Isolation of a Homoleptic Non-oxo Mo(V) Alkoxide Complex: Synthesis, Structure, and Electronic Properties of Penta-tert-Butoxymolybdenum

Treatment of [MoCl4(THF)2] with MOtBu (M = Na, Li) does not result in simple metathetic ligand exchange but entails disproportionation with formation of the well-known dinuclear complex [(tBuO)3Mo≡Mo(OtBu)3] and a new paramagnetic compound, [Mo(OtBu)5]. This particular five-coordinate species is the first monomeric, homoleptic, all-oxygen-ligated but non-oxo 4d1 Mo(V) complex known to date; as such, it proves that the dominance of the Mo=O group over (high-valent) molybdenum chemistry can be challenged. [Mo(OtBu)5] was characterized in detail by a combined experimental/computational approach using X-ray diffraction; UV/vis, MCD, IR, EPR, and NMR spectroscopy; and quantum chemistry. The recorded data confirm a Jahn–Teller distortion of the structure, as befitting a d1 species, and show that the complex undergoes Berry pseudorotation. The alkoxide ligands render the disproportionation reaction, leading the formation of [Mo(OtBu)5] to be particularly facile, even though the parent complex [MoCl4(THF)2] itself was also found to be intrinsically unstable; remarkably, this substrate converts into a crystalline material, in which the newly formed Mo(III) and Mo(V) products cohabitate the same unit cell.


■ INTRODUCTION
Any systematic exploration of the prodigiously rich (bio)inorganic, organometallic, organic, catalysis, and material chemistry of molybdenum has to cope with a surprising dearth of practical entry points. MoCl 5 is one of them, because this compound is available in bulk quantities at fairly low cost. 1 In solid form, MoCl 5 is manageable, 2 despite the rather aggressive chemical character that it draws from the combination of strong Lewis acidity, exceptional oxophilicity, and a pronounced oxidizing power; moreover, MoCl 5 is an effective chloride and/or chlorine source. Actually, only few of the commonly used solvents are inert toward MoCl 5 : benzene and related aromatic hydrocarbons succumb to oligomerization and/or chlorination; 3 ethereal solvents face rapid cleavage with concomitant formation of oxo-molybdenum species such as 2 (THF, dioxanes, hexamethyldisiloxane) (Scheme 1) and/ or serve as single electron reducing agents, entailing the release of Cl 2 (Me 2 O, DME); 4−7 acetonitrile also leads to reduction with formation of [MoCl 4 (MeCN) 2 ] and chlorinated acetonitrile byproducts. 8 These examples highlight the vigorous reactivity of MoCl 5 , which can be intentionally used for synthetic purposes, such as the oxidative coupling of (electronrich) arenes or the formation of esters by catalytic acylative ether cleavage. 7,9 Ordinary alcohols do not withstand this powerful oxidant and strong Lewis acid either: even primary alcohols R−OH are readily converted into the corresponding alkyl chlorides R−Cl (Scheme 1). The reaction is thought to proceed by partial ligand exchange with formation of transient Mo (V) alkoxides (e.g., 3) followed by rupture of a MoO−R bond; the resulting oxo-molybdenum complexes of type [OMoCl 3 ] (4) dimerize and can lead to adducts such as 5 or even complex polynuclear arrays. 10−13 The ease of formation of the [Mo O] group manifest in these examples is a hallmark of (highvalent) molybdenum chemistry; 10,14,15 this prevalent functionality dominates the field and plays a pivotal role in biological and industrial catalysis alike. 16,17 Only highly fluorinated alcohols were found to subsist to such degradation: trifluoroethanol, for example, on treatment with MoCl 5 , affords the stable heteroleptic complex 6. 18,19 The fact that even the electron-poor O atom of trifluoroethoxide serves as a bridging ligand highlights the very strong bias for dimerization, which is yet another characteristic trait of the coordination chemistry of molybdenum. A closely related binuclear molybdenum (V) alkoxide 7 was prepared by salt metathesis between MoCl 5 and NaOMe, but complete ligand exchange could not be accomplished either. 20,21 Analogous reactions with electronrich alkoxides are unlikely in view of the ease of reduction of MoCl 5 even by much milder agents; 22 in any case, no followup investigation describing the preparation of analogues of 7 has been published.
It is against this backdrop that the isolation and full characterization of [Mo(OtBu) 5 ] (1) must be seen: even oxofree alkoxides of Mo(V) with a mixed ligand sphere, such as 6 and 7, are exceedingly rare chemical entities, but homoleptic alkoxides of Mo(V) are elusive and may perhaps even be deemed inaccessible. It is therefore perplexing that complex 1 as the first incarnation of this previously unknown class of compounds carries tertiary alkyl residues on all oxygen atoms: a priori, these ligands are particularly prone to [MoO] formation, independent of whether the actual O−C bond cleavage proceeds in a heterolytic or homolytic manner with release of a stabilized tertiary carbocation or tertiary radical, respectively; 7,23 most notably, the clean and efficient radical breakdown of tert-butoxide at a Mo(V) center has been explicitly mentioned in the literature. 24 The fact that 1 is monomeric even in the solid state is equally striking if one considers that a d 1 electron count is potentially conducive to metal−metal bonding. 25,26 Moreover, the sheer size of a tBuO group certainly does not exempt it from serving as a bridging ligand; 27 in general, the formation of μ-bridged dimers (oligomers) is so favorable that even very poorly donating fluorinated alkoxides do not get away (cf. 6). Finally, the synthesis of 1 by disproportionation tells fundamental lessons about the meta-stability of low-valent molybdenum complexes in general and, in doing so, casts doubts on previous reports claiming the isolation of putative tetra-valent [Mo-(OtBu) 4 ]. 28 42−44 and a new, green, highly air-sensitive and paramagnetic species, which was isolated by taking advantage of its high solubility in pentane; purification of the crude product by sublimation in high vacuum (23°C, 10 −7 mbar) furnished complex 1 in 30% yield (Scheme 2). It is emphasized that formation of 1 in appreciable amounts was observed only under the specified conditions: changing the stoichiometry of the reagents and/or the solvent (THF, npentane) resulted in lower yields (see the Supporting Information).
Compound 1 is well-soluble in n-pentane, benzene, toluene, and CH 2 Cl 2 but rapidly decomposes in MeCN at room temperature; it also degrades in [D 8 ]toluene when the solution is warmed to ≥50°C. 45 Mo 20.58). The NMR spectroscopic fingerprint of this paramagnetic species (broad signals in C 6 D 6 at δ H = 7.43 ppm and δ C = 54.9, 30.3 ppm) is also in accord with the proposed homoleptic structure. This assignment was confirmed when single crystals suitable for X-ray diffraction were obtained from a saturated solution in npentane at −35°C.
The unit cell contains two independent molecules, which differ from each other only in conformational detail [for details, see the Supporting Information (SI)]. 46 As can be seen from Figure 1, the coordination geometry about the Mo center can be described as approximately trigonal bipyramidal, although notable deviations from the idealized structure are on record: specifically, the O2−Mo1−O2′ angle is only 169.1(5)°rather than 180°, and the angles between the O atoms forming the equatorial plane are uneven and irregular As will be discussed in detail below, the overall distorted structure of 1 results from a symmetry-lowering Jahn−Teller distortion of the nuclear framework, befitting this d 1 complex (see below). 47 The fairly clean formation of 1 by disproportionation of [MoCl 4 (THF) 2 ] in the presence of NaOtBu at low temperature was unexpected in view of two earlier literature reports. Specifically, it had been reported that treatment of [MoCl 4 (THF) 2 ] with LiOtBu in THF/hexane entails simple ligand exchange; a workup with CH 2 Cl 2 followed by recrystallization from cold n-pentane supposedly gave [Mo-(OtBu) 4 ] in 25% yield. 29 This assignment, however, had been based on a rather rudimentary characterization by elemental analysis and the fact that a single resonance in the 1 H NMR spectra was observed; the compound was reported to be EPRsilent. 29 In our hands, treatment of [MoCl 4 (THF) 2 ] with either NaOtBu or LiOtBu invariably resulted in disproportionation with formation of 8 48 and 1; the latter is EPR-active (see below). Unfortunately, the paucity of data reported for [Mo(OtBu) 4 ] precludes a detailed comparison with 1 and an accurate assessment. We cannot exclude that [Mo(OtBu) 4 ] had previously been formed by ligand exchange, 29 yet we are also unable to confirm it, despite considerable experimentation.
A second report describes [Mo(OtBu) 4 ] as a thermally unstable green-brown solid material formed on reaction of [Mo(NMe 2 ) 4 ], with tBuOH. 28 Its characterization was based upon elemental analysis, IR, and mass spectrometry: it is stunning, however, that the peak in the reported mass spectrum with the highest m/z 461 actually fits to [Mo-(OtBu) 5 ] + , 49 although the observed base-peak at m/z 388 matches the mass of [Mo(OtBu) 4 ] + . 28 The reported IR data have not been analyzed in any detail and therefore do not allow one to make a final judgment; we note that the reported bands are close to those observed for 1 (see below). In consideration thereof, we tried to reproduce this literature route (Scheme 3). Although it cannot be rigorously excluded that [Mo(OtBu) 4 ] is present in the crude mixture, the only complexes that we were able to isolate in pure form in several independent runs were, once again, [Mo(OtBu) 5 ] (1) and the new dinuclear species 9. From the structure in the solid state it is apparent that 9 contains a bridging amide and a bridging oxo-ligand; moreover, a molecule of Me 2 NH released upon reaction of [Mo(NMe 2 ) 4 ] with tBuOH now serves as a donor ligand to one the Mo centers ( Figure 2). The short Mo1−Mo2 distance [2.486(3) Å] speaks for a metal−metal bonding interaction within the core of this paramagnetic species. 50 The question as to how this peculiar product is formed has to remain open at this point. The fact that the μ-oxo atom can only derive from a tBuO precursor ligand could imply an oxidative cleavage mechanism, which might transform a transient dinuclear Mo(III) precursor as the expected low-valent product of the actual disproportionation reaction into the ultimately isolated complex 9.
Collectively, these results suggest that alkoxide ligands favor the disproportionation of Mo(IV), even though a discrete and well-characterized homoleptic Mo(IV) enolate complex is known that has been made from [MoCl 4 (THF) 2 ] by ligand exchange with the corresponding bulky alkali-metal enolate. 51 [MoCl 4 (THF) 2 ] itself also shows the propensity to disproportionate, although more latently. 52 This complex had previously been recognized as unstable under nitrogen atmosphere, but the decomposition products had not been identified. 8b When a solution of [MoCl 4 (THF) 2 ] in dichloromethane was kept at ambient temperature for ≥4 d, a solid material started to precipitate that is composed of a 1:1 mixture of the oxo-species 10, crystallized with one THF and one adventitious water ligand, 53 and the known dinuclear complex 11 (Scheme 4). 54 The coexistence of these Mo (V) and Mo(III) complexes in a single unit cell is unprecedented ( Figure 3); 55 it provides compelling evidence for the notion that Mo(IV) is inherently  In any case, [Mo(OtBu) 5 ] (1) is the prototype of a new class of high-valent molybdenum complexes, the existence of which is rather counterintuitive; such monomeric, homoleptic, five-coordinate, all-oxygen-ligated but non-oxo 4d 1 Mo(V) species were previously unknown. For this unique status, it was deemed appropriate to scrutinize the structure and bonding of this new chemical entity in more detail by a combined spectroscopic and theoretical approach.
Computational Study: Structure. The geometry-optimized structure 1a of [Mo(OtBu) 5 ] (1) starting from the crystal structure geometry is shown in Figure 4. The computed structure (see the Supporting Information for computational details) is intermediate between square pyramidal and trigonal bipyramidal, thus rendering all O atoms inequivalent. The distortion from trigonal bipyramidal and square pyramidal can be quantified by a Berry pseudorotation, where the O3−Mo− O1 angle has decreased from 180°(square pyramidal limit) to 135.7°. Since the complex is homoleptic, multiple minima likely exist on the potential energy surface; 47 as such, the system is Jahn−Teller-active.
A comparison between the crystal structure and the optimized geometry based on the crystal structure shows a good agreement (Table 1). The axial direction of the trigonal bipyramid is given by the O2−Mo and/or Mo−O2′ direction (O2−Mo−O2′ = 160.6°). The low symmetry of the molecule also leads to an inequivalence of the x and y directions. This geometrical aspect is indeed reflected, for example, in the observation of a rhombic EPR spectra (see below).
In addition, in an effort to explore the potential energy surface in more depth, additional geometry optimizations were performed with modified starting structures, by rotating the tBu groups about the Mo−O direction in various permutations. While some calculations ended up converging to identical minima, a range of stable, slightly different conformations have indeed been found, indicative of a potential energy surface with multiple close-lying minima. Figure 4 features two additional optimized structures (1b, 1c) that clearly show structural differences. The geometry optimization starting from the crystal structure 1a gave rise to the most square pyramidal structure. The middle and rightmost structures are significantly more trigonal bipyramidal. Structure 1b turned out to be the most stable structure in vacuo. However, the energy difference with respect to 1a and 1c is only 2.5 and 1.6 kcal/mol, respectively, which is well within the accuracy of the employed DFT methodology; as such, the observation that structure 1b is most stable with the present choice of functional, method, and basis set should not be overinterpreted. Rather, these calculations should be viewed of as an in silico confirmation of the presence of multiple, closely related local minima that differ by rotation of the tBu groups about the respective Mo−O bond directions.
Given the small energy differences of the structures in Figure  4, it is by all means conceivable that local minima become thermally populated at elevated temperature and depopulated at low cryogenic temperature, which is in-line with the observed temperature dependence of the EPR signal (vide infra). Moreover, when the tBu groups rotate about the Mo− O direction, the hybridization of the oxygen orbitals, and thus the covalency of the Mo−O bonds, is expected to be altered, resulting in changed Mo−O covalencies. This in turn may lead to a continuous or discrete distribution in spectroscopic parameters like EPR g-values and chemical shifts, in the ligand field splitting parameters and transition energies observable in MCD, and even in the powder diffraction pattern.
Powder Diffraction. In order to investigate possible structural heterogeneities experimentally, we have recorded powder diffraction patterns of several batches. The results, shown in Figure 5, indeed confirm that a discrete structural heterogeneity may be present: whereas the measured pattern for the batch shown in Figure 5b matches the expected pattern calculated from the crystal structure reasonably well (Figure   Journal of the American Chemical Society pubs.acs.org/JACS Article 5a), the pattern of a second batch in Figure 5c shows a splitting of the band at 5.5°. Moreover, this sample also seems to contain some dimer, [Mo 2 (OtBu) 6 ]. While the powder pattern clearly confirms the presence of two discrete conformers, the measured data does not contain enough information to allow a confident identification of the second structure of the monomer. Electronic Structure. In spite of the significant Berry rotation of the optimized crystal structure 1a, the computed dorbital scheme of 1a shown in Figure 6 closely resembles that  represented by the geometry-optimized crystal structure 1a. Also included are four representative oxygen-based doubly occupied orbitals (at about −6.8 eV). Since the orbital structure largely corresponds to that of a square pyramid, the molecular z-axis is chosen as parallel to the Mo−O3′ direction and the x-axis as along the O2−O2′ direction. Symmetry labels under approximate C 2v symmetry were added accordingly for the orbitals and for the transitions; the tBu groups are omitted for clarity. Upon consideration of the ligand character, the low-lying y-polarized 2 B 2 and x-polarized 2 B 1 d−d transitions (blue) are expected to gain oscillator strength owing to an overlap of the ligand part of the donor orbital with the metal part of the acceptor orbital, leading to a symmetry-allowed, nonzero transition dipole moment upon admixture of ligand character into the orbital. Electronic excitations from the oxygencentered doubly occupied orbitals with approximate C 2v symmetry labels are included in the dashed boxes.  Journal of the American Chemical Society pubs.acs.org/JACS Article expected for a square pyramidal coordination geometry. The unpaired electron is located in the nonbonding d xy orbital. The other orbitals are destabilized owing to the metal−ligand interactions: specifically, the SOMO is followed by the d xz /d yz pair, the degeneracy of which has been lifted by the Jahn− Teller distortion. Next comes the d z 2 orbital and finally, as the most destabilized orbital, d x 2 −y 2 . The symmetry labels of the orbitals under approximate C 2v symmetry (vide infra) are included in Figure 6. Inspection of the doubly occupied orbital structure reveals a plethora of "nonbonding" linear combinations of oxygen-centered orbitals (the first four are included in Figure 6) that do not directly overlap with the Mo 4d orbitals owing to symmetry. This set of orbitals occurs at about 3.5 eV (28 000 cm −1 ) in energy below the SOMO. The first oxygenbased orbital with bonding character to Mo occurs even lower, at −8.50 eV, in which the oxygen orbitals are π-bonding to the 4d xy orbital (not shown in the figure). While it is in principle possible to induce a z-polarized 2 A 1 charge-transfer transition from this orbital into the SOMO, this electronic transition is expected at an energy above 40 000 cm −1 , where it would be difficult to detect. A comparison of the electronic differences in terms of ligand field splittings of complexes 1a−1c is given in the Supporting Information. UV/Vis and MCD Spectroscopy. The UV/vis spectrum displays a broad and relatively featureless absorption in the blue part of the spectrum (Figure 7 Figure S9 (SI) and vide infra]. Thus, this observation provides additional evidence for the coexistence of several conformers under the experimental conditions. Attempts to address the structural heterogeneity by additional MCD experiments of deposited and smeared-out monocrystalline material were not successful due to scattering, as were experiments with frozen solutions, since n-pentane and toluene or mixtures thereof produced a poor glass at low temperature. The calculations shown in Figure 7 will be discussed in the Theoretical Spectroscopy section.
EPR Spectroscopy. The Q-band continuous wave (cw) EPR spectrum (Figure 8) has been well-fitted by using a minimalistic spin Hamiltonian that included the three g-values and the Mo hyperfine coupling constants, as well as three line width parameters. The fitted g-values amount to 1.818, 1.903, and 1.936. The observed rhombicity of the g-tensor provides direct confirmation of the symmetry-lowering of [Mo(OtBu) 5 ] (1) as compared to the structure of this complex in the solid state. It is worth noting that the g x feature of the EPR spectrum shows a splitting in two poorly resolved bands. A comparison of spectra recorded at the X-band and Q-band (Supporting Information) indicates that the two bands are not caused by hyperfine splitting and must stem from slightly different conformers present in frozen solution, with possibly slightly altered orientation of the tBu groups and thereby slightly changed Mo−O covalencies and thus g-values.
Being aware of possible structural heterogeneity, we investigated the EPR spectrum as a function of temperature. The temperature dependence in Figure 8 (right) measured at the X-band displays a drastic change of the spectrum upon raising the temperature from 30 to 100 K, in the form of the disappearance of the intense signal at 360 mT and the merging together of the g z feature at 380 mT. The change is completely reversible, as subsequent cooling of the sample back to 30 K recovered the original signal. The spectrum at 4.2 K is essentially identical to the one at 30 K. The change in shape of the signal with temperature and, in particular, the presence of a split g z signal at low temperature provide a rather strong experimental indication of the presence of two conformers corresponding to local minima of the potential energy surface of the Jahn−Teller-active molecule, the thermal populations of which change.
Theoretical Spectroscopy. In order to interpret the recorded spectra, a detailed comparison with quantum chemical calculations is necessary. A particular problem occurs in interpreting the six bands in the MCD spectrum ( Figure 7). First, the AILFT calculation ( Figure S9, SI) that included the five 4d orbitals and one electron features only three d−d transitions below 20 000 cm −1 . The natural transition orbitals for the lowest two transitions are included in Figure 7. These are the 1-2 B 2 and 1-2 B 1 transitions featuring the d xy orbital as the donor orbital and the d xz , d yz pair as the acceptor orbitals. Out of all four d−d transitions these two are expected to have the largest oscillator strength. The oxygen-centered nonbonding orbitals were not included in the CAS calculation, because they are located more than 3.5 eV below the SOMO.
The seemingly doubled number of bands in the MCD spectrum, in addition to the observed splitting of the g z feature in the EPR spectrum ( Figure 8) and the splitting observed at 5.5°for one of the batches in the powder diffraction pattern (Figure 5c), is compatible with the presence of two conformers with perhaps slightly changed orientations of the OtBu groups and, concomitantly, slightly changed Mo−O covalencies. This is additionally confirmed by the calculations of the electronic spectrum, where structures 1a, 1b, and 1c each display two bands in the near-infrared region at different positions ( Figure  7, bottom, right), which is a direct consequence of the slightly changed ligand field splitting of the conformers ( Figure S9, SI). While a definite assignment of the bands in the MCD spectrum seems presently not feasible owing to the Jahn−Teller-induced heterogeneity, the mere observation of these bands in the near-IR region shows unambiguously that the molecular species does not feature an oxo-ligand. An oxo-ligand, in turn, would lead to an MCD spectrum where the lowest d−d transition would be significantly higher in energy, e.g., 20 000 cm −1 in [Mo(O)Cl 3 ·(dppe)]) 56 or larger than 23 000 cm −1 in a derived in silico [MoO(OtBu) 4 ] − complex contained in the Supporting Information ( Figure S7).
Analysis of the g-values in terms of ligand field theory (see the Supporting Information) yielded good agreement with experiment. However, a confident, detailed assignment of all low-energy MCD bands could not be achieved. The calculated g-values for structure 1a amount to 1.87, 1.94, and 1.95, in reasonable agreement with experiment, whereby in particular the middle g-shift is calculated slightly too small. The Journal of the American Chemical Society pubs.acs.org/JACS Article calculated Loẅdin spin population at Mo amounted to 85%, and the calculation was largely spin-uncontaminated (⟨S 2 ⟩ = 0.754). The more-trigonal geometries 1b and 1c found on the potential energy surface gave rise to slightly different g-values, typically changed by about 0.02 (Table S7, SI). It is instructive to investigate the observed splitting of the least-shifted g-value, g z . In perturbation theory, the g z shift in structure 1a arises from the matrix elements of spin−orbit coupling and the orbit−Zeeman interaction between the (xy) 1 ground state and the (x 2 −y 2 ) 1 excited state [3a 1 , ΔE(xy,x 2 −y 2 ) = 34 540 cm −1 ]. In structures 1b and 1c, the least-shifted gvalue occurs between the (xz) 1 ground state and contributions of both the (z 2 ) 1 (1a′) and the (x 2 −y 2 ) 1 (1e″) excited states. The excitation energy of the 1e″ contribution changes from 17 667 to 19 883 cm −1 upon going from 1b to 1c (cf. Figure  S9, SI), i.e., a change of 12.5%, which would lead to a change of the g-values of up to 0.008, compatible with experiments and thus providing yet another confirmation of the Jahn−Teller heterogeneity between the C 2v and C 3h structures.
Subsequently, a TDDFT calculation has been performed in order to assign bands 7−10 in the UV/vis spectrum shown in Figure 7. The agreement between experiment and theory is good for structure 1a, providing additional confidence that the chosen model accurately mimics the actual structure of [Mo(OtBu) 5 ] (1). Structures 1b and 1c, although essentially equal in energy to structure 1a, seem to provide a less accurate description of the UV/vis spectrum. The plethora of individual transitions (depicted as sticks) included in the diagram gives information that a large number of transitions contribute to the Gaussians used in the deconvolution. As already noted, the large number of transitions originates from the presence of a large number of ligand-centered orbitals, mainly with oxygen character, that are located about 3.5 eV below the SOMO. Difference densities for representative sticks are included in Figure 7. Although the relative phases of the oxygen orbitals inhibit admixture of Mo(4d) character, all of these orbitals principally contribute to the rich manifold of excited states that are all formally of oxygen-to-molybdenum charge-transfer character. The more intense transitions involve the B 1 and B 2 transitions of the 1a 2 , 1b 2 , 1a 1 , and 1b 1 orbitals as donor orbitals (cf. Figure 6).
Most convincingly, the calculated IR spectra for all three model structures included in Figure S11 (SI) reproduce virtually all bands observed in the experimental spectrum very nicely and hence allows for their assignment (see the Supporting Information).
Overall, we conclude that despite the observed complications owing to temperature-dependent structural heterogeneity, generally excellent agreement between experiment and theory is found for all spectroscopic methods employed in this study. This provides confidence in the model structures used in the calculations. The calculated geometric and electronic structure therefore accurately represents the so far unique homoleptic Mo(V) alkoxide [Mo(OtBu) 5 ] (1), including the symmetry-lowering as a result of Berry rotation.

■ CONCLUSION
The chemistry of Mo(V) is dominated by oxo-complexes, which are readily formed by formal oxygen atom abstraction even from substrates as stable as ethers, phosphine oxides, sulfoxides, and, most commonly, alcohols. Due to the proclivity to form and maintain [MoO] groups, Mo(V)alkoxides devoid of at least one additional oxo-ligand in general are exceedingly rare, and homoleptic representatives were entirely unknown prior to the present study. It is now shown that disproportionation of Mo(IV) provides access to this previously elusive class of compounds under notably mild conditions. The first embodiment is [Mo(OtBu) 5 ] (1). This compound is a monomeric entity with a Jahn−Teller-distorted trigonal bipyramidal structure in the solid state but is subject to facile symmetry-lowering Berry pseudorotation in solution. The geometric and electronic structures of this unique complex are accurately described by DFT, which was calibrated against experimental spectra (IR, UV/vis, MCD, EPR). 57 Indirect evidence suggests that alkoxide ligands render the disproportionation reaction of Mo(IV), leading to the formation of [Mo(OtBu) 5 ], remarkably facile, even though [MoCl 4 (THF) 2 ] was also found to be intrinsically unstable toward decay into Mo(III) and Mo (V) in CH 2 Cl 2 solution at ambient temperature. The question whether the increased reaction rate in the presence of tert-butoxide has to do with the formation of the dinuclear complex [(tBuO) 3 MoMo-(OtBu) 3 ] as a particularly favorable low-valent product of the disproportionation process or if stabilization of [Mo-(OtBu) 5 ] itself by dispersive forces within the ligand sphere 58,59 plays any significant role will be the subject of future studies.