Heavy Alkali Metal Manganate Complexes: Synthesis, Structures and Solvent‐Induced Dissociation Effects

Abstract Rare examples of heavier alkali metal manganates [{(AM)Mn(CH2SiMe3)(N‘Ar)2}∞] (AM=K, Rb, or Cs) [N‘Ar=N(SiMe3)(Dipp), where Dipp=2,6‐iPr2‐C6H3] have been synthesised with the Rb and Cs examples crystallographically characterised. These heaviest manganates crystallise as polymeric zig‐zag chains propagated by AM⋅⋅⋅π‐arene interactions. Key to their preparation is to avoid Lewis base donor solvents. In contrast, using multidentate nitrogen donors encourages ligand scrambling leading to redistribution of these bimetallic manganate compounds into their corresponding homometallic species as witnessed for the complete Li ‐ Cs series. Adding to the few known crystallographically characterised unsolvated and solvated rubidium and caesium s‐block metal amides, six new derivatives ([{AM(N‘Ar)}∞], [{AM(N‘Ar)⋅TMEDA}∞], and [{AM(N‘Ar)⋅PMDETA}∞] where AM=Rb or Cs) have been structurally authenticated. Utilising monodentate diethyl ether as a donor, it was also possible to isolate and crystallographically characterise sodium manganate [(Et2O)2Na( n Bu)Mn[(N‘Ar)2], a monomeric, dinuclear structure prevented from aggregating by two blocking ether ligands bound to sodium.


Introduction
Manganese occupies a unique position in the centre of the periodic table as it exists in oxidation states ranging from À 3 to + 7, [1] indicating a most versatile reactivity in reactions with organic substrates. The oxidation state + 2 illustrates its uniqueness in particular, distinguishing it from its transition metal neighbours by significant ionic contribution of the metal-carbon bond. This has been attributed to the significant harder character of Mn(II) when compared to other first row transition metal ions. [2] This "main-group-mimic" characteristic in combination with accessible redox-chemistry and the fact that manganese is i) earth abundant ii) non-toxic and iii) inexpensive compared to other transition metals [3,4] has led to a wide range of applications of organometallic Mn(II) compounds, as for example in acylation, addition and other CÀ C bond forming or oxidation reactions. [5,6] The first applications of organomanganese(II) rest on the pioneering reports by Normant and Cahiez in the 1970s. [7] Subsequently, four main organomanganese(II) reagent types have received most attention: namely, organomanganese halides, RMnX; [8] diorganyl manganese compounds MnR 2 ; [9][10][11] lower-order triorganomanganates [MnR 3 ] À , [12] and higher order tetraorganomanganates [MnR 4 ] 2À . [11][12][13] The third and fourth types require cations for charge balance. In general, the lowerand higher-order ate compounds benefit from increased thermal stability and solubility compared to the monometallic analogues, hence the chemistry of polyalkyl-manganates has been extensively studied and reviewed by Cahiez, [14] Duplais and Buendia, [7] and by Oshima. [15] Despite their interesting organic applications, the identity of these species has remained largely concealed, although recent studies by Uzelac and Hevia have shown the applications of structurally-defined lithium manganate [(TMEDA) 2 Li 2 Mn(CH 2 SiMe 3 ) 4 ] to promote oxidative homocoupling of aryliodides. [16] An interesting feature is the unique synergic reactivity exhibited by certain heterobimetallic or mixed metal (ate) compounds that occurs upon mixing Mn(II) [17][18][19] or other divalent metal complexes (e. g., Mg [20][21][22][23][24][25][26] or Zn [23,24,[27][28][29] ) with alkali metal precursors AMÀ R (with AM = usually Li, Na, or K, but rarely Rb and Cs). [22,30] The synergistic reactivity seen with alkali metal manganates (notably monoalkyl-bisamido manganates) has been labelled alkali-metal-mediated manganation (AMMMn), [18,19,31] in respect to metallation (CÀ H to C-metal exchange) reactions, which fail to work at all or to work efficiently without the intervention of an alkali metal.
Though not studied in much detail, AMMMn bears a close similarity to the more extensively studied alkali-metal mediated magnesiations (AMMMg) due to shared features between Mn(II) and Mg(II). It is therefore not surprising that analogous Mg(II) complexes (or vice versa) can be found in the literature for a large number of Mn(II) structures. Prominent examples include Power's mixed lithium-manganate trisamide complex [Mn{N-(SiMe 3 ) 2 } 3 Li(THF)] [32] and isostructural lithium-magnesiate trisamide complex [Mg{N(SiMe 3 ) 2 } 3 Li(THF)] [33] or the hydride encapsulated inverse crown sodium-magnesiate [Na 2 Mg 2 (μ-H) 2 {N(iPr) 2 } 4 (toluene) 2 ] [34] and its manganese(II) analogue [Na 2 Mn 2 (μ-H) 2 {N(iPr) 2 } 4 (toluene) 2 ] [35] reported by Mulvey. Recently, as part of a growing movement to develop the use of main group compounds in catalysis, chemists have begun to study magnesiate compounds in homogeneous catalysis. [36][37][38] Interestingly, heavier alkali metals, nearly always potassium, but also to a much lesser extent, rubidium and caesium, have played key roles in this development. [39] In this context Guan and co-workers have successfully utilised mixtures of saline potassium hydride KH and group 2 metal bis-hexamethyldisilazides M(HMDS) 2 [M = Mg, Ca; HMDS = N(SiMe 3 ) 2 ] that form effective catalysts for hydrogenation of olefins. [40] The active species in these transformations in the case of magnesium (M = Mg) is evidenced to be the mixed amido-/hydridopotassium magnesiate [KMg(H)(HMDS) 2 ] 2 , which was first reported and structurally characterised by Hill. [41] We would advocate that it is now considered vital to include the whole of group 1 (Li -Cs) when investigating catalytic performance, since gradations in reactivity can take place depending on which alkali metal is used. Investigations to gain more structural insights in whole series of group 1 complexes often exclude the heavier homologues due to the synthetic challenge or inaccessibility, [42,43] hence studies such as that by Schulz, [44] Roesky, [45] or Liddle, [46] in which light is shed on the whole of group 1 within the same ligand framework, remain exceptional. One of the most widely utilized ligands to study alkali metals in complexes are amide derivatives since they are of central importance in many chemical transformations as metallation and amide-transfer reagents or as Brønsted bases. [47][48][49] Again, the amide chemistry of the heaviest group 1 elements rubidium and caesium is less developed with structurally characterised compounds being restricted to few examples of i) silyl-stabilised (trimethylsilyl)-or (phenyl-dimethylsilyl) amides, [50][51][52][53][54][55][56][57][58][59] ii) electron withdrawing perfluorinated alkyland aryl-substituted amides [60] or iii) sterically hindered 2,2,6,6tetramethylpiperidides. [61] Silyl-stabilised bulky amides, especially silyl-aryl-substituted derivatives, may offer a lower degree of aggregation and enable solubility in common organic solvents. This is particularly important for the heavier alkali metal amides as the increasingly electrostatic metal nitrogen bond favours the formation of insoluble salt like compounds. The possibility of the alkali metal to engage in inter-and intramolecular anagostic as well as AM···π-arene interactions with the ligand framework can be the decisive factor in the formation of fundamentally different complex structures as highlighted in recent advances in low-valent main group metal chemistry. [62,63] Investigation and deeper insight into the class of Mn(II) compounds remains challenging, since stringent exclusion of air to prevent unwanted oxidation to Mn(IV) species is paramount, while characterisation by NMR spectroscopy is hindered by the paramagnetism of Mn(II). Additionally, isolation and handling of the mediating alkali metal alkyl precursor is impeded with increasing electropositivity of the group 1 metal (Cs > Rb > K > Na > Li), as has recently been shown in the challenging structural elucidation of the heaviest alkali-metalbenzyl complexes by Strohmann. [64] Hence, manganate(II) compounds bearing the heaviest alkali metals are scarce and only combinations of rubidium with manganese in mixed or higher oxidation states, [65] in oxidation state + 1 [66] or in hydrophilic coordination compounds [67] and metal-organic frameworks [68] (MOFs) are structurally identified. Similarly, while Cs + is commonly found as a counter-cation in water soluble coordination compounds or caged in cluster complexes containing manganese in the oxidation state + 2 or + 3, [69] only one study from Berke reveals structural insight of air-and moisturesensitive organocaesium compounds in the tris(cyclopentadienyl) manganates. [70] With this background, this study sets out to merge these different aspects by attempting to synthesise and characterise a series of new alkali metal manganates, with particular emphasis on rubidium and caesium examples. First, we highlight the influence of the coordinating donor ligands, THF, Et 2 O, TMEDA (N,N,N',N'-tetramethylethylenediamine) and PMDETA (N,N,N',N",N"-pentamethyldiethylenetriamine) in the success or failure to isolate challenging manganate-complexes instead of a mixture of their heteroleptic, monometallic alkyl-amido Mn(II) compounds and alkali metal amides, derived from manganatecomplex destruction. The heaviest homologues of these alkali metal amides are also discussed extensively in terms of their structural diversity when solvated with said donors. We also reveal how exclusion of donor solvents allowed the successful synthesis and structural characterisation of the, to the best of our knowledge, first Rb-manganate with manganese solely in the oxidation state + 2 and its heavier Cs-homologue prepared in hydrocarbon solvents.

Results and Discussion
Initial attempts to prepare alkali metal manganate complexes We started our investigations by mixing stoichiometric amounts of the literature known bis-arylsilyl amide Mn(N 'Ar ) 2 [71] [N 'Ar = N(SiMe 3 )(Dipp), where Dipp = 2,6-iPr 2 -C 6 H 3 ] and commercially available n-butyl lithium solution in hexane. Unfortunately, these co-complexation attempts, that is the formal addition of two monometallic species to yield a new bimetallic complex, repeatedly resulted in the formation of mixtures of lithium amide LiN 'Ar or its monometallic coordination compounds with various donor solvents and yellowish residue, which failed to deposit any crystalline material for further investigation (see Scheme 1).
Changing the lithium alkyl precursor to the neosilyl lithium LiCH 2 SiMe 3 reinforces this observation as, in separate experiments, crystals of homometallic LiN 'Ar · (Et 2 O) 2 , [72] LiN 'Ar · (THF) 3 , [73] LiN 'Ar · (TMEDA) [74] and LiN 'Ar · (PMDETA) [74] could be identified upon repeating the reaction with addition of Et 2 O, THF, TMEDA and PMDETA respectively, consistent at least in part with manganese-lithium exchange phenomena. The total absence of additional donor solvent promoted the formation of formal twofold ligand exchange to yield one equivalent of the [LiN 'Ar ] 2 [75] dimer with one equivalent of polymeric [{Mn-(CH 2 SiMe 3 ) 2 } ∞ ], [12,76] which can be evidenced by X-ray analysis and is observed by a colour change to the distinctive orange colour of [{Mn(CH 2 SiMe 3 ) 2 } ∞ ].
As this ligand redistribution seemingly follows Pearson's acid-base concept with the hard Li + cation preferring the harder nitrogen of the amide ligand, we next investigated the influence of heavier alkali metals on this metathesis reaction with respect to the donor solvents applied. Treatment of Mn(N 'Ar ) 2 dissolved in benzene with freshly prepared n BuNa under reflux conditions leads to precipitation of a white solid of unknown composition. Attempts to recrystallize this material in aromatic solvents failed, but slow diffusion of n-hexane into a concentrated Et 2 O solution yielded colourless crystals in 71 % yield and suitable for X-ray crystallographic determination, revealing the bimetallic sodium manganate [(Et 2 O) 2 Na( n Bu)Mn-[(N 'Ar ) 2 ] (1) (see Scheme 2).
Compound 1 was characterised by combining X-ray crystallographic studies (see Figure 1) with elemental analysis and infrared spectroscopy (see Supporting Information for details). Attempts to record meaningful NMR spectra in benzene-d 6 and THF-d 8 solutions were unsuccessful due to the paramagnetic nature of Mn(II). However, Evans methods could be used to assess its magnetic moment (effective magnetic moment μ eff = 5.94 μ B ) which is consistent with five unpaired electrons in a high-spin state for the Mn(II) centre. Manganate 1 crystallised in the monoclinic primitive space group P2 1 /c and can be classified as a contacted ion pair type complex, containing a 2 : 1 amido/butyl stoichiometric ratio, similar to that of previously reported Mn(II) synergistic metalating agents. [19,31] In comparison to these systems, the arrangement of the amide ligand in 1 is strikingly different though, as none of the amide ligands lie in a bridging position between the Mn and Na metal but solely bind to the transition metal. Thus, Mn1 occupies a N 2 C trigonal planar coordination (sum of bond angles: 360.0°) comprising two amide ligands (mean MnÀ N distance, 2.0523 Å) and one butyl ligand [Mn1-C1, 2.1668(13) Å]. This three-coordinate Mn moiety is connected through two bridges to the Na1 centre, which overcomes intermolecular bonding tendencies by coordination to two diethyl ether molecules. One bridge consists of the α-carbon of the bridging butyl ligand and the other is the η 3 -coordination (Na1-C25: 3.1190(14) Å, Na1-C26: 2.8122(14) Å, Na1-C27: 3.0180(13) Å) to one of the amido ligand's Dipp groups. These alkali metal-arene interactions are decisive for the formation of the polymeric helical chain structure in the related magnesium compound [74] and are a common feature in main group metal chemistry in general, thus they have been discussed extensively in the literature. [77][78][79] It should be noted that adding the more strongly donating oxygen donor solvent THF did not allow for isolation of a defined species. This has been the incentive to investigate for possibly complex formation tendencies with multidentate donor ligands.
Isolation of neutral manganese complex 2 a in its pure form is rendered difficult as it exhibits similar solubility to those of the corresponding alkali metal amide by-products. In general, mixed amido-alkyl Mn(II) species containing non-stabilised nbutyl ligands remain scarce in literature, [14,80] since β-hydride abstraction, followed by alkene dissociation is the principal decomposition pathway often observed in these compounds. Crystalline samples of 2 a did not show any signs of decomposition in argon atmosphere at room temperature over the course of six weeks.
Similarly, "β-stable" Mn-(N 'Ar )(CH 2 SiMe 3 )·TMEDA (2 b) cannot be isolated in pure form from this complex destruction pathway and attempts to synthesise 2 b from stoichiometric mixtures of [{Mn-(CH 2 SiMe 3 ) 2 } ∞ ] and free amine H-N 'Ar with one equivalent of TMEDA led to intractable product mixtures.
Determined by X-ray crystallography 2 a displays a distorted-tetrahedral coordination geometry in the solid state, which is typical for four-coordinate Mn(II) complexes (see In order to probe the stability, that is, the tendencies for complex dissociation with chelating N-donor solvents for the heavier group 1 metals, the relevant alkali metal reagent needs to be optimised. The lower stability of the butyl compounds of potassium, rubidium and caesium necessitated the switching to more stable neosilyl-alkyl reagents AM(CH 2 SiMe 3 ) [61] (AM = Na, K, Rb, Cs), in which β-hydrogen elimination side reactions are impossible. Following the same methodology, AM(CH 2 SiMe 3 ) reacts at room temperature with Mn(N 'Ar ) 2 to afford the cocomplexed products as a white precipitate from a yellow benzene reaction mixture. However, recrystallisation from a TMEDA\n-hexane solution leads repeatedly to ligand scrambling to yield the monometallic complexes Mn(N 'Ar )(CH 2 SiMe 3 )·TMEDA Determined by X-ray crystallography, the molecular structure of 2 b (see Figure 2b) has a Mn atom in a distorted tetrahedral environment made up by the anionic amide and the Scheme 3. Ligand redistribution of mixed monoalkyl/bisamide sodium, potassium, rubidium, and caesium manganates in presence of TMEDA to give alkali metal amide and heteroleptic monoalkyl/monoamide manganese complexes in 1 : 1 ratios.

Chemistry-A European Journal
Research Article doi.org/10.1002/chem.202201716 neosilyl-alkyl ligands and is, like compound 2 a, coordinatively satisfied by TMEDA, which chelates to the metal via its two nitrogen atoms. Akin to its butyl analogue, 2 b crystallises in a monoclinic space group and the short Mn1-C1 [2.143(2) Å] bond and Mn1-N1 [2.0623(15) Å] bond compare favourably to the structurally similar complex 2 a, with the main difference lying in the wider N1-Mn1-C1 bite angle of 135.19 (7)°in the neosilyl-substituted complex 2 b, which is a consequence of the increased steric bulk induced by the CH 2 SiMe 3 group.

Synthesis and characterisation of unsolvated and solvated rubidium and caesium amides and their role in manganate formation
In accordance with previous observations regarding ligand scrambling, carrying out this set of co-complexation reactions in the presence of tridentate PMDETA, instead of TMEDA, leads to formation of a yet unidentified Mn(N 'Ar )(CH 2 SiMe 3 )·PMDETA species and homometallic AM(N 'Ar )·PMDETA (AM = Na: Na-(N 'Ar )·PMDETA, [74]  Intrigued by the repeated isolation of solvated heavier group 1 amide by-products, especially the rubidium and caesium homologues, we set out to find a rational synthetic pathway to increase the body of these species and gain access to the unsolvated parent alkali metal amides. Therefore The rubidium and caesium amide complexes 3-AM could be identified by X-ray crystallography, as well as by one-and twodimensional NMR and infrared spectroscopy (see Supporting Information for detailed analysis). 3-Rb crystalises in the monoclinic space group P2 1 /c and 3-Cs in the tetragonal space group P-42 1 m while both complexes adopt a polymeric chaintype structure interconnected by attractive intermolecular anagostic interactions in the solid state (see Figure 3).
For rubidium amide 3-Rb the formally one-dimensional polymeric chains propagate parallel to the crystallographic caxis through stabilizing Rb···π-arene interactions which are best described as η 6  Suspending 3-AM in hexane and subsequent addition of chelating TMEDA led to transparent solutions, which yielded colourless crystals after storage at À 20°C. X-ray crystallography and one-and two-dimensional NMR spectroscopy confirmed the same products 3-AM·TMEDA, as have been found in the corresponding dissociation of the co-complexed manganate. In contrast to the lighter Li to K congeners, 3-Rb·TMEDA and 3-Cs·TMEDA are not deaggregated by the donor solvent, but a polymeric chain structure is retained (see Figure 4).
The crystallographic determination of 3-Rb·TMEDA established it as a contacted TMEDA solvate where the metal centre binds to the neutral nitrogen donor, that is syn orientated towards the ligand's N(SiMe 3 ) moiety. In contrast to parent amide 3-Rb, the absence of interconnecting anagostic contacts leads to an infinite zig-zag chain in 3-Rb·TMEDA that propagates parallel to the crystallographic b-axis through one η 6 -coordinating Rb···π-arene interaction [Rb1···C centroid : 3.05 Å, range of Rb1···C arom distances: 3.323(2) Å -3.413(2) Å] and one η 4 -coordinating Rb···π-arene interaction [Rb1···C centroid : 3.14 Å, range of Rb1···C arom distances: 3.260 (19) As a previous study has shown, amide salts of the lighter homologues lithium, sodium and potassium AM(N 'Ar ) (AM = LiÀ K) can be similarly deaggregated by adding tridentate PMDETA. In the case of Li(N 'Ar )·PMDETA and Na(N 'Ar )·PMDETA this deaggregation takes place all the way down to the monomers; whereas the heavier potassium amide PMDETA adduct [{K(N 'Ar )·PMDETA} 2 ] forms a dimer in the solid state. [74] Descending down group 1, suspensions of parent amides 3-AM (AM = Rb, Cs) in benzene or hexane treated with PMDETA yielded clear solutions and storing mixtures of 3-AM in hexane/ PMDETA at À 20°C yielded colourless crystals of the corresponding alkali metal amide donor adduct complexes [{AM-(N 'Ar )·PMDETA} ∞ ] (3-AM·PMDETA, AM = Rb, Cs). The compounds could be structurally identified by X-ray crystallography and one-and two-dimensional NMR spectroscopy, confirming the previous findings from the complex destruction pathway in the corresponding bimetallic manganate complexes (see Scheme 4). Unlike in the lighter alkali metal adduct complexes PMDETA is

Chemistry-A European Journal
Research Article doi.org/10.1002/chem.202201716 not sufficient to coordinatively saturate the heaviest group 1 metals and therefore 3-Rb·PMDETA and 3-Cs·PMDETA exhibit polymeric chain structures similar to their TMEDA analogues in the crystal as depicted in Figure 5.
Unlike in 3-AM and 3-AM·TMEDA, in which a significant structural difference is exhibited between the rubidium and caesium homologue, 3-Rb·PMDETA and 3-Cs·PMDETA are structurally more closely related and both compounds crystallise in the monoclinic space group P2 1 /c. Note that in 3-Cs·PMDETA the asymmetric unit cell contains two crystallographic independent molecules with slightly different structural parameters, but for brevity only one is discussed here. In the Rb   in s-block chemistry in general [85] and rely heavily on the induced polarisation by the main group metal, so they are especially strong regarding the large difference in electronegativity of the Si atom and the heavier group 1 metals Rb and Cs. [52,86] Exposure of manganates 4-AM (AM = K, Rb and Cs) in the aromatic solvents benzene or toluene to the coordinating oxygen donors Et 2 O or THF did not yield crystalline material suitable for X-ray crystallographic analysis, which may be explained by a lower tendency of these non-chelating ligands to disintegrate the bimetallic species. In contrast to previously reported Mn(II) metallators [18] extended heating of benzene or toluene solutions of manganates 4-AM did not show any signs of direct manganation or group 1 arene metallation reactions, which emphasises the importance of template effects that play a crucial role in AMMMn reactions.

Conclusion
In conclusion, we have shown that treatment of literature known manganese-bisamide Mn(N 'Ar ) 2 with alkyl lithium reagents give reaction mixtures with limited stability under investigated conditions and exclusively led to twofold ligand exchanged monometallic lithium amide and bis-alkyl-manganese compounds. Addition of the polydentate coordinating solvents TMEDA and PMDETA allowed for more detailed structural identification of heteroleptic, monometallic alkylamido Mn(II) complexes 2 a and 2 b which were crystallographically characterised, as well as the monometallic alkali metal amide complexes 3-AM·TMEDA and 3-AM·PMDETA (AM = Rb, Cs), which in contrast to their lighter homologues adopt infinite linear or zig-zag arrangements as identified in their solid state structures. Structural diversity in these adductcomplexes and in the parent, donor-free alkali metal amide compounds 3-Rb and 3-Cs, obtained from a transamination reaction, is discussed, emphasising the complexity in seemingly simple alkali metal amide systems. The co-complexation of butyl-sodium with said Mn(II) amide yields monomeric, bimetallic sodium manganate 1 in diethyl ether, a potential metalating agent in AMMMn chemistry, which will be targeted in future work. The first structurally characterised Rb-manganate 4-Rb with manganese solely in the oxidation state + 2 is reported, together with its heavier Cs-homologue 4-Cs, which represents the first example of a mixed amido-alkyl caesium manganate. Presumably the AM-arene interactions of the heavier AM metals Rb and Cs favour formation of the bimetallic species, contrary to the lighter Li and Na homologues in which these arene interactions are less pronounced, which leads to ligand redistribution reactions. Likewise, adding the donor solvents TMEDA and PMDETA may block the coordination sites of the AM metals, which competes with the AM-arene interactions, thus favouring AM cleavage and complex destruction.
With these alkali metal complexes now in hand, future work will investigate the application of these complexes in stoichiometric and catalytic reactions.

Experimental Section
Experimental details can be found in the Supporting Information.