal−nitrogen double bond, the ScN bond is more ionic and polarized, which originates from a mismatch in orbital energies. The ionic and polarized ScN bond exhibits versatile reactivity that shows both similarities and distinctions from those of other transition metal imides. Now the scope of rare-earth metal terminal imido species has been extended from ScIII to include YIII, LuIII,44 and CeIV.45 However, there are still a range of rare-earth metals and oxidation states that have yet to be realized for the LnN functionality (Chart 2), especially given that the +2 oxidation

possible mechanistic scenarios for the selenation: (1) [2 + 1] cycloaddition between ScN and selenium provides a [Sc− N−Se] three-membered-ring intermediate, as observed in the reaction of selenium with the scandium alkylidene complex [{MeC(NDipp)CHC(Me)(NCH2CH2N(iPr)2)}Sc{C(SiMe3)PPh2S}].42 The [Sc−N−Se] intermediate subsequently rearranges into the product. (2) Intramolecular C−H activation of DMAP or L2− by the ScN bond generates a scandium amido 2 3 pyridyl/alkyl intermediate with a Sc−Csp /sp bond, followed by 2

3

insertion of selenium into the Sc−Csp /sp bond. The reaction of 10 with CuI provided ScIII/CuI heterobimetallic bridging imide 41 (Scheme 19),43 with the nucleophilic Nimide in 10 coordinated to CuI. This heterobimetallic complex 41 features a short Sc···Cu distance

Chart 2. Prospects of Rare-Earth Metal Terminal Imido Chemistry

Scheme 19. Reactions of Scandium Terminal Imides toward CuI or [M(COD)Cl]2 (M = Rh, Ir)

state is now available for all rare-earth metals.46 On the imido side, other than the aryl substituents, a scandium borylimide was recently reported.47 By manipulation of the substituent on Nimide, [NR]2− ligands of bespoken electronic/steric profiles can be envisaged, which will expand both the structural and reactivity scopes of the LnN chemistry. Although the stoichiometric reactivity of scandium terminal imides has been well-explored in last 7 years, the catalytic reactivity of these complexes as well as that of other rare-earth metal terminal imides is almost untouched. This deserves special 564

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research

(2) Schrock, R. R. Multiple Metal-Carbon Bonds for Catalytic Metathesis Reactions (Nobel Lecture). Angew. Chem., Int. Ed. 2006, 45, 3748−3759. (3) Grubbs, R. H. Olefin-Metathesis Catalysts for the Preparation of Molecules and Materials (Nobel Lecture). Angew. Chem., Int. Ed. 2006, 45, 3760−3765. (4) Giesbrecht, G. R.; Gordon, J. C. Lanthanide Alkylidene and Imido Complexes. Dalton Trans. 2004, 2387−2393. (5) Trifonov, A. A.; Bochkarev, M. N.; Schumann, H.; Loebel, J. Reduction of Azobenzene by Naphthaleneytterbium: A Tetranuclear Ytterbium(III) Complex Combining 1,2-Diphenylhydrazido(2−) and Phenylimido Ligands. Angew. Chem., Int. Ed. Engl. 1991, 30, 1149− 1151. (6) Liu, Z. X.; Chen, Y. F. One Frontier of the Rare-earth Organometallic Chemistry: The Chemistry of Rare-earth Metal Alkylidene, Imido and Phosphinidene Complexes. Zhongguo Kexue: Huaxue 2011, 41, 304−313. (7) Summerscales, O. T.; Gordon, J. C. Complexes Containing Multiple Bonding Interactions between Lanthanoid Elements and Main-Group Fragments. RSC Adv. 2013, 3, 6682−6692. (8) Lu, E. L.; Li, Y. X.; Chen, Y. F. A Scandium Terminal Imido Complex: Synthesis, Structure and DFT Studies. Chem. Commun. 2010, 46, 4469−4471. (9) Beetstra, D. J.; Meetsma, A.; Hessen, B.; Teuben, J. H. (Cyclopentadienylamine)scandium(2,3-dimethyl-1,3-butadiene): A 1,3-Diene Complex of Scandium with Sc(I)- and Sc(III)-like Reactivity. Organometallics 2003, 22, 4372−4374. (10) Bailey, B. C.; Fan, H.; Baum, E. W.; Huffman, J. C.; Baik, M.-H.; Mindiola, D. J. Intermolecular C−H Bond Activation Promoted by a Titanium Alkylidyne. J. Am. Chem. Soc. 2005, 127, 16016−16017. (11) Scott, J.; Basuli, F.; Fout, A. R.; Huffman, J. C.; Mindiola, D. J. Evidence for the Existence of a Terminal Imidoscandium Compound: Intermolecular C−H Activation and Complexation Reactions with the Transient ScNAr Species. Angew. Chem., Int. Ed. 2008, 47, 8502− 8505. (12) Conroy, K. D.; Piers, W. E.; Parvez, M. Nucleophilic Degradation of a β-Diketiminato Ancillary by a Transient Scandium Hydride Intermediate. Organometallics 2009, 28, 6228−6233. (13) Xu, X.; Xu, X. Y.; Chen, Y. F.; Sun, J. Dialkyllanthanide Complexes Containing New Tridentate Monoanionic Ligands with Nitrogen Donors. Organometallics 2008, 27, 758−763. (14) Lu, E. L.; Chu, J. X.; Chen, Y. F.; Borzov, M. V.; Li, G. Y. Scandium Terminal Imido Complex Induced C−H Bond Selenation and Formation of an Sc−Se Bond. Chem. Commun. 2011, 47, 743− 745. (15) Jian, Z. B.; Rong, W. F.; Mou, Z. H.; Pan, Y. P.; Xie, H. Y.; Cui, D. M. Intramolecular C−H Bond Activation Induced by a Scandium Terminal Imido Complex. Chem. Commun. 2012, 48, 7516−7518. (16) Rong, W. F.; Cheng, J. H.; Mou, Z. H.; Xie, H. Y.; Cui, D. M. Facile Preparation of a Scandium Terminal Imido Complex Supported by a Phosphazene Ligand. Organometallics 2013, 32, 5523−5529. (17) Knight, L. K.; Piers, W. E.; Fleurat-Lessard, P.; Parvez, M.; McDonald, R. β-Diketiminato Scandium Chemistry: Synthesis, Characterization, and Thermal Behavior of Primary Amido Alkyl Derivatives. Organometallics 2004, 23, 2087−2094. (18) Chu, T.; Piers, W. E.; Dutton, J. L.; Parvez, M. Synthesis and Reactivity of a Terminal Scandium Imido Complex. Organometallics 2013, 32, 1159−1165. (19) Chu, J. X.; Han, X. H.; Kefalidis, C. E.; Zhou, J. L.; Maron, L.; Leng, X. B.; Chen, Y. F. Lewis Acid Triggered Reactivity of a Lewis Base Stabilized Scandium Terminal Imido Complex: C-H Bond Activation, Cycloaddition, and Dehydrofluorination. J. Am. Chem. Soc. 2014, 136, 10894−10897. (20) Wigley, D. E. Organoimido Complexes of the Transition Metals. Prog. Inorg. Chem. 1994, 42, 239−482. (21) 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-Interscience: New York, 1998.

attention in the future, and those studies will likely be rewarded by the discovery of interesting catalytic properties.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Yaofeng Chen: 0000-0003-4664-8980 Present Addresses †

E.L: School of Chemistry, The University of Manchester, Oxford Road, Manchester M13 9PL, UK. ‡ J.C.: Department of Chemistry and Biochemistry, University of California, Santa Barbara, Goleta, CA 93106-9510, USA. Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest. Biographies Erli Lu received a B.Eng. from Tianjin Polytechnic University in 2006 and a Ph.D. from Shanghai Institute of Organic Chemistry with Prof. Yaofeng Chen in 2012. In 2012 he moved to the U.K. and has worked in the research group of Prof. Stephen T. Liddle at the University of Nottingham and at the University of Manchester as a research fellow, investigating actinide chemistry. Jiaxiang Chu received a B.S. from Shandong University in 2008. He was then admitted to Shanghai Institute of Organic Chemistry and obtained his Ph.D. in 2013 under the supervision of Prof. Yaofeng Chen. He worked as a postdoctoral fellow in Prof. Guy Bertrand’s group at the University of California, San Diego, from 2014 to 2016 and is now a postdoctoral fellow in Prof. Gabriel Menard’s group at the University of California, Santa Barbara. Yaofeng Chen received a B.S. from Hangzhou Normal University in 1993, a Master’s degree from Hangzhou University with Prof. Yuqiu Gong in 1996, and a Ph.D. from Zhejiang University with Prof. Zhiquan Shen in 1999. After that, he did postdoctoral studies at the Shanghai Institute of Organic Chemistry (1999−2002, Prof. Changtao Qian’s group), the University of Montreal (2002−2003, Prof. Davit Zargarian’s group), and the University of California, Santa Barbara (2003−2005, Prof. Guillermo C. Bazan’s group). Since 2006 he has been a research professor at Shanghai Institute of Organic Chemistry. His current research interest is focused on the synthesis and reactivity of rare-earth metal complexes.



ACKNOWLEDGMENTS We are indebted to Prof. Yuxue Li, Prof. Laurent Maron, and the members of their research groups for theoretical studies related to this work. This work was supported by the National Natural Science Foundation of China (Grants 21325210, 21421091, and 21732007) and the Strategic Priority Research Program of the Chinese Academy of Sciences (Grant XDB20000000).



REFERENCES

(1) Chauvin, Y. Olefin Metathesis: The Early Days (Nobel Lecture). Angew. Chem., Int. Ed. 2006, 45, 3740−3765. 565

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566

Article

Accounts of Chemical Research (22) Schrock, R. R. Recent Advances in Olefin Metathesis by Molybdenum and Tungsten Imido Alkylidene Complexes. J. Mol. Catal. A: Chem. 2004, 213, 21−30. (23) Hazari, N.; Mountford, P. Reactions and Applications of Titanium Imido complexes. Acc. Chem. Res. 2005, 38, 839−849. (24) Mindiola, D. J. Oxidatively Induced Abstraction Reactions. A Synthetic Approach to Low-Coordinate and Reactive Early Transition Metal Complexes Containing Metal-Ligand Multiple Bonds. Acc. Chem. Res. 2006, 39, 813−821. (25) Fout, A. R.; Kilgore, U. J.; Mindiola, D. J. The Progression of Synthetic Strategies to Assemble Titanium Complexes Bearing the Terminal Imide Group. Chem. - Eur. J. 2007, 13, 9428−9440. (26) Duncan, A. P.; Bergman, R. G. Selective Transformations of Organic Compounds by Imidozirconocene Complexes. Chem. Rec. 2002, 2, 431−445. (27) Chu, J. X.; Lu, E. L.; Liu, Z. X.; Chen, Y. F.; Leng, X. B.; Song, H. B. Reactivity of a Scandium Terminal Imido Complex Towards Unsaturated Substrates. Angew. Chem., Int. Ed. 2011, 50, 7677−7680. (28) Mindiola, D. J.; Waterman, R.; Iluc, V. M.; Cundari, T. R.; Hillhouse, G. L. Carbon−Hydrogen Bond Activation, C−N Bond Coupling, and Cycloaddition Reactivity of a Three-Coordinate Nickel Complex Featuring a Terminal Imido Ligand. Inorg. Chem. 2014, 53, 13227−13238. (29) Blum, S. A.; Walsh, P. J.; Bergman, R. G. Epoxide-Opening and Group-Transfer Reactions Mediated by Monomeric Zirconium Imido Complexes. J. Am. Chem. Soc. 2003, 125, 14276−14277. (30) Jin, J. Z.; Mariott, W. R.; Chen, E. Y. X. Polymerization of Methyl Methacrylate by Metallocene Imido Complexes and Tris(pentafluorophenyl)alane. J. Polym. Sci., Part A: Polym. Chem. 2003, 41, 3132−3142. (31) Scandium terminal imido complex 10 reacts with phenylacetylene, which has a relatively acidic proton, to form a deprotonated product of scandium amido phenylacetylide. The same reaction with scandium terminal imido complex 20 was reported by Piers and coworkers (see ref 18). (32) Wicker, B. F.; Fan, H. J.; Hickey, A. K.; Crestani, M. G.; Scott, J.; Pink, M.; Mindiola, D. J. Evidence for the Existence of Terminal Scandium Imidos: Mechanistic Studies Involving Imido-Scandium Bond Formation and C-H Activation Reactions. J. Am. Chem. Soc. 2012, 134, 20081−20096. (33) Chu, J. X.; Kefalidis, C. E.; Maron, L.; Leng, X. B.; Chen, Y. F. Chameleon Behavior of a Newly Synthesized Scandium Nitrilimine Derivative. J. Am. Chem. Soc. 2013, 135, 8165−8168. (34) Wicker, B. F.; Scott, J.; Fout, A. R.; Pink, M.; Mindiola, D. J. Atom-Economical Route to Substituted Pyridines via a Scandium Imide. Organometallics 2011, 30, 2453−2456. (35) Chu, J. X.; Lu, E. L.; Chen, Y. F.; Leng, X. B. Reversible Addition of the Si−H Bond of Phenylsilane to the ScN Bond of a Scandium Terminal Imido Complex. Organometallics 2013, 32, 1137− 1140. (36) Obora, Y.; Ohta, T.; Stern, C. L.; Marks, T. J. Organolanthanide-Catalyzed Imine Hydrogenation. Scope, Selectivity, Mechanistic Observations, and Unusual Byproducts. J. Am. Chem. Soc. 1997, 119, 3745−3755. (37) Takaki, K.; Kamata, T.; Miura, Y.; Shishido, T.; Takehira, K. Dehydrogenative Silylation of Amines and Hydrosilylation of Imines Catalyzed by Ytterbium-Imine Complexes. J. Org. Chem. 1999, 64, 3891−3895. (38) Mellino, S.; Stevenson, L. C.; Clot, E.; Mountford, P. Reactions of Titanium Imides and Hydrazides with Boranes. Organometallics 2017, 36, 3329−3342. (39) Chu, J. X.; Wang, C.; Xiang, L.; Leng, X. B.; Chen, Y. F. Reactivity of Scandium Terminal Imido Complex toward Boranes:C(sp3)-H Bond Borylation and B−O Bond Cleavage. Organometallics 2017, 36, 4620−4625. (40) Luo, Y. R. Comprehensive Handbook of Chemical Bond Energies; CRC Press: Boca Raton, FL, 2007. (41) Han, X. H.; Xiang, L.; Lamsfus, C. A.; Mao, W. Q.; Lu, E. L.; Maron, L.; Leng, X. B.; Chen, Y. F. Synthesis and Reactivity of a

Scandium Terminal Hydride: H2 Activation by a Scandium Terminal Imido Complex. Chem. - Eur. J. 2017, 23, 14728−14732. (42) Wang, C.; Zhou, J. L.; Zhao, X. F.; Maron, L.; Leng, X. B.; Chen, Y. F. Non-Pincer-Type Mononuclear Scandium Alkylidene Complexes: Synthesis, Bonding, and Reactivity. Chem. - Eur. J. 2016, 22, 1258−1261. (43) Lu, E. L.; Zhou, Q. H.; Li, Y. X.; Chu, J. X.; Chen, Y. F.; Leng, X. B.; Sun, J. Reactivity of Scandium Terminal Imido Complexes Towards Metal Halides. Chem. Commun. 2012, 48, 3403−3405. (44) Schädle, D.; Meermann-Zimmermann, M.; Schädle, C.; Maichle-Mössmer, C.; Anwander, R. Rare-Earth Metal Complexes with Terminal Imido Ligands. Eur. J. Inorg. Chem. 2015, 2015, 1334− 1339. (45) Solola, L. A.; Zabula, A. V.; Dorfner, W. L.; Manor, B. C.; Carroll, P. J.; Schelter, E. J. Cerium(IV) Imido Complexes: Structural, Computational, and Reactivity Studies. J. Am. Chem. Soc. 2017, 139, 2435−2442. (46) MacDonald, M. R.; Bates, J. E.; Ziller, J. W.; Furche, F.; Evans, W. J. Completing the Series of + 2 Ions for the Lanthanide Elements: Synthesis of Molecular Complexes of Pr2+, Gd2+, Tb2+, and Lu2+. J. Am. Chem. Soc. 2013, 135, 9857−9868. (47) Clough, B. A.; Mellino, S.; Clot, E.; Mountford, P. New Scandium Borylimido Chemistry: Synthesis, Bonding, and Reactivity. J. Am. Chem. Soc. 2017, 139, 11165−11183.

566

DOI: 10.1021/acs.accounts.7b00605 Acc. Chem. Res. 2018, 51, 557−566