Coordination of LiH Molecules to Mo≣Mo Bonds: Experimental and Computational Studies on Mo2LiH2, Mo2Li2H4, and Mo6Li9H18 Clusters

The reactions of LiAlH4 as the source of LiH with complexes that contain (H)Mo≣Mo and (H)Mo≣Mo(H) cores stabilized by the coordination of bulky AdDipp2 ligands result in the respective coordination of one and two molecules of (thf)LiH, with the generation of complexes exhibiting one and two HLi(thf)H ligands extending across the Mo≣Mo bond (AdDipp2 = HC(NDipp)2; Dipp = 2,6-iPr2C6H3; thf = tetrahydrofuran, C4H8O). A theoretical study reveals the formation of Mo–H–Li three-center–two-electron bonds, supplemented by the coordination of the Mo≣Mo bond to the Li ion. Attempts to construct a [Mo2{HLi(thf)H}3(AdDipp2)] molecular architecture led to spontaneous trimerization and the formation of a chiral, hydride-rich Mo6Li9H18 supramolecular organization that is robust enough to withstand the substitution of lithium-solvating molecules of tetrahydrofuran by pyridine or 4-dimethylaminopyridine.


■ INTRODUCTION
Along with noble gas helium, hydrogen and lithium are the simplest, lightest elements and the only ones that existed in the young universe. 1 Helium hydride, HeH + , is a molecule of astrophysical importance, 2 whereas LiH is the lightest metal hydride and arouses considerable interest due to its many applications. 3−5 In the gas phase, molecules of LiH exist as a result of the overlap of the singly occupied H 1s and Li 2s atomic orbitals, 6 with an experimentally determined interatomic distance of ca. 1.60 Å. 7 In the solid state, LiH adopts a cubic NaCl-type structure, characterized by long Li···H contacts of approximately 2.04 Å. 3 Molecular hydrides of the s-block elements have been intensively investigated in recent years. For group 2 metals, new, uncommon structures and a diversity of useful applications in hydrometalation, hydrogenation, and other reactions have been uncovered, thanks in no small part to the use of sterically encumbered auxiliary ligands. 3,8−19 Progress for the alkali metals has been more limited, although with notable exceptions. These include Stasch's hydrocarbonsoluble lithium hydride cluster [{(DippNPPh 2 )Li} 4 (LiH) 4 ], containing a (LiH) 4 central cube (Dipp = 2,6-i Pr 2 C 6 H 3 ), 20 as well as the generation by Mulvey and co-workers of hexanesoluble lithium hydride transfer reagents. 21,22 Of particular relevance is the synthesis of the dilithiozincate hydride [(tmed)Li] 2 [{ i PrNCH 2 CH 2 N( i Pr)}Zn( t Bu)H] that retains the Li−H bond in solution and undergoes the dynamic association and dissociation of (tmed)LiH. 21 Also noteworthy are reports on hydride encapsulation by molecular alkali metal clusters, 23 the structural characterization of the LiH and LiO t Bu agglomerate Li 33 H 17 (O t Bu) 16 , 24 and the synthesis of a (LiH) 4 cube coordinated to three bis(amido)alane units. 25 Transition-metal complexes allegedly containing coordinated monomeric molecules of LiH are sparse. There are, however, some reports describing M−H−Li systems where a degree of covalent bonding within the bridging bond can be proposed on the basis of the observation of one-bond 1 H− 6,7 Li NMR coupling constants. 21,26−35 Despite the scarcity of complexes of this type, it is conceivable that, like other main-group metal−hydrogen bonds (e.g., Mg−H, Al−H, and Zn−H), 36−44 a molecule of lithium hydride might bind to a transition-metal fragment through its Li−H bond, assisted by an interaction with an adjoining ligand that could compensate for the unsaturation of the lithium coordination shell and further heighten the σ-donor strength of the polar Li−H bond.
In this context, we envisioned that quadruply bonded hydride central units [Mo 2 (H) n ] (n = 1, 2) could be utilized to build the target molecular architectures. As represented in structure A of Figure 1, such dimolybdenum dihydride units possess strong hydride character 45 and feature Mo−Mo separations of close to 2.10 Å, with Mo−H vectors nearly perpendicular to the Mo−Mo bond. 45 Here, we discuss the synthesis and structure of complexes 3·thf and 4·thf ( Figure 1) that contain one and two formally monoanionic, bridging H− Li(thf)−H ligands, respectively, spanning the Mo≣Mo bond. We also study the unexpected formation of a unique, hydrocarbon-soluble, hydride-rich Mo 6 Figure 1). 45 The Mo 2 (H) 2 functionality of this complex was engendered by hydrogenolysis of the Mo−C bonds of the [(Me)Mo≣Mo-(Me)] homologue, 47 a method that continues to be a main vehicle for the synthesis of transition-metal hydrides. Searching for a related monohydride [(H)Mo≣Mo] core, we carried out the two-step transformation shown in Scheme 1a. Low-temperature alkylation of [Mo 2 (μ-O 2 CH) 2 (μ-Ad Dipp2 ) 2 ] with equimolar amounts of LiEt yielded an ethyl-formate intermediate that was reacted in situ with H 2 and converted to the hydride-formate product, 2·thf (Scheme 1), in good isolated yields (ca. 70%). The coordinated tetrahydrofuran molecule of 2·thf is highly labile, and it was readily replaced by Lewis bases such as 4-dimethylaminopyridine (dmap), 1,3,4,5-tetramethylimidazol-2-ylidene (IMe 4 ), and PMe 3 , giving complexes 2·L (Scheme 1a, top). Similarly, the use of LiAlH 4 as a source of LiH permitted the isolation of complex 3·thf that was obtained as a yellow solid in yields of around 60%. This reaction was not, however, simple and also produced related derivative 4· thf, along with minute amounts of a tetrahydroaluminate complex to be described elsewhere. Complex 3·thf possesses a H−Mo≣Mo−H−Li(thf) chelate moiety resulting from the substitution of the coordinated tetrahydrofuran of 2·thf by a molecule of (thf)LiH, with the formation of a σ-Li−H complex, that becomes stabilized by the concomitant formation of a 3c−2e Mo−H⇀Li bond involving the adjacent Mo−H terminus.
Next, 1·thf was utilized as a source of the [Mo 2 (H) 2 ] center (Scheme 1b). Mixing a tetrahydrofuran solution of this complex with a solution of LiAlH 4 in the same solvent caused the immediate precipitation of a bright-yellow solid that was ] cores by the incorporation of one, two, or three molecules of (thf)LiH (n = 1, complex 3·thf; n = 2, A and complex 4·thf; when n = 3, the unobserved monomer trimerizes to complex 5·thf with the loss of a molecule of tetrahydrofuran). In the structure of 5·thf, symmetry-related lithium atoms share the same color.
identified as dilithium tetrahydride dimolybdenum complex [Mo 2 {μ-HLi(thf)H} 2 (μ-Ad Dipp2 ) 2 ] (4·thf), that is, as a Mo 2 Li 2 H 4 cluster. As drawn in Scheme 1b, the compound contains two trans-[μ-HLi(thf)H] ligands that extend across the Mo≣Mo bond. Thus, it can be related to 3·thf by means of formal replacement of the bridging formate of the latter by a second μ-Li(thf)H 2 − three-atom chelating ligand. In agreement with this rationale, complexes 3·thf and 4·thf were generated in high yields (70−85%) by the more direct method summarized in Scheme 1c, based on the reaction of readily available [Mo 2 (μ-O 2 CH) 2 (μ-Ad Dipp2 ) 2 ] with LiAlH 4 , under appropriate conditions. Complexes 2·L, 3·thf, and 4·thf were characterized with the aid of microanalytical, spectroscopic, and X-ray data and were additionally studied by computational methods. For molecules of 2·L, the proposed structure is based on IR and NMR data and was unmistakably confirmed by X-ray crystallography for 2·IMe 4 ( Figure S1). Regarding complexes 3·thf and 4·thf, their hydride signals were not readily apparent in the IR spectra, possibly because of the Mo−H−Li bridging character, so the unambiguous identification of the three-atom HLiH chains in 3·thf and 4·thf owes much to the 1 H and 7 Li NMR experiments developed. Surprisingly more soluble in benzene and toluene than in tetrahydrofuran, the H atoms of the HLi(thf)H ligand in 3·thf resonate at δ 4.33 (C 6 D 6 ), appearing as a partially resolved multiplet due to coupling to the 7 Li (92.6%, I = 3/2) and 6 Li (7.4%, I = 1) nuclei. As can be seen in Figure 2, this signal becomes a singlet in the 1 H{ 7 Li} NMR spectrum. Moreover, the 4.33 multiplet is absent in the 1 H NMR spectrum of the DLiD isotopologue of 3·thf, prepared by the reaction of [Mo 2 (μ-O 2 CH) 2 (μ-Ad Dipp2 ) 2 ] 48 with LiAlD 4 . The 7 Li{ 1 H} NMR spectrum is a somewhat broad singlet at 3.6 ppm that transforms into a 1:2:1 triplet in the proton-coupled 7 Li NMR experiment, with a one-bond 7 Li− 1 H coupling constant of 16 Hz.
Complex 4·thf is only scarcely soluble in common solvents such as benzene, toluene, and tetrahydrofuran, but it is just sufficiently soluble in C 6 H 5 F for NMR studies. Pertinent NMR data are also included in Figure 2. In particular, the Mo 2 LiH 2 moieties exhibit comparable 1 J( 7 Li− 1 H) couplings of 17 Hz. These observations categorically demonstrate the existence of HLiH entities coordinated to the Mo−Mo quadruple bond in the 3·thf and 4·thf molecules. Besides, they attest without a doubt to the fact that, although probably mainly Coulombic in nature (vide infra), the Mo−H−Li−H−Mo bonding interactions involve a considerable degree of covalency, that is, of substantial electron density shared among the molybdenum, hydrogen, and lithium valence orbitals. It is pertinent to remark that the observation of scalar coupling in lithium hydride complexes is rare, to the point that few 1 J( 6,7 Li− 1 H) values can be found in the literature. 21  Complexes 3·thf and 4·thf possess good thermal stability. Their C 6 D 6 and C 6 D 5 F solutions appear to be stable for 1 day at room temperature, though decomposition occurs upon heating at 70°C for 3 to 4 h. In the solid state, decomposition is apparent only at T ≥ 150°C. The two compounds behave as soluble LiH carriers, particularly 4·thf, which is the more reactive of the two. For instance, complex 4·thf reacted with Ph 2 C(O) to give the expected alkoxide Ph 2 C(H)(OLi). 20, 22 Their molecular structures were investigated by X-ray crystallography and optimized by means of DFT calculations. Owing to poor crystal properties, the data collected for the former do not permit a rigorous structural discussion, particularly with respect to what concerns the geometric parameters of H atoms. Nonetheless, they allow us to define beyond any doubt the connectivity represented in Figure S2. Figure 3 contains an ORTEP representation of the molecules of 4·thf, along with selected metrics. A more complete set of bond distances is collected in Table 1, which contains both experimental and computational data. When this manuscript was being prepared, there was no precedent for a structural motif of this kind in the Cambridge Structural Database (CSD). 50 Table 1 is most likely due to the incertitude in the localization by X-ray diffraction of hydride ligands bound to a heavy atom such as molybdenum. The computed distances are 1.85 and 1.78 Å, respectively. The first is almost coincident with the average Mo−H−Mo bond lengths determined by neutron diffraction, 54 while the second is somewhat longer than the 1.60 Å value measured for the molecule of LiH in the gas phase but significantly shorter than the interatomic separation of 2.04 Å found for this hydride in the solid state. 3 Regarding the Mo−Li distances, the experimental values of 2.91(2) and 2.97(2) Å are indistinguishable within experimental error, whereas in the optimized structure this slight asymmetry vanishes, leading to a separation of ca. 2.97 Å. For comparison, the sum of the covalent radii of the atoms is 2.82 Å. 55 We have carried out geometry optimization and an NBO analysis of chemical bonding within the Mo−H−Li−H−Mo rings of 4·thf. For simplicity, we describe here the comparable    (Figure 4a). In addition, we find that the dx 2 −y 2 orbitals, not involved in Mo−Mo bonding, form spd hybrids directed toward the hydrides 47 and combine with s(H) orbitals to form the two Mo−H bonds (one of which is shown in Figure 4a). The NBO approach results in limited participation of the lithium atomic orbitals in occupied MOs. However, this does not mean that its interactions with the hydrides and the molybdenum atoms are strictly ionic, since the calculated charge on Li is +0.67, indicative of a non-negligible covalent contribution. The reduced charge of the lithium "ion" is thus associated, in addition to thf → Li donation, with two sets of donor−acceptor interactions:   covalent radii sum of 2.82 Å. 55 Although of lesser quantitative importance, the existence of non-negligible electron donation from the bonding π and δ(Mo≣Mo) orbitals is worth being stressed. The fact that the calculated dissociation energy of 4· thf′ into (thf)Li−H and dihydride [Mo 2 (H) 2 (μ-Ad Dipp2 ) 2 ] is 27.9 kcal/mol, smaller than the sum of NBO interaction energies shown in Figure 5a (98.7 kcal/mol), is explained by the high energy required to deform the (thf)Li−H group from linear in the free molecule to a highly bent (120°) geometry in 4·thf′ as well as to modify the second coordination sphere of the Mo atoms to make room for the Li−thf moiety.
Having successfully built Mo 2 LiH 2 and Mo 2 Li 2 H 4 platforms based on Mo≣Mo bonds coordinated to one and two H− Li(thf)−H units, respectively, our next goal was to explore the possibility of reaching a Mo 2 Li 3 H 6 organization in a purported [Mo 2 {HLi(thf)H} 3 (μ-Ad Dipp2 )] complex. To this end and taking into account the successful synthesis of complexes 3·thf and 4·thf by the procedure shown in Scheme 1c, we prepared tris(acetate) precursor [Mo 2 (μ-O 2 CMe) 3 (μ-Ad Dipp2 )] and performed its reaction with an excess of LiAlH 4 . Although the above Mo 2 Li 3 H 6 complex could not be observed, the transformation led to complex 5·thf, identified as a polymetallic hydride cluster Mo 6 Li 9 H 18 (Scheme 2), that probably results from spontaneous trimerization of the targeted Mo 2 Li 3 H 6 monomer, with the loss of a molecule of tetrahydrofuran. The reaction was, however, complex and gave in addition compound [Mo 2 (μ-O 2 CMe) 2 (μ-Ad Dipp2 ) 2 ] through an undisclosed reaction path. Like the bis(formate) analogue (Scheme 1c), the latter may react further with LiAlH 4 , justifying that isolated yields of 5·thf are about 25%. Complex 5·thf is very air-sensitive and decomposes instantly in the presence of oxygen and water, both in solution and in the solid state. Under strict anaerobic conditions, solutions in tetrahydrofuran or aromatic hydrocarbons remain unchanged at 25°C for at least 24 h, although decomposition is fast above 50°C.
The new supramolecular entity can be understood as a triangular array of [Mo 2 (μ-Ad Dipp2 )] 3+ components 51,57,58 connected by a [Li 9 H 18 ] 9− linker in a fairly robust manner. The Li-coordinated molecules of tetrahydrofuran were readily substituted by pyridine and 4-dimethylaminopyridine, giving complexes 5·py and 5·dmap without the alteration of the molecular skeleton. Notwithstanding the foregoing, complex 5· thf acted as an efficient source of LiH in the hydrolithiation of Ph 2 C(O) to give Ph 2 C(H)(OLi). 20,22 Somewhat unexpectedly, solutions of 5·thf decomposed gradually upon stirring at room temperature under an atmosphere of H 2 , generating LiAd Dipp2 as a byproduct. Dideuterium acted similarly and showed that H/D exchange took place, as attested to by NMR detection of HD along with H 2 . The H 2 -promoted cluster breakup was not investigated any further. Nevertheless, it seems plausible that H 2 may disrupt the cluster structure by displacing LiH molecules from the [Li 9 H 18 ] 9− linker, eliminating LiAd Dipp2 . As an extension of these studies, various attempts were made to produce an alleged {Mo 2 (H) 8 [Li(thf)] 4 } complex (i.e., the hydride analogue of known methyl compound {Mo 2 (CH 3 ) 8 [Li(OEt 2 )] 4 }). 59 As detailed in the SI, all essayed trials were unsuccessful.
The room-temperature 1 H NMR spectra of complexes 5·L in C 6 D 6 or thf-d 8 solution show two septets and four doublets for the 12 isopropyl groups of the amidinate spectator ligands, in accordance with the proposed D 3 molecular symmetry (details in Supporting Information). The 18 H atoms that make up the [Li 9 H 18 ] linker are expected to give rise to three resonances of equal relative intensity. Whereas for 5·thf one of these signals seems to be hidden underneath other resonances, the three are clearly observed for complex 5·py with chemical shifts of 2.04, 5.21, and 5.41 ppm. They appear as broad multiplets, but while the 2.04 peak becomes a singlet in the 1 H{ 7 Li} NMR spectrum, the other two are converted to doublets with 2 J HH = 4 Hz. The 7 Li NMR spectrum contains three resonances centered at 5.4, 4.7, and 2.7 ppm, with relative intensities approaching roughly 6:2:1, once more in agreement with the proposed structure.
The molecular structure of complex 5·thf was determined by X-ray crystallography ( Figure 6) and computational studies. Since the calculated and experimental structures are very Scheme 2. Formation of Hexamolybdenum Nonalithium Dodecaoctahydride Cluster 5·thf  (Table 1), all of the features that are discussed here based on the X-ray data also apply to the optimized geometry. The whole cluster is built up by successive concentric groups around a central Li 3 unit (Figure 7, left) formed by Li7, Li9, and Li8 with a nearly linear arrangement (176(1)°) and distances of 2.50(2) and 2.45(2) Å, which are slightly shorter than twice the lithium covalent radius (2.56 Å). 55 We have been unable to locate a solid-state or gas-phase structure in which such a Li 3 rod is present. The only Li 3 group whose structure we are aware of appears in the crystal structure of  (7) Å is coherent with 4-fold bonding. 51 Leaving aside the Li atoms, an ionic description of the cluster leaves us with three [Mo 2 (μ-Ad Dipp2 )H 6 ] 3− blocks, in which each Mo atom bears two cis hydrides and one trans hydride relative to the N atoms of the μ-Ad Dipp2 ligand. The latter have just been described as forming an H 6 octahedron around the inner Li 3 rod and being bonded to the three Mo 2 units as well. The 12 cis hydrides can be described as distorted trigonal prisms, one with the trigonal faces roughly at the height of the external atoms of the central Li 3