Heterolytic Si−H Bond Cleavage at a Molybdenum‐Oxido‐Based Lewis Pair

Abstract The reaction of a molybdenum(VI) oxido imido complex with the strong Lewis acid B(C6F5)3 gave access to the Lewis adduct [Mo{OB(C6F5)3}(NtBu)L2] featuring reversible B−O bonding in solution. The resulting frustrated Lewis pair (FLP)‐like reactivity is reflected by the compound's ability to heterolytically cleave Si−H bonds, leading to a clean formation of the novel cationic MoVI species 3 a (R=Et) and 3 b (R=Ph) of the general formula [Mo(OSiR3)(NtBu)L2][HB(C6F5)3]. These compounds possess properties highly unusual for molybdenum d0 species such as an intensive, charge‐transfer‐based color as well as a reversible redox couple at very low potentials, both dependent on the silane used. Single‐crystal X‐ray diffraction analyses of 2 and 4 b, a derivative of 3 b featuring the [FB(C6F5)3]− anion, picture the stepwise elongation of the Mo=O bond, leading to a large increase in the electrophilicity of the metal center. The reaction of 3 a and 3 b with benzaldehyde allowed for the regeneration of compound 2 by hydrosilylation of the benzaldehyde. NMR spectroscopy suggested an unusual mechanism for the transformation, involving a substrate insertion in the B−H bond of the borohydride anion.


Introduction
The chemistry of frustrated Lewis pairs (FLPs) has received large attention in recent years, with as ignificant spark being the discovery of metal free hydrogen splitting in 2006 by Stephan and co-workers. [1,2] The term FLP refers to Lewis acidbase pairs that contain unquenched acidic and basic centers. [3] Usually,t hese Lewis acid-base pairs combine to aL ewis adduct by forming ac hemical bond between the Lewis functionalities. The deliberate introduction of steric hindrance into bimolecular Lewis pairs or the introduction of spatially separated Lewisa cid and base functionalities in one molecule (an ambiphilicc ompound), however,r esults in the formation of FLPs. [2][3][4] The corresponding chemistry has evolvedi nto much more than mere curiosity,a sc ountlessr eports of small molecule activations andc atalytic applicationsh ave been given, with the metal-freeh ydrogenation of olefins being just one prominente xample. [1,[3][4][5][6][7][8][9][10] The present understanding of FLP chemistry was notably influenced by findings made several years earlier in the investigation of borane-mediated hydrosilylation reactions by Piers and co-workers. [11,12] In recenty ears, the combination of unquenched Lewis pair reactivity and transition-metal chemistry has received considerable attention,b yd eveloping transition-metal-based FLPs. [6,[13][14][15][16] The benefitso fu sing transition metals include increased reactivity in small molecule activation, reactivity not observed with main group frustrated Lewis pairs and the advantage of an easy variation of the FLP components, for example, by ligand modifications. [6,13,17] High valent metal oxido fragments usually reacta se lectrophiles, asf or example observed in oxygen-atom transfer reactions to Lewis basic phosphanes. [18][19][20][21][22][23][24] Nevertheless, there have been reports of metal oxides exhibiting ambiphilic reactivity. [25,26] The inherent nucleophilicity of M=Of ragments has been showcased by Green and co-workers, who reactedv arious transitionm etal oxidesw ith B(C 6 F 5 ) 3 to obtain the corresponding Lewis adducts. However,t hey restricted their investigationst ot he thorough characterization of the compounds, while reactivity studies were not performed. [27][28][29] On the other hand, examples of transition-metal FLPs involving the versatile metal oxido functionality are surprisingly scarce. Very recently, Ison andc o-workers disclosed the capability of rhenium(V) oxido based Lewis adducts to catalytically hydrogenate alkenes, an unprecedented behaviorf or high valent metal oxido compounds. [30,31] The postulated mechanism is based on the formation of a FLP with the metal oxido group as the Lewis basic component (Scheme 1), corroborated by mechanistics tudies with H 2 /D 2 mixtures that revealed isotopic scrambling (formation of HD) under catalytic conditions. [30] FLPs in equilibrium with the corresponding classical Lewis adducts can be seen as quasi-frustrated Lewis pairs, underlin-ing that stable Lewis acid-base adducts can exhibit FLP-like reactivity under certain conditions. [1,7,9,16,30,32] For several years we have used molybdenuma sa ne arth abundant and non-toxic transitionm etal [33] and develop high oxidation state molybdenumc ompounds, primarily for oxygen activation,o xygen-atom-transfer and oxidation reactions. [18-20, 23, 34-36] With this in mind, we are exploring new concepts for the activation of metal peroxidoa nd metal oxido moieties, with the latter being important intermediates in catalytic aerobic oxidation reactions, to increase the reactivity of this notoriously inert groups. [23,36] Ap romising approach seems to us the addition of as trong Lewis acid for the intermoleculara ctivation. For am etal oxido moiety,the coordination of aLewis acid shouldlead to aweakening of the M=Ob ond thus enabling novel reactivity,a dditionally aided by the prospect of FLP formation. The feasibility of the concept has very recently been demonstrated in dinitrogen complexes in whichaLewis acid led to the activation of the inert N 2 molecule. [37] Furthere xamples of reactivity enhancement upon coordination of B(C 6 F 5 ) 3 at metal oxido moieties have been reported by the groups of Schrock [38] and Ison. [39] Catalytic hydrosilylationo fo rganic carbonyl functions using B(C 6 F 5 ) 3 has been intensively investigated. [12,[40][41][42] In fact, the general mechanism of B(C 6 F 5 ) 3 -catalyzed hydrosilylation, elucidated by the work of Piers and Oestreich, is based on the formation of aF LP with the substrate. [11,12,[40][41][42][43][44] In addition, some transition metal oxido complexes are knownt ob ea ctive catalysts in hydrosilylation reactions. [45] While av ast number of transition metal compounds are known to react with hydrosilanes, [46] only few examples of well-characterizeds peciesresulting from the reaction of hydrosilanes with metal oxido groups have been disclosed by Abu-Omar, [47] Hayton [48,49] and To ste. [50] For this reason, we found hydrosilanes as uitable choice as test reagentst oa ssess the envisioned Lewisa cid activation of the molybdenum(VI) oxido group.
Herein we report the cleavage of SiÀHb onds at an oxidomolybdenum/borane Lewis pair,a llowing fort he isolationo f novel cationic molybdenum(VI) silanolate species, showcasing the anticipated Lewis acid assisted increaseo fr eactivity.T he molybdenum(VI) oxido imido complex with B(C 6 F 5 ) 3 ,aclassical Lewis acid-base adduct with the Lewis acid exclusively bound to the oxido ligand,i sf ully characterized. However,r eactivity studies showed the BÀOb onding to be reversible, corroborating the formation of ar eactive Lewis pair in solution,c apable to heterolytically cleave SiÀHb onds. Subsequentr eactivity studies towardsb enzaldehyde uncovered an unusual substrate insertions tep into the borohydride bond under reformation of the Lewis adduct. The research presented here likewise offers the prospect of the templated transfer of not only silyl groups but also other electrophiles to avariety of substrates.

Results and Discussion
Lewis adduct synthesis Additiono fo ne equivalent of the Lewis acid B(C 6 F 5 ) 3 to the yellow solution of the molybdenum(VI) oxido imido precursor [MoO(NtBu)L 2 ]( 1) [35] in pentane led to an immediate color change to deep-red ands ubsequent formation of the Lewis adduct [Mo{OB(C 6 F 5 ) 3 }(NtBu)L 2 ]( 2)a sareddish precipitate. Compound 2 was isolated as ab rick-reds olid in very good yield after purification (Scheme2).
The slightly moisture-sensitive complex 2 is soluble in most polar organic solvents, sparingly soluble in benzene and toluene and practically insoluble in acetonitrile and DMSO.
Interestingly,N MR spectroscopy revealed compound 2 to exist as as ingle isomer in solution,w hich contrasts with the isomeric equilibrium observed for 1.W ea ttributet his not only to the increased steric demandi nt he coordination sphere, but also to the pronouncede lectron-withdrawing effect of the oxidoborane unit, leading to increased s-donation of the other ligands and thus decreased flexibility at the metal center.T he single isomer furtheri ndicates that the Lewis acid exclusively coordinates at one terminal ligand,t he Mo=Om oiety,a s shown by X-ray crystallography (vide infra). The coordination of the Lewis acid is confirmed by an ew set of signals correspondingt ot he meta, ortho and para fluorines of B(C 6 F 5 ) 3 ,r espectively,o bservable by 19 FNMR spectroscopy.T he pronounced shift of the para Fresonance (À161.9 ppm) compared to free borane (À142.0 ppm) is in agreement with literature. [25] Compound 2 was further characterizedv ia 11 BNMR spectroscopy,w hich features ab road singlet resonance at 2.5 ppm, indicative of adduct formation.
To assesst he stability of the BÀOb ond, that is, possible dissociation to form aF LP,t he reactivity of 2 towards donor solvents known to form adducts with free B(C 6 F 5 ) 3 ,s uch as THF or MeCN, was examined. Whereas addition of THF did not lead to ar eaction, addition of MeCN to aC 6 D 6 solution of 2 led to an equilibrium between 2 and B(C 6 F 5 ) 3 ·MeCN.T he ratio was found to be dependent on the added amount of MeCN (FigureS1, Supporting Information), demonstrating ar ather stable BÀO bond butnevertheless reversibility of the adduct formation.

Reactivity of 2t owards hydrosilanes
Complexes 3a and 3b were obtained in good yields after purification. They are very soluble in most polar and non-polar organic solvents and only sparingly soluble in alkanes. The formation of the diamagnetic cationic Mo VI specieso ft he general structure [Mo(OSiR 3 )(NtBu)L 2 ] + was confirmed by distinct signal sets for two iminophenolate, one imido, one silanolate ligand and ab road (1:1:1:1) quartet resonance for the [HB(C 6 F 5 ) 3 ] À anion in the 1 HNMR spectra. The structure of the anion is further confirmed by three new resonances in the 19 FNMR spectra, and ac haracteristicd oublet resonance in the 11 BNMR spectra,b othm atching literature. [8] Similart o2, 3a and 3b are present as as ingle isomer in solution,f ollowing the same reasoning presenteda bove.
It is interesting to note that in C 6 D 6 ,anon-polar solvent, 1 HNMR shifts of the Mo VI cation are highly dependento nt he concentration, while the 19 FNMR signals are only slightly affected.
In contrast, in ap olar solvent (CD 2 Cl 2 ), 1 HNMR resonances are not shiftedu pon change of concentration. This can be attributed to the increased solvation capability of the polar solvent and thusdecreased ionic interaction ( Figure 1).
The stability of complexes 3a and 3b in solution is limited as they are very sensitive towards moisture. Thus, 1 HNMR spectra of 3a or 3b,r espectively,i nt he presence of H 2 O, show an ew set of 1 HNMR signals, which can be attributed to HLH 2 , [51] the protonated ligand with ar educed C=Nm oiety,underscoring the reactivity of the hydridoborate anion. [14,[52][53][54] Infrared spectroscopy of 3a and 3b revealed no assignable Mo= Os tretch. However,ab road peak in the MoÀOr egion at 567 cm À1 (3a)a nd 568 cm À1 (3b)i ndicates overlap of the stretches of three MoÀObonds in total ( Figure S3).
In complexes 1 and 2 on the otherh and, as ingle sharp band was found at 544 and 546 cm À1 ,r espectively,a ssigned to the two phenolate MoÀOs tretches. The assignment of ana dditional MoÀOf requency is in good agreement with the bond lengthso btained via single-crystal X-ray diffraction analysis (vide infra). The infrareds pectra of 3a and 3b also feature a characteristic broad BÀHb and at approximately 2380 cm À1 , consistentwith the hydridoborate anion( Figure S2).
Formation of complexes[ Mo(OSiR 3 )(NtBu)L 2 ][HB(C 6 F 5 ) 3 ]( 3a and 3b)u pon reaction of 2 with R 3 SiH is reminiscent of the ionic intermediate in the FLP-based hydrosilylationm echanism formulated by Piers and Oestreich. [12,43] Am ajor differencei s the possibility of chargec ompensation by the metal in our case. The cationic molybdenum centeri slikely energeticallyf avored over ah ypothetic oxonium-based ion. AB (C 6 F 5 ) 3 assisted reactiono fametal oxide with ah ydrosilane has been disclosed previously for ah igh valent uranium system,h owever the reactionw as accompaniedb yt he reduction of the metal center and led to different productsw ith Et 3 SiH and Ph 3 SiH. [48,49] Scheme3.Formation of the cationic silanolate complexes 3a and 3b. In order to evaluate the strengtho ft he formed Mo-OSiR 3 bonds, s-bond metathesis with Me 3 SiCl was investigated. [55] Upon reaction of 3a with three equivalents of Me 3 SiCl, silyl group exchange was observedt of orm the complex [Mo(OSi-Me 3 )(NtBu)L 2 ][HB(C 6 F 5 ) 3 ]( 3c)r ather than metathesis to form (Me 3 Si) 2 Oa nd am olybdenum chlorido species (Scheme 4). This is in agreement with the higher average bond dissociation energy of MoÀO( 560 kJ mol À1 ) [56] in comparison to SiÀO (452 kJ mol À1 ). [57] The observed reactivity represents ap ractical way to synthesize the trimethylsilanolatec omplex 3c in high yield withoutt he need for the volatile Me 3 SiH. Nevertheless, (Me 3 Si) 2 Ow as frequently encountered as ab y-product in the reaction, af act that is attributedt ot he pronounced sensitivity towardswater.

Anion exchange
As above mentioned, complexes 3a-3c are very sensitive towards moisture, which is predominantly ar esult of the inherent reactivity of the hydridoborate anion,a sp reviously discussed in literature. [14,[52][53][54] Furthermore, 3a-3c were obtained as amorphous solidsf rom which X-ray quality crystalsc ould not be grown.
To stabilize the complexes and possibly enable crystallization, the exchange of the anion was investigated.
The use of [Ph 3 C][BF 4 ]s hould abstract the hydride from the hydridoborate and introduce the smaller and more stable anion [BF 4 ] À .T hus, reaction of complexes 3a and 3b,r espectively,w ith [Ph 3 C][BF 4 ]a tl ow temperatures afforded after work-up two new species 4a and 4b in good yields as purple solids. ProtonN MR spectroscopy revealed resonances that are virtually identicalt ot hose of 3a and 3b,r espectively,a lbeit the BÀHr esonance is absent, consistent with an anion exchange. Thea bsence of aB ÀHs tretch is also confirmed via in-frared spectroscopy. However,c areful examinationo f 19 FNMR data revealed an unexpected fluorination of the anion after hydride abstraction,l eading to the formation of complexes [Mo(OSiR 3 )(NtBu)L 2 ][FB(C 6 F 5 ) 3 ]( 4a,R = Et and 4b,R = Ph) (Scheme 5), indicated by an additional broad resonance for the BÀFm oiety.T he 11 BNMR resonance observed for 4b (br d, À0.5 ppm) is in good agreement with literature, supporting the structureo ft he anion. [58] The obtained complexes exhibit increased solubility in alkanes, compared to 3a and 3b.T hus,t he anion modification allowed for the growth of single crystals suitable forX -ray diffraction analysis for 4b from saturated pentanes olutions. The stabilityofc ompounds 4a and 4b,e specially towards moisture is, however,still low.
We attribute the fluoridet ransfer to the significantly higher Lewis acidity of B(C 6 F 5 ) 3 in comparison to BF 3 ,w hichi si ng ood agreement with the fluoridei on affinities (FIA) of the two species. [59] Ac omparable behavior has been observed by Parkin and co-workers for aZ n ÀFc omplex, in which the fluoridew as largely transferredu pon exposure to B(C 6 F 5 ) 3 . [60] We found that treatment of Na[HB(C 6 F 5 ) 3 ] [61] 3 ], [62] indicating that the fluoridet ransfer occurs independently of the used cation.

Hydrosilylation reactivity of 3a and 3b
To evaluate the suitability of the formed ionp airs for hydrosilylation, the reaction of 3a with 2equivalents of benzaldehyde as model substrate wasi nvestigated. Reaction progress was monitored by 1 Ha nd 19 FNMR spectroscopy in C 6 D 6 and CD 2 Cl 2 , respectively,r evealing at wo-step process via an intermediate species Int 3a' and finally the formation of benzyl silyl ether under regeneration of the Lewis adduct 2 (Scheme 6).
In the polar solventC D 2 Cl 2 againc onversion to Int 3a' occurs, albeit significantly faster.A fter 6h,7 5% of Int 3a' is formed and after 24 hf ull conversion is observed. Both 19 Fa nd 1 HNMR spectroscopy indicatet he formation of Int 3a' due to unchanged resonances for the complex cation and no observable BÀHr esonance (Figures S6 and S7). Also in this solvent the 1 HNMR resonance at 4.3 ppm is distinctive for the CH 2 group in [PhCH 2 OB(C 6 F 5 ) 3 ] À ,ing ood agreement with literature. [63] Additionally, 11 BNMR spectroscopy of Int 3a' in CD 2 Cl 2 revealed the disappearance of the doublet corresponding to the borohydride moiety of 3a (À25.4 ppm) and an ew sharp singlet at À2.7 ppm, matching literature data for the [PhCH 2 OB(C 6 F 5 ) 3 ] À anion ( Figure S8). [62] The nature of Int 3a' was furthere lucidated by in situ addition of Et 3 SiD to as olution of 2 in CD 2 Cl 2 ,r esulting in the formation of 3a-d 1 (Figure S17), and subsequent addition of benzaldehyde. The resulting 1 HNMR spectrum revealed the -(CHD)-methylene resonance integrating for only one proton, compared to two in the non-deuterated compound (Figure S9). Additionally,t he corresponding 13 CNMR spectrum showedr esonances for the aromatic benzylate protonsi n agreement with literature as wella satriplet for the -(CHD)moiety ( 1 J (CÀD) % 21 Hz), confirming insertion into the BÀDb ond ( Figure S10). [62] It is interesting to note, that in CD 2 Cl 2 furtherr eactiono fInt 3a' formingB nOSiEt 3 virtually does not occur (yield of 2 and BnOSiEt 3 after 72 h % 5%). Possibly,t he polar solvent leads to a larger separation of the ion pair and thus inhibits reaction of the benzyloxy boratew ith the silyl group (Scheme 6).
The reactivity of compound 3b is similar to that of 3a,a lbeit with considerably lower chemoselectivity.N evertheless, 19 FNMR data revealed resonances for Int 3b' virtually identical to those of Int 3a'.A lso the 1 HNMR spectrum in CD 2 Cl 2 shows the distinct resonancea t4 .3 ppm for the CH 2 group of the benzaldehyde inserted anion.
We were interested whether benzaldehyde insertion into the BÀHb ond also occurs in other borate salts such as Na[HB(C 6 F 5 ) 3 ] [61]  Based on the data described above the following steps are suggested for the reactionshown in Scheme 6: i) the Lewis acidic molybdenum cation activates the substrate which allows,u nder nucleophilic attack of the hydride, insertioninto the BÀHb ond; ii)int he apolar solvent, the ion pair is in closer contact so that the electrophilic silicon (activated by the electron poor molybdenum center) reacts with the benzyloxy borate,l eading to the formation of the hydrosilylated product BnOSiEt 3 and the Lewis adduct 2.T he presumably lown ucleophilicity of [PhCH 2 OB(C 6 F 5 ) 3 ] À leads to the observed slow product formation.
Although am echanism based on direct attack of the benzaldehydea tt he silyl group and subsequenth ydride transfer could be envisioned, our spectroscopic evidencer ules against it.
The presentedf indingst huss uggestamechanism different to traditional FLP-basedr eaction pathways, such as for example reported for rheniumo xido/borane Lewis pair catalyzed alkeneh ydrogenation [31] or catalytic hydrosilylation using a nickelo rc obalt/boraneL ewis pair. [64] Whereas mechanisms involvings uch an initial formal hydroboration of the substrate are scarce, they have been described previously in FLP-based CO 2 hydrogenation and alkyne hydroboration reactions, the latter also requiring substrate activation by aL ewis acid. [10,65] A comparable mechanism has very recently been proposed for a Mg/Zn mediated CO 2 hydrosilylation reported by Parkin and co-workers. [66] UV/Vis spectroscopy  (Figure 2). The spectral properties of 2 and 3a-3c are unusualf or molybdenum(VI)c omplexes and indicative of charge-transfer phenomena, which is corroborated by the order of magnitude of the correspondingm olar extinction coefficients (Table 1).
While al igand to ligand charget ransfer cannotb er uled out, the electropositive metal centerf avors al igand to metal (LMCT) transition, originating either from as iloxido or imido based lone pair.T he observed transition energies reflect the electronic situation at the molybdenum metal center. Whereas the corresponding transition found for complex [MoO(NtBu)L 2 ] (1)i sc omparatively weak and of high energy,i ti sr edshifted in complex 2,c orrespondingt ol ower transition energy.T hose for complexes 3a-3c are shiftede ven more to lower energy/ higher wavelength, with the more electron-withdrawing triphenyls iloxide ligand in 3b resulting in the lowest transition energy.T he trend correlates with energeticallyl ow,m ore accessibleu noccupied molecular orbitals in the compounds with ah ighly electropositive metal centera nd thus substantiates the pronounced effect of Lewis acid coordination, as well as ionization, on the metal center.

Moleculars tructures
The molecular structures of 2 and 4b were determined by single-crystal X-ray diffraction analysis. Selected bond lengths are given in Ta ble 2, molecular views of 2 and 4b are given in Figure 3. Full crystallographic detailss uch as structure refinement and experimental details are provided within the Supporting Information.
The Mo=Ob ond length in 2 is substantially elongated in comparison to the parentc omplex 1,1 .8221(9)v ersus 1.7198 (18) . [35] Previously reported molybdenum(VI) oxido borane adducts showed as imilarb ond elongation. [28,29,67] The Figure 2. UV/Vis absorptionspectra of 1, [35] 2 and 3a-3c in CH 2 Cl 2 and photograph of solutionso f2 and 3a (CH 2 Cl 2 ).   (2) [a] MoÀOSiR 3 in 4b. bonds from all other donor atoms to the metal center are slightly shortened, causing ad ecreased flexibility around the metal center,t hereby mostl ikely preventing isomerization (vide supra). The MoÀOSi bond length in 4b, 1 .8975 (19) ,i s comparable to Mo VI complexes bearing triphenylsilanolate ancillary ligands, corroborating ar eduction of the molybdenum oxido bond. [68,69] Also, the clearly bent Si-O-Mo angle of 155.54 (12)8 causesalimitation in orbital overlap between oxygen and molybdenum, in good agreement with as ingle bond to an electrophilicm etal center. [69] All other ligands are bound tighter to the metal centert oc ompensatef or the strongly increased electropositive nature, which is reflectedb y am ean shortening of the MoÀO, MoÀNa nd Mo=N imide bond lengths of approximately 0.02-0.15 per bond, going from 1 to 4b.O verall, the ensemble of bond lengths in the first coordination spherew ell reflects the decreasing electron density at the molybdenum centerint he series 1-2-4b.

Electrochemistry
To get ad eeper insight into the electronic situation at the metal center,t he electrochemical behavior of complexes 1, 2 and 3a-3c was investigated by cyclovoltammetry.C yclovoltammograms of 1 and 3a-3c were recorded in acetonitrile, while for complex 2,C H 2 Cl 2 was used for solubility reasons. Data was referenced to the ferrocene Fc/Fc + redox couple in MeCN and CH 2 Cl 2 ,respectively,u sing the same conditions. Whereasc omplex 1 exhibits no redox process in the observed potentialr ange, the CV of complex 2 depicts ar edox event at À1.37 Vv ersus Fc/Fc + .F or the cationic Mo VI imido silanolate complexes,t he Mo VI /Mo V redox couples experience a large anodic shift (À0.62 V, À0.55 Vand À0.63 Vv s. Fc/Fc + for 3a-3c,r espectively) in comparison to the peak potential found for complex 2,p ointing towards am uch more electro-positiveM o VI metal center (Figure 4).
Given the assumption that the Mo VI /Mo V couplef or complex 1 lies outside of the experimentally accessible voltage range, this clearly corroborates that increased electrophilicity of the molybdenum metal centerf acilitates one-electronr e-ductionofM o VI to Mo V .
In general,t he potentialf ound for 2 is comparable to redox couples previously reported for dioxido molybdenum(VI) compounds, with oxido imido compounds usually exhibiting significantly higherp otentials due to ap otent electron donating capabilityo ft he imido ligand. [22,24,70] However,t he anodic shift of more than 0.7 Vfor the redox couples of 3a-3c,incomparison to 2,l eads to remarkably low reduction potentials (Figure 4).
To evaluatet he reversibility of the redox processes found for complexes 2 and 3a-3c (Table3), the scan rate (n)d ependence of 2 and 3a was investigated.
For compound 3a,t he peak separation in the cyclic voltammogram is scan rate independenta nd the peak current linear dependento nn 1/2 ,c orroborating ar eversible process. In contrast, assessment of the redox couple found for the oxido/ borane adduct 2 reveals an increase in peak separation at highers can rates pointing towards electrochemical irreversibility (FigureS3).
Such irreversibility is often caused by chemical instability of the analyte, which can likely be attributed to the dynamics of the system in solution (lability of the BÀOb ond) although solvent specific phenomenaa sw ell as slow electron transfer rates cannotb ee xcludedapriori. Due to the low value for I pa /I pc for compound 3c,t he same scan rate study was also performed, showingscan rate independence thereby suggesting reversibility ( Figure S3).
These findings indicate that the redox couples for 3a-3c are truly reversible and likely metal based. This is highly interesting, given that the reduction of molybdenum(VI) to molybdenum(V) is frequently hampered by irreversible dimerization. An additional benefit of the system is the possibility for an electronic fine tuning by simply varying the silanolate group.

Conclusions
The reported molybdenum oxido based Lewis adduct [Mo{OB(C 6 F 5 ) 3 }(NtBu)L 2 ]f eaturing reversible BÀOb onding reacts with tertiary silanes to form highly unusual ion pairs of the type [Mo(OSiR 3 )(NtBu)L 2 ][HB(C 6 F 5 ) 3 ]( R= Et, Ph). This is not only ar aree xample of FLP-like reactivity involving the widespreadt ransition metal oxido functionality but also gives accesst oh igh valent molybdenums pecies with unique spectroscopic and electronic properties of potential interest to a broad field of chemistry. Furthermore, it represents ar are instanceo faLewis acidm ediated activation of the oxido ligand.  The described ion pairs are furtherr eactive towards benzaldehyde, regeneratingt he initial Lewis adduct via formation of the respective benzyloxy silane in as tepwise manner.T his reactivity is highly dependent on the polarity of the employed solvent, with polar solvents inhibiting the formation of the hydrosilylation product. This supports mechanistic considerations suggesting an unusual two step pathway involving insertiono f the benzaldehyde into the borohydride bond of the anion, forming the intermediate species [Mo(OSiR 3 )(NtBu)L 2 ][PhCH 2 OÀ B(C 6 F 5 ) 3 ], which is supported by spectroscopici nvestigations. In summary,t he presented system combines the advantage of aL ewis adduct with the reactivity of ar eversibly formed FLP. We believe the remarkable stabilityo ft he isolated silanolate borohydride ion pairs to be ar esult of electronic stabilization caused by the metal center ancillary to the Lewis basic oxido group. The research disclosed here is of particulari nterest because of the abundance of metal oxido motifs and the potential suitability to other M=Oc ompounds. The here presented step-wise reactivity also offers the prospect of ac ontrolled, metal-templated transfer of silyl groups and possibly other electrophiles.

Experimental Section
General Unless specified otherwise, all experiments were performed under inert conditions using standard Schlenk equipment or aN 2 -filled glovebox. Commercially available chemicals were used as received.
Air and moisture sensitive chemicals were stored in Schlenk flasks or under N 2 atmosphere in ag lovebox;l iquids were additionally stored over molecular sieves. The metal precursor [MoO(N-tBu)Cl 2 (dme)], [71] the ligand HL [51] as well as B(C 6 F 5 ) 3 [72] were synthesized according to known procedures. Solvents were purified via a Pure-Solv MD-4-EN solvent purification system from Innovative Te chnology,I nc. The 1 H, 11 B, 13 C, 19 Fa nd HSQC NMR spectra were recorded on aB ruker Optics instrument at 300/96/75/282 MHz. Peaks are denoted as singlet (s) doublet (d), doublet of doublets (dd), triplet (t), quartet (q) and multiplet (m), broad peaks are denoted (br) and all peaks are referenced to the solvent residual signal. Shifts in 11 Ba nd 19 FNMR are referenced to external standards (BF 3 ·Et 2 Oa nd CFCl 3 ,r espectively). Used solvents and peak assignment are mentioned at the specific data sets. HR-MS (ESI + /ESI À )m easurements were performed at the University of Graz, Department of Analytical Chemistry,u sing aT hermo Scientific Q-Exactive mass spectrometer in positive and negative ion mode, the used solvent was acetonitrile. Peaks are denoted as ionic mass peaks, and the unit is the according ions mass/charge ratio. Calculated and found isotopic patterns are provided within the supporting information. Samples for infrared spectroscopy were measured on aB ruker Optics ALPHA FT-IR Spectrometer.I Rb ands are reported with wavenumber (cm À1 )a nd intensities (s, strong;m ,m edium; w, weak). Elemental analyses were measured at the Graz University of Te chnology,I nstitute of Inorganic Chemistry using aH eraeus Vario Elementar automatic analyzer.D eviations in the found elemental compositions of ionic compounds (low carbon content) are attributed to the pronounced water sensitivity as also observed by HR-MS measurements. In all cases, addition of H 2 Ot ot he elemental compositions would diminish the observed error. UV/Vis spectroscopy UV/Vis spectra were recorded on aV arian Cary 50 spectrophotometer in aq uartz cuvette with an optic path length of 1mm. Analyte solutions were prepared in CH 2 Cl 2 near 1mm.P eak maxima are reported with wavelength (nm) and molar exctinction coefficient (M·cm À1 ), overlapping peak maxima are denoted as shoulders. All maxima and the corresponding absorptivities were obtained via deconvolution in the SciDaVis [73] software using as caled Levenberg-Marquardt algorithm. [74] X-ray diffraction analyses Single-crystal X-ray diffraction analyses were measured on a BRUKER-AXS SMARTA PEX II diffractometer equipped with aC CD detector.A ll measurements were performed using monochromatized Mo Ka radiation from an Incoatec microfocus sealed tube at 100 K( cf. Ta ble S1). Absorption corrections were performed semiempirical from equivalents. Structures were solved by direct methods (SHELXS-97) [75] and refined by full-matrix least-squares techniques against F 2 (SHELXL-2014/6). [75] Full experimental details for single-crystal X-ray diffraction analyses of all compounds are provided in the Supporting Information.

Electrochemistry
Electrochemical measurements were performed in ag lovebox under N 2 atmosphere in dry solvents with aG amry Instruments Reference 600 Potentiostat using at hree-electrode setup. Used electrodes were glassy carbon as working electrode, Pt wire (99.99 %) as supporting electrode and an Ag wire immersed in a solution containing 10 mm AgNO 3 and 100 mm [NBu 4 ][PF 6 ]i n CH 3 CN, separated from the analyte solution by aV ycor tip, as ar eference electrode. Analyte concentrations ranged from 0.2 to 1.0 mm in CH 3 CN or CH 2 Cl 2 ,r espectively.T he supporting electrolyte used was [NBu 4 ][PF 6 ]( 100 mm). Cyclic voltammetry data was smoothed in the SciDaVis [73] software using am oving average filter. Full sweep-width cyclic voltammograms are provided in the Supporting Information. Syntheses 2,4-Di-tert-butyl-6-((phenylimino)methyl)phenol (HL):A nalytical data is in agreement with literature, [51] additional 1 HNMR data in CD 2 Cl 2 is given for comparison reasons. 1 3 :A nalytical data is in agreement with literature, [72] additional 11 Ba nd 19 FNMR data in CD 2 Cl 2 is given for comparison reasons. 11  Complex syntheses:A ll complexes except 2 are very sensitive towards moisture in solution and solid state, 2 is sensitive towards moisture in solution. They can be stored at ambient temperature in aN 2 -filled glovebox for several weeks without decomposition.
Improved synthesis of [MoO(NtBu)L 2 ]( 1): [35] For the synthesis of 1,as olution of 1equiv of [MoO(NtBu)Cl 2 (dme)] (1.13 g, 3.28 mmol) in acetonitrile (20 mL) was added dropwise to as uspension of 2equiv HL (2.03 g, 6.57 mmol) and 2. quently the red mixture was evaporated in vacuo. To luene (25 mL) was added to the red residue, resulting in ar ed solution and areddish solid that was filtered off and washed with toluene (2 10 mL) until it was essentially colorless (NEt 3 ·HCl). The red filtrate was evaporated in vacuo, followed by the addition of acetonitrile (10 mL). After 15 minutes of stirring, ab right yellow solid precipitated, which was isolated by filtration. The solid was re-dissolved in n-pentane (50 mL), filtered, and the resulting solution was evaporated in vacuo to obtain 1 as yellowish-orange colored solid (2.11g,8 0%). Analytical data is in agreement with literature, [35] additional 1 HNMR an UV/Vis data in CD 2 Cl 2 is given for comparison reasons. 1  Synthesis of [Mo{OB(C 6 F 5 ) 3 }(NtBu)L 2 ]( 2):F or the synthesis of 2,a solution of 1equiv B(C 6 F 5 ) 3 (313 mg, 0.61 mmol) in dry pentane (5 mL) was added to as olution of 1equiv of 1 (490 mg, 0.61 mmol) in the same solvent (10 mL). The addition was accompanied by an immediate color change from yellow to dark red. The reaction mixture was subsequently stirred at room temperature for 6h,w hereupon al arge quantity of ar ed crystalline precipitate had formed. The precipitate was subsequently filtered off, washed with cold pentane (10 mL) and acetonitrile (5 mL) and dried in vacuo to yield 2 as am icrocrystalline brick-red solid (649 mg, 81 %). Single crystals suitable for X-ray diffraction analysis were obtained via crystallization from ac oncentrated pentane solution of 2 at À35 8C. 1

Synthesis of [Mo(OSiEt 3 )(NtBu)L 2 ]
[HB(C 6 F 5 ) 3 ]( 3a):F or the synthesis of 3a,5equiv of Et 3 SiH (61.2 mL, 0.38 mmol) were added to a solution of 1equiv of 2 (100 mg, 0.08 mmol) in toluene (15 mL). After stirring for 16 ha tr oom temperature, the color had changed to purple. After removal of all volatiles, the residual sticky substance was re-dissolved in acetonitrile (6 mL) and filtered. The solvent was subsequently evaporated and the residual solid washed twice with pentane (2 5mL) and thoroughly dried in vacuo to obtain 3a as af luffy dark purple solid (104 mg, 82 %). 1