Structural Motifs of Alkali Metal Superbases in Non‐coordinating Solvents

Abstract Lochmann–Schlosser superbases (LSB) are a standard reagent in synthetic chemistry to achieve an exchange of a proton on an organic framework with an alkali metal cation, which in turn can be replaced by a wide range of electrophilic groups. In standard examples, the deprotonating reagent consists of an equimolar mixture of n‐butyllithium and potassium t‐butoxide. However, the nature of the reactive species could not be pinned down either for this composition or for similar mixtures with comparable high reactivity. Despite the poor solubility and the fierce reactivity, some insights into this mixture were achieved by some indirect results, comparison with chemically related systems, or skillful deductions. Recent results, mainly based on new soluble compounds, delivered structural evidence. These new insights lead to advanced and more detailed conclusions about the interplay of the involved components.

Abstract: Lochmann-Schlosser superbases (LSB) are as tandard reagent in synthetic chemistry to achieve an exchange of ap roton on an organic framework with an alkali metal cation, which in turn can be replaced by aw ide range of electrophilic groups. In standard examples, the deprotonating reagent consists of an equimolar mixture of n-butyllithium and potassium t-butoxide. However,t he nature of the reactive speciesc ould not be pinned down either for this composition or for similarmixtures with comparable high reactivity.D espite the poor solubility and the fierce reactivity, some insights into this mixture werea chieved by some indirect results, comparison with chemically related systems, or skillful deductions. Recent results, mainly based on new soluble compounds, delivered structural evidence. These new insights lead to advanced and more detailed conclusions about the interplay of the involved components.

The Organometallic Chemistry of Alkali Metals
The organometallic chemistry in general,and the organometallic chemistry of the alkali metalsi np articular, is based on the seemingly unfavorable bond between the carbon atom of an organic group and the metal atom. At first glance,t his arises from the tendency of carbon to form covalent bonds, while many metalsp refer more ionic interactions. Schlosser considered this explanation as being too simple. [1] He points out the highly polar character of the metal-carbon bond. This polarity is based on the differenceo ft he correspondinge lectro-negativities, leadingt oanegatively polarizedc arbona se videnced by experimental facts. [2] Schlossere mphasizes that the situation with the alkali metal atom should not be neglected;e ven with the polar bond, the metal atom is not keen to go without furtherb ondede lectrons. Yet, as ingle, weak bond cannot serve the electronic needs of the metal atom. The solution is in the formation of multiple weak bonds. The metal atom is flexible in its coordination sphere,b oth in the number and the geometrica rrangemento fl igands. This is not the case for the carbon of an organic group attached to the metal atom. Due to the covalently attached atoms on the carbon atom (e.g. carbon or silicon centered groups), its coordination to metal atoms is limited in terms of direction.
In other words, the negative charge of these organic groups is spatially directed towards one, two, or three metal atoms. This electronic situation leads to the formation of oligomers or larger aggregates,w hich can be derived as sections from salt structures, which was aptly demonstratedi nareview by Stalke et. al. [3] Both effects, the metal-carbon bond polarization and the tendency to form oligomers, play important roles with respect to the reactivity of organometallic alkali metal compounds or their interaction with other molecules. The negative charge on the carbon atom renders it av ery strong reducing agent on one side, as well as ar eactive Lewis base, leading to nucleophilic behavior or high Brønsted basicity on the other. The degree of aggregation of organometallic alkali metal compoundsh as al arge influence on the solubility of such compounds; quantitative effects on the reactivity are also discussed. [4] However,t he influence of different grades of aggregation on reactionr ates is far from being simple. [5] At this point, it is very important to mention the necessary distinction between different classes of organic groups in organometallic compounds. The sp 3 -hybridized carbon atom of an aliphatic alkyl group bondedt ot he metal shows low group electronegativity,a nd acts as ad onor with the hapticity h 1 ; nevertheless, it can adopt bridging positions to metals. sp 2and sp-hybridized carbon atoms of allyl, aryl, or ethynyl groups show highere lectronegativity due to the greater s-character of the involved orbitals. In addition, they can accept further, bridging interactions to metals using p-orbitals. [2b] This greatly affects the reactivity and aggregation of such compounds. In the following, the focus is set on the organometallicc hemistry of alkyl groups. This includes their high basicity and their characteristic modes of interaction with alkali metals in solids tate and in solutionsofnon-coordinating solvents. [6] during synthesis, [10] leading to low or negligible yields.T he corresponding lithium compoundsa re more robusttothis specific side-reaction, [11] which allows their isolation with good yields and high purity.H owever,t he organometallic chemistry of the heaviera lkali metals [12] with their highly ionic metal-carbon interaction is furtherc omplicated by two other fundamental characteristics:a ggregation and reactivity.T he ionic interactions and the larger size of the metal atoms lead to high coordination numbers and polymeric aggregates, resulting in negligible solubility. [13] Coordinatings olvents such as benzene, toluene, or ethers, [14] which could dissolve these compounds and facilitatet heir characterization and synthetic use, suffer chemical attacks caused by the inherent reactivity. [15] Another route of decomposition is b-elimination, which affects alkyllithium compounds [16] and becomes even more destructive for the heaviera lkali metal alkyl congeners. [17] Despite these problems, it is attractive to harvest this formidable reactivity in hydrogenmetal exchange/deprotonation/metalation reactions, which cannotbea chieved by corresponding lithium compounds.

Access to Highly Reactive Alkali Metal Bases
Twoc hemically possible options can be considered to achieve this goal:E ither the organometallicc ompoundso ft he heavier alkali metals are broughti ntoam ore controllable and synthetically usable form,o rt he reactivity of organolithium compounds is increased significantly without sacrificing their already positive and very useful characteristics. [18] Schlosser described how this was achieved by the search for activating ligands for organolithium compounds. [19] Ethers, [14] (including crown ethers) and (chelating) tertiary amines, [20] are able to cause this activation of organolithium, but suffer metalation themselves under these conditions. Chelating di-alkoxides offer the Lewis donor capabilities and chemical inertness, but lack the necessary solubility in hydrocarbon solvents such as nhexane. The only remaining possibility is the use of tertiarya lkoxides with feasible solubility,s uch as lithium, sodium,o rp otassium tert-butoxides. Thec ombination of alkyllithium with lithium alkoxides produces compounds, [21] the reactivity of which (in addition-reactions) [22] does not substantially differ from the alkyllithium itself. [23] Potassium tert-butoxide as "ligand" in mixtures with alkyllithium ultimately leads to the desired increased reactivity (Scheme 2). Exactly the same reaction is used to produce alkylpotassium from the corresponding lithium compound. This metal-metal exchange reactionh ad already been investigated in depth by Lochmann and his group in the mid-sixties. [24] The positive effect of alkoxide on metalation reactions of alkenes with amylsodium (n-pentylsodium) had already been studied by Morton two decades earlier. [25] Here, the sodiuma lkoxide was formed in situ by the sacrificial reaction of ap art of the amylsodiumw ith iso-propanol. Due to the limited access to amylsodium, this line of research was not developed further.T he majority of examples refers to n-butyllithium,b ut examples using methyllithium, [26] iso-butyllithium, [27] or tert-butyllithium [28] are also reported.
An umber of excellent reviewsh ave been published by Lochmann, [23,29] Schlosser, [15,30] and others, [31] discussing the nature of these superbases, their reactivity towards organic substrates, and changes in the composition by variations of alkyl/alkoxy ratios. An umber of studies with more or less related systems were aimed at finding out the chemical nature of these superbases. Ap ublication by Bauer and Lochmann provides an excellent overview of experimentsc onducted until then. [32] However,t he majority of conclusions are drawn from products,selectivities, and yields of the reactionofcorresponding basic systems with organic substrates andn ot from the LSBs themselves. According to the interaction of the base components with each other,t he results are often inconclusive or contradictory.N evertheless,t he review [29] by Lochmann summarizes some importantc onclusions:T he reaction of alkyllithium with potassium alkoxides produces alkylpotassium, the products of subsequentr eactions also give rise to corresponding potassium compounds.T he potassium alkoxide hast ob e used in at least equimolar compounds, so its role cannot be described as merely catalytic or activating;t he contribution of the participating lithium alkoxide is less dominant. The reactivity benefitsf rom the use of potassium alkoxides bearing more branched groups to increase solubility and concentration. In summary,t hese mixtures are highly flexible systems, in which the metal-metal interchange plays an "integral part in the reaction".

Alkali Metal Alkoxides
In the vast majority of examples, the alkali metal alkoxides used in LSBs are potassium compounds (LiR/KOR'). Investigations also included mixtures of alkyllithium with sodium alkoxides [33] (LiR/NaOR') or combinations of alkyl sodium with sodiuma lkoxide [34] (NaR/NaOR') or potassium alkoxide [35] (NaR/ KOR'). Very few examples involve rubidiumo rc esium alkox-ides [36] (LiR/RbOR'a nd LiR/CsOR'). Secondary alkoxides [25] or bifunctional alkoxides (such as pinacolate) [35] are rarely used. This can be attributed to low solubility or al ack of chemical inertness. Most studies cover reactions with tertiary alkoxides. In some cases, the use of branched tertiary alkoxide (such as 2methyl-2-butoxide or 3-methyl-3-pentoxide) instead of tert-butoxide led to better results. Therefore, the mixtures featuring branched alkoxides with improved solubility are an advancement and hence called LSBs of the second generation. [23] Structural motifsf ound in solid-state structures of relevant alkali metal alkoxides (Figure 1) might be reflected in the structures of corresponding mixed aggregates. [37] In tertiary alkoxides, every alkali metal atom interacts with three oxygen atoms of three alkoxy groups, and vice versa.T his alwaysr esults in the formation of M 2 O 2 four-membered rings;a lso M 3 O 3 six-membered rings are possible. [38] LiOtBu is found in hexameric aggregates, [39] but also octameric aggregates [40] are observed.N aOtBu crystallizes as hexamer and nonamer side by side. [39] The corresponding potassium, rubidium, and cesium tert-butoxides form regularh eterocubanes. [41] The structures of alkali metal 2-methyl-2-butoxides( tert-amyloxide) follow this pattern but show increased solubility. [42] The reason fort he limitation to tetrameric structuresi nt he case of potassium, rubidium, and cesium originates from the possibility to create mutuali ntermolecular interactions, which can be optimally arranged in ap acking of tetramers. Tetramers are also found in lithium [43] and sodium [44] alkoxides, if more bulky alkoxy groups such as OCH(tBu) 2 or OC(CF 3 ) 3 are used. The co-existenceo fh examersa nd octamers (in the case of LiOtBu) providesaclue, how smaller units such as dimers are transferred between oligomersa nd mixed alkyl/alkoxy aggregates.

Working around the Chemistry of Alkali Metal Superbases
Twoi mportants tructural examples pointed in the direction of which structural motifs are to be expected in bi-metallic and hetero-anionic systems( Scheme 3). Harder and Streitwieser used the reactiono fs odium phenoxide with n-butyllithium to produce an intramolecular combination of lithium phenoxide with as odium-metalated benzyl-position. [45] This structurally characterized molecule combines ah eaviera lkali metal (sodium) carbon interaction with lithium oxygen interactions, two expectable arrangements in LSBs.
As tep further is ar esult by Mulvey et. al.,w hich consisted of the combination of al ithiated primary amine with potassium tert-butoxide, which is able to perform am etalation of toluene. [46] Here, the metal atoms (lithium/potassium) are combined with an alkoxide/amide framework;t he amide anion is isoelectronic to the corresponding alkyl groups (e.g. in n-butyllithium). However,b oth examples, thoughp resenting relevant metal-element interactions, also feature the possibility of additional Lewis-base/metal interactions (p-o ra dditional free electron pairs) combined with less Brønsted-basic groups (benzylic M-CH 2 Ph and M-NR 2 versus aliphatic M-CR 3 ).
An ew level was reached in the publication by Strohmann et al.: [47] here, as tructure wasp resentedt hat incorporates a combination of phenyll ithium and phenyl potassium with lithium tert-butoxide, with THFa sa dditional donor (Scheme 4). Both groups of metalse xhibit interactions with alkoxide oxygen atoms and phenylcarbon atoms at the same time.
Though the basicity towards toluene could be demonstrated and relevant structuralg oals were achieved, three more obstacles are evident in the way towards structurally identified Lochmann-Schlosser superbases:( i) The basicity of the phenyl group can be expected to be lower in comparison to the corresponding alkyl groups, based on the highere lectronegativity of the phenylc arbon (vide supra). (ii)The phenyl groups can interact with metal cations both through the electron pair of the ipso-carbon atom and as well through the p-systemo ft he Scheme4.Schematicrepresentation of the mixed aggregate compoundr eported by Strohmann. Only one of the three relevant metal-element interactions is highlighted;additional coordinating THF and benzene molecules are omitted for clarity. phenylr ing. [2] This creates structuralm otifs, which cannot be present in alkyl systems. (iii)The use of THF adds an ew component to the system that adds additional structural properties as well as proton-acidicr eactivity,w hich must be suppressed by low temperatures when used in Lochmann-Schlosser superbasic mixtures. However, the relevance of this example, which no longer represents am ere model system, cannotb eu nderestimated with respectt oa romatic reactionp roducts resulting from superbasic metalations,o rf or the Lochmann-Schlosser superbase chemistry in THFa tlow temperatures.
What is the main obstacle on the way to achievet he isolation and characterization of compounds very similar (or even identical) to those expected to be present in LSBs?T he main challenge is the extremely low solubility of the corresponding alkylpotassium, whichi su ltimately always formed in these superbasicm ixtures (Scheme2). This forces every equilibrium involvinga lkylpotassium to the product side, making it impossible to identify relevant compounds due to their very low concentration.A ne xample of ah ighly soluble Lochmann-Schlosser superbase, whichi sf ormed in hexane by combining 2-ethylhexyllithium with potassium tert-amyloxide [KOtAm] (Scheme 5), is reported, [48] however,w ithout furtherd ata on the mixture itself. This might be caused by the low thermal stabilityo ft he formed 2-ethylhexylpotassium, which is reported to be soluble in hexane.
In our own research, we found that mixtures of potassium tert-b utoxide with neopentyllithium [49] [LiCH 2 tBu, LiNp] also produced solid neopentylpotassium [50] [KCH 2 tBu, KNp].T he precipitate was isolated by filtration.H owever,t he yield of isolated KNp was considerably lower comparedt ot he results of similar potassium compounds. [51] In fact, it was possible to isolate ac rystalline solid from the filtrate, which contained all four components expected to be presents in LSBs:Lithium, potassium,a lkyl groups, and alkoxide groups. [52] One reason for the unexpected high solubility of the neopentyl/alkoxide mixed aggregates is the structural similarity between tertbutoxy groups [O-tBu]a nd Np groups [CH 2 -tBu]( Scheme 6). The structuralm imicryo ft he Np group leads to as tatistical replacemento fO tBu groups, leadingt od ecreased symmetry of the resulting molecules, and in turn to an increased solubility.

Possible Combinations in aFour-Component System
In order to obtain ab etter overview of the available results, a general reflection of the possible methods to combine lithium and potassium atoms with alkoxy and alkyl groupsi no ne compound might be helpful. Schlosser did not expect the formationo fasingle mixed aggregate, but an entire family of such adducts. [19] In this respect, as ystematic approacht o which familym embers can be expected in such systems may produce additional insights.
The hypothetical combinationo fa ll four components, lithium, potassium, alkyl groups [R],a nd alkoxy groups [OR'],w ill result in the formation of ac ompound with the formula Li x K y R z (OR') x + yÀz .B earing in mind that the single compounds LiR, KR, LiOR', and KOR' possess lower or at least different reactivity than mixed aggregates;t he simplest relevant combination is formed by one equivalent LiR and KOR' each:L iKR(OR'). This compound could also be seen as hetero-dimer,w ith two metal atoms present: LiR·KOR' (or LiOR'·KR). In reality,t he number of metalsw ill be two or higher.E xamples from alkali metal alkoxide compounds suggest the formation of heterotetramers, -hexamers,-octamers, or -nonamers. However, also hetero-forms of trimers, pentamers, or heptamers are feasible. Hetero-octamers are of particular interest, because two model compounds possesst he composition of the mixed alkyl/alkoxy lithium compound Li 8 (nBu) 4 (OtBu) 4 (or 4LinBu·4LiOtBu), [21] and the lithium/potassium alkoxide Li 4 K 4 (OtBu) 8 (or 4LiOtBu·4KOt-Bu). [53] Both compounds may offer ap ossible structural design for aggregatesc ombining all four components.T he question is whether or not the introductiono fa lkyl groups or potassium atomsw ill have am ore dominante ffect on the resulting structure.
To obtain ac learer picture of the possible hetero-oligomers, one should keep in mind that the number of cationic and anionic units must be the same. This obviousc ondition simplifies the formula Li x K y R z (OR') x + yÀz with three variables to a system with only two fractional variables:L i a K (1-a) R b (OR') (1Àb) .T o avoid fractional numbers of atoms or groups, it is possible to add av ariable m, which also reflects the degree of oligomerization:L i ma K m(1Àa) R mb (OR') m(1Àb) .T he simplification of af ractional Li/K ratio (a)a nd af ractional R/OR' ratio (b) allows plotting all possible variations in at wo dimensionalc oordination system referring to aand b( Figure2). [54] In the chosen representation, pure lithium compounds can be found on the left, pure potassium compounds on the right border.P ure alkoxides are found at the top of the diagram, alkyl compounds at the lower end.
The fractional ratios shown on the diagram axes are defined by the number n of each component:L i/K: n(K)/[n(Li) + n(K)] and R/OR': n(R)/[n(OR') + n(R)].A ccordingly,t he pure alkyl or alkoxy compounds can be found in the corners of the diagram:l ithium alkoxide LiOR' at the top left, potassium alkoxide KOR' at the top right corner,a lkyllithium LiR and alkylpotassium KR in the bottom left and bottom right corners, respectively.E very mixed aggregate consistingo ft hree or more of these four components (Li, K, R, OR'), independentf rom its existence in solution or solid state, can be placed on the edges (three components) or the area (four components)o ft his diagram. This raw diagram can then be populated with substances relevant for this type of system ( Figure 3): lithium tert-butoxide LiOtBu, whichc an be found both in hexameric [39] [Li 6 (OtBu) 6 ]o r octameric [40] [Li 8 (OtBu) 8 ]f orm;t etrameric potassium tert-butoxide [K 4 (OtBu) 4 ]; [41a] and hexameric butyllithium[ Li 6 (nBu) 6 ]. [55] No relevant alkylpotassium compound is known,b ut examples of alkylsodium [9c] or donor coordinated alkylpotassium [51] hint towards the possible existence of tetrameric units [K 4 R 4 ].
The two substances Li 8 (nBu) 4 (OtBu) 4 [21] and Li 4 K 4 (OtBu) 8 , [53a] which act as potentials tructural modelsf or Lochmann-Schlossers uperbases, are found at the center of the respective edges of the diagram. The first compound has as tructure similar to octameric [Li 8 (OtBu) 8 ], with the n-butyl groups on the peripheral positions of the molecule. AL ochmann-Schlosser superbase, consisting of an ideal 1:1c ombination of n-butyllithium and potassium tert-butoxide, would be situatedi nt he middle of the diagram. The lack of compounds populating the diagram area can be attributed to the removal of alkyl potassium (right lower corner) from every conceivable mixed aggregate containing both potassium and alkyl groups. The very low solubility in most solvents removes it from the correspondinge quilibria (Scheme 2). The remaining compounds withoutpotassium or alkyl group are enriched with lithium alkoxide, which placest hem on the edges of the diagram. The mixed aggregate of the potassium-rich compound reported by Strohmann [47] (Scheme 4) could be seen as an exception. However,i nt his case, the strongly coordinating solvent THF plays a dominant role and this discussion is restricted to systems in the absence of donors olvents.
The same positiono fm ixed aggregates with the same fractional composition of each component,b ut with differentd egrees of oligomerization (e.g. Li 2 K 2 R 2 (OR') 2 ,L i 3 K 3 R 3 (OR') 3 ,o r Li 4 K 4 R 4 (OR') 4 ), is ad rawback of this plot, but experimental results will show whether these combinations exist side by side.

The Neopentyl Group as Key to Mixed Alkyl/ Alkoxy Aggregates
After the preparationo ft he field, it is necessary to find aw ay to solubilize alkyl potassium to lift the restriction to the sideline. As described above, we managed this by using neopentyllithium insteado fn-butyllithium as alkyl source in LSB mixtures. [52] The combination of structurals imilarity to the tertbutoxy group and the introduction of structurald isorder hinderingc rystallization and precipitation (vide supra) consider-   4 ,and (LinBu) 6 )] in the corners (circles). Int he middle of the right and top edges are the compounds Li 8 (nBu) 4 (OtBu) 4 [21] and Li 4 K 4 (OtBu) 8 . [53a] The molecular 1:1m ixed aggregate of n-butyllithium and potassium tert-butoxider epresents an ideal LSB situatedi nthe middle of the diagram (cross).
ably increases thes olubility of the corresponding compounds. An additional advantage of neopentyl and tert-butoxy from a practical point of view is the simplicity of the resulting 1 HNMR spectra.T he presence of three resonances in 1 HNMR (Np:C H 2 and C(CH 3 ) 3 ;O tBu:C (CH 3 ) 3 )m akes it easier to identify these groups even in structurally different species.
In the attempt to produce neopentylpotassium [KNp] in a reactionu sing neopentyllithium and potassium tert-butoxide (Scheme 2), we noticed the poor yield of KNp (< 30 %). Considering the low solubility of pure samples of KNp in C 6 D 12 for NMR spectroscopy,w er ealized that the excess KNp must be part of as oluble mixed aggregate. From the mother liquid of the reaction mixture, we obtained crystals at À30 8Cw ith the approximate formula Li 4 K 4 Np 3 (OtBu) 5 (1,Scheme7).

Behavior of Mixtures of Neopentyllithium and Potassium tert-Butoxide in Solution
The disordered and fractional compositiono f1 and its close relationship with Mulvey's mixed alkoxy compound Li 4 K 4 (OtBu) 8 (by analogy compound 1 with no neopentyl group: 1 0 )o pened up interesting possibilities. Disorderb etween two chemically different groups is not desirable in terms of structurala ccuracy.H owever here, the structure of 1 represents two or more members of the same structural family with different numberso fn eopentyl groups at the same time. 1 can be seen as ap resentation of different orientationso f1 3 but, at the same time, it also represents the compounds 1 2 and potentially 1 4 with am issing or an additional neopentyl group, respectively.B ya dding increasing amountso fL iNp to Li 4 K 4 (OtBu) 8, 1 0 (continuousv ariation [56] ), it was possible to study its transformationi nto 1 by 1 HNMR spectroscopy. [52] LiNp was added to equimolar mixtures of LiOtBu and KOtBu in n-hexanei ni ncreasing amounts; the crystals obtained at À30 8Cw ere characterized by 1 HNMR spectroscopy in deuterated cyclohexane [C 6 D 12 ]. The results show that the intensity of one of the two distinguishable tBuO signals of Li 4 K 4 (OtBu) 8 is reduced whileb oth singlet resonances of corresponding neopentyl groups (CH 2 and tBu)i ncrease in intensity.T his observation suggests the presence of Li 4 K 4 Np(OtBu) 7 (1 1 with one alkyl group) in solution with the arrangements imilar to Li 4 K 4 (OtBu) 8 (1 0 ), as was anticipated based on the group disorder (Np/OtBu) present in the structure of 1.A ccordingly,t his replacement of OtBu by Np can be expected to occur exclusively in the peripheral position of a[ (OtBu)Li(OtBu) 2 Li(OtBu)] 2À unit (Scheme 8). Further addition of Np leads to the formation of Li 4 K 4 Np 2 (OtBu) 6 (1 2 ). However,t he addition of the second neopentyl group leads to the formation of two isomers: introduction of as econd Np in the same Li(OtBu) 2 Li centered unit togetherw ith the first Np group (1 2 -I), or it can be placed next to the Li(OtBu) 2 Li unit on the other side of the K 4 plane (1 2 -II). While 1 2 -I exhibits as ymmetric arrangement, in 1 2 -II,t he two protons of the CH 2 -Np group have ad ifferent chemical environment,regardless of their rotational orientation.
This break in symmetry manifests itself in the 1 HNMR spectrum by the diastereotopic splitting of the CH 2 signal into two symmetric duplets. In the 1 HNMR spectrum, 1 2 -I just shows a singlet replacing or adding to the singlet of 1 1 .T hese findings support the existence of 1 1 , 1 2 -I,a nd 1 2 -II in solution at room temperature. When even more LiNp is used in preparation of the crystalline samples, the corresponding 1 HNMR spectra become more complicated. For Li 4 K 4 Np 3 (OtBu) 5 (1 3 % 1)i n C 6 D 12 ,t he expected signals in the CH 2 -Np region are found (singlet + two diastereotopic duplets), but an ew broad signal appearsa lso. This shows that in solution compound 1,w hich was isolated as pure crystalline solid, partially falls apart into other species. In such solutions the outcomeo fc rystallization dependso nt he concentration and the solubility of the compounds formed in the solution equilibrium at ag iven temperature. The replacement of more than two or three of the peripheral OtBu groups in Li 4 K 4 (OtBu) 8 leads to as tructurali nstability due to weaker metal-Np interactions. Accordingly,e ven higher Np contents did not lead to the formation of the "ideal" neopentyl LSB, but to the formation of mixed aggregates enriched with lithium and alkyl/alkoxy ratios close to 1/1. The two compounds Li 4 K 3 Np 3 (OtBu) 4 , 2 (% 2 3 )( Figure 5) and Li 4 KNp 2 (OtBu) 3 , 3 (% 3 2 )( Figure6)w ere isolated from such mixtures and characterized by X-ray crystallography and NMR spectroscopy.A gain, both compounds exhibit partial group disorderb etween OtBu and Np as seen before in compound 1, suggesting the existence of neopentyl-rich compounds such as Li 4 K 3 Np 4 (OtBu) 3 , 2 4 and Li 4 KNp 3 (OtBu) 2 , 3 3 .
The structure of 3 shows the formation of dimers. Additional intermolecular interactions are found between the potassium atom and the CH 2 unit of the neighboring Np group. The two protons are orientated towards the potassium in an agostic type interaction. This weaker long-range interaction is possible because both the potassium atoms and the neopentyl groups are found in exposed positions, whichm akes them more accessiblef or intermolecular interactions.
The interplay of compounds 1, 2,a nd 3 can be interpreted graphically (Figure 7): Starting from Mulvey's compound Li 4 K 4 (OtBu) 8 (1 0 ,situated on the middle of the top edge of the diagram) it is possible to perform ap rogressive replacement of OtBu groups with Np. The result is am ovement downwards towardst he middle of the diagram, passingc ompounds 1 1 , 1 2 , and 1 3 .S tartingw ith compound 1 3 ,t he system is affected by emerging equilibria in solution. This prevents reachingh ypothetical LSB Li 4 K 4 Np 4 (OtBu) 4 , 1 4 ,a nd leads to the formation of compounds enriched with lithium such as compounds 2 and 3 instead. According to their composition,c ompounds of the family 2 and 3 are found in the left half of the diagram.
While the compounds Li 4 K 4 (OtBu) 8 , 1 1 ,a nd 1 2 can be observed in solution with some confidence,c ompounds 1, 2,a nd 3 were isolated as crystals from such solutions. In solution,t he broader 1 HNMR signals of 2, 3,a nd 1 4 are indistinguishable or cannotb ea ssigned because of their participation in fast equilibria. [57] However,t he absence of strongly coordinating solvents and the structural consistency of solid state and solution in the case of 1,t heir presence in non-donating solvents [6] such as nhexane or cyclohexane can be anticipated. In reference to the classical LSB, which uses n-butylg roups,i ti sl ikely that the stability of an n-butyl compound of af ormulation similart o1, loses its structurali ntegrity in an even earliers tage. In analogy Scheme8.
Step by step exchange of tert-butoxy groups in Li 4 K 4 (OtBu) 8 (1 0 )with neopentyl groups, leadingt ot he formationo fneopentyl-enriched compounds 1 1 , 1 2 ,and 1 3 (the superscripted number describes the number of alkyl groups in the mixeda ggregate).Further introductiono fn eopentyl groups leads to the formation of the lithium-rich compounds 2 3 and 3 2 beforereachingthe ideal combination of hypothetical 1 4 .  to 1,h igher n-butyl contents in such compounds would cause ejectionofn-butylpotassium units due to its negligible solubility.S till, the presence of ac ompound such as Li 4 K 4 (nBu)(OtBu) 7 can be anticipated in low concentrations in solution.

Structural Motifs, Part 1: Mixed Aggregates
An umber of structural motifs potentially present in Lochmann-Schlosser superbases were discussed by Schlosser, [27] which he deduced from the combinationo ft he involved components.T hese motifs (Scheme 9) range from merely "activated alkyllithium"t op urea lkylpotassium. Schlosser suggested the existence of an ate-complex, ap otassium alkyl/alkoxy lithiate (B), alkyllithium coordinated to potassium alkoxide (C), alkylpotassium coordinated to lithium alkoxide (E), and also a mixed aggregate [37] (or symmetrical adduct) of alkyllithium and potassiuma lkoxide (D). Ad escription as "co-complex" [37d] (of LiR and KOR') emphasizes the mixed-metal charactero ft hese combinations; the notation "mixed aggregate" also involves homo-oru ni-metallic systems. [31] Am ixed lithium alkyl/alkoxy compound [57] (A)w as not considered because of its lower reactivity. [23] With ah igher degree of aggregation (hetero-tetramer or highera ggregated), it is likely to find two or more of these motifs in mixed aggregates; there is also room for interpretation. In the following this is demonstrated on the structure of hypothetical compound 1 4 derived from structure of 1 (Figure 8).
By regarding the interactions of the four potassium atoms as purely electrostatic,t he two remaining units have to be anionic, making the compound ap otassium lithiate (B). The structure of hypothetical 1 4 can also be derived from a( open) tetrameric potassium alkoxide, [41a, 42] which is hostingf our monomeric alkyllithium units (C). Similarly,t he structure of 1 4 can also be regarded as tetrameric alkylpotassium, which accommodates two dimeric lithium alkoxide units (E). The motifs C and E show the characteristics of both, the startingm aterials [LiR and KOR'] and the products [KR and LiOR'],a tt he same time. The same situation was demonstrated by Lochmann on the basis of ah etero-dimer. [23] The separation into lithium alkoxide units and alkylpotassium can be seen even clearer in Li 4 K 3 Np 4 (OtBu) 3 , 2 4 ,w here one potassium atom of at etrameric K 4 Np 4 unit (or K 4 Np 3 OtBu unit in 2 3 )i sr eplaced by ac ationic Li 4 (OtBu) 3 unit. Similarly,d imeric 3 2 can be considered as ad imeric[ KNp] 2 unit, which is coordinated by two Li 4 Np(OtBu) 3 units. Motif D is present in compounds of the families 1, 2,a nd 3 as ad istorted square, where lithium and potassium are bridged by both alkoxide oxygen and an alkyl carbon atom. Another arrangement present in 1 4 is as quare formed by two potassium atoms and the alkoxide oxygen and the alkyl carbon atom (F). This motif was not anticipated before.I nc ompound 1 4 ,i tw ould be consistentw ith ac hemically less meaningful lithium potassiate. [58] However,i n am ixed alkyl/alkoxy potassium compound (in the absence of lithium)i twould be the only relevant motif.

Degradation of Mixed Aggregates Following Metalation Reactions
The transformation of hypothetical 1 4 into 2 3 by loss of one unit of neopentylpotassium (Scheme 8) provides an insight into what happens to the base, when it is consumed in ar eaction with an acidic substrate. Ap roton is transferredf rom the acidic substrate to an alkyl group,w hich is then released to the solution as alkane. Thed eprotonated substrate anion replaces the alkyl group in the mixed aggregate (Scheme 10).
In contrastt ot he "monodentate" alkyl group with as ingle directedi nteraction, the anion of the substrate anion (e.g. a phenyla nion) will exhibit additional electron pairs allowing secondary (intermolecular) interactions to potassium cations. [2b] This will result in the precipitation of the insoluble corresponding potassium-substrate compound, leaving behind am ixed aggregate lacking one alkylpotassium unit. The consequence is an ew compound enriched by both lithium and alkoxide, resulting in am ovement to the upper left corner away from KR in the mixed aggregate diagram (Figure 9). This reactionp ath can be conceived for every known or hypotheticalm ixed aggregate containing highly basic alkyl groups. Ap ossible, but not yet observed, mixed aggregate in the sequence Li 4 K 4 Np z (OtBu) (8Àz) (z = 3: 1 3 )a nd Li 4 K 3 Np z (OtBu) (7Àz) (z = 3: 2 3 ) would be hetero-hexameric Li 4 K 2 Np z (OtBu) (6Àz) (Q Z ). This hypotheticalc ompound Q of unknown structure would lead, after formal loss of another KNp unit, to Li 4 KNp z (OtBu) (5Àz) (z = 2: 3 2 ). After af inal step, LiOtBu will stay behind,w hich in turn can interfere with all the other neopentyl-containing speciest of orm lithium alkoxide enriched species.

Homometallic Potassium Mixed Aggregates
The successive transformation of one mixed aggregate into another during ar eaction with an organic substrate and the participation of thesec ompounds in interchanging equilibria is a major obstacle when it comes to the description of the reactivity of the involved bases.T he assignment of both NMR and vibrational spectroscopic data of involved chemical groups to distinguishable species will not be an easy task. Isotopically enriched compoundsm ay help to decipher such systems.
The formation of mixed aggregatesw ithout lithium will simplify mattersconsiderably.Leaving out lithium as afourth component, it is possible to find out whether the cooperativity (or synergy) of two different metals [59] or the presence of both alkyl and alkoxy groups side by side is required to obtain superbasicity.T he synergyo fn umerous mixed-metal systems usually depends on the reactivity-enhancing effect of ap olar metal compound on as econd less reactive, less polar organometallicc ompound, whilet he reactivity of bi-and homo-metallic LSBs or relatedm ixed aggregates is described by the taming effect of the added alkoxide on the fiercely reactive alkali metal alkyl compound. The potential of homometallic bases wasd emonstrated by the outstanding reactivity of sodium alkyl/alkoxy mixtures [25] or even by al ithium amide/ alkyl mixture, which was able to metalate cyclopentadienyllithium for as econd time. [60] Bases consisting of potassium alkoxide and alkylpotassiumcan be expected to show less structural diversity and even higherb asic reactivity compared to systems using lithium and potassium. In the diagram representation of mixed aggregates, thesepotassium compounds are found only on the right edge of the diagram (Figure 10.) The same is true for products based on ar eaction of this base with organic substrates and subsequent removal of insoluble potassium product compound (Scheme 11). The residual mixed aggregates will contain more alkoxide in relation to alkyl groups. Ultimately,asolution of potassium alkoxide will be left, whichc an easily be separatedf rom the insoluble po-Scheme10. Reactionofam ixed aggregate with an organic substrate S-H leadingtot he formationa nd elimination of ap otassiateds ubstrate. The reaction leads to af ormal loss of alkylpotassium from the mixed aggregate. The residualmixed aggregate can act as ab ase itself again or take part in equilibria, forming new basic species enriched with lithium and alkoxide. Figure 9. Graphical representation of reaction sequencesd uring the reaction of mixeda ggregates with an organic substrate. The arrows represent the formal loss of an alkylpotassium unit, leadingtoanew mixeda ggregate.
[Q z ] represents ah ypothetical hetero-hexamer Li 4 K 2 R z (OR') (6Àz) .The final product will be LiOR', or considerably less basic mixed aggregates of the formula Li x K y (OR') (x + y) or Li x R z (OR') (xÀz) ,d epending on the original composition. tassium compound. The latter compound could also contain stoichiometric amountso fa lkoxide in somecases. [61] The formation of ap otassium mixed alkyl/alkoxy aggregate was achieved by dissolving KNp in as olution of potassium tert-amyloxide (potassium 2-methyl-2-butoxide, KOtAm). [54] The latter compound was used because of its increased solubility in non-donating solvents such as n-hexane. [42] However,t he poor thermals tability of KNp madei tf avorable to use LiNp with an excesso fK O tAm (Scheme 12). This produces KNp in situ, the by-product LiOtAm is trappeds imultaneously by the excess KOtAm to form Li 4 K 4 (OtAm) 8 .I nterestingly,n of ormation of an aggregate waso bserved by mixing n-butylpotassium and KOtAm. [61a] The product, which was isolated as large yellowish crystals from as olution of n-hexane at À30 8C, was identified as K 4 Np(OtAm) 3 (4)b yN MR spectroscopy and X-ray diffractometry.T he compound consists of ah etero-tetramer of one KNp unit and three KOtAm units arrangedi nahetero-cubus ( Figure 11).
In contrastt op ure potassium alkoxides, which preferably crystallize in tetrameric form, [41a, 42] the mixed aggregate 4 (= 4 1 )o ffers an additional (nucleophilic) coordination site for the coordinatively unsaturated potassium atoms. The interaction is bilateral, which results in the formation of ad imer of hetero-tetramers (similar to compound 3). Thisi nteraction is easily released in solution, as could be demonstrated by DOSY NMR [62] of 4 in deuterated cyclohexane. Ad ecreasei nc oncentrationr esulted in the formation of al ighter,m ore mobile neopentyl-containing species. The formation of ah exameric form cannotb er uled out,b ut the preference of potassium alkoxide [41a, 42] and alkylpotassium [13,51] for the formation of tetramers makes this unlikely. The example of this dimerization might also hint to what happenst om ore neopentyl-rich hetero-tetramers. K 4 Np 2 (OtAm) 2 (4 2 )c an be expected to form badly soluble linear or zig-zag polymericc hains, similar to the structure of [NaCH 2 SiMe 3 ] 1 . [9c] Accordingly,t he likewise hypothetical forms K 4 Np 3 (OtAm) (4 3 )a nd tetrameric KNp (by analogy: 4 4 ) might form two-o rt hree-dimensional networks in the same way.
In contrast to pure KNp, [50] compound 4 has ah igher thermal stability,d emonstrated by as low decomposition over several hours as solid or in solution. This adds as tabilizing effect to the solubilizing capabilities of excesspotassium alkoxide.
In addition, the comparable simple arrangemento f4 exhibits motif F (Scheme 9), af our-membered ring of two potassium atoms bridged by an alkyl group and an alkoxy group, respectively.I nt his instance, it is comparable to compounds 1 and 2; however,b ecause of the absence of lithium,a ll the other motifs (A-E)a re excluded in 4.O verall, the presence of com- Figure 10. Graphicr epresentation of the chemical composition of compound 4,respectively 4 1 .The arrow represents the formal loss of an eopentylpotassium unit duringareaction with an organic substrate, producing potassium alkoxide.
Scheme11. Reactiono fapotassium mixeda ggregate with an organic substrate S-H leadingtot he formation of ap otassiated substrate. The reaction leads to aformal elimination of alkylpotassium from the mixed aggregate, the residualcompound is inactive as base in this context if y = 1.
Scheme12. Formation of compound 4 by reaction of neopentylpotassium with three equivalents potassium tert-amyloxide or by reaction of neopentyllithium( LiNp) withe xcess potassium tert-amyloxide (KOtAm). pounds similar to 4 can be expected in superbasic mixtures using an excess of soluble potassium alkoxides. [61] Structural Motifs, Part 2: Alkyl-Metal Interactions Ta king into account the presence of several motifs (Scheme 9A-F)i nasingle compound and the possible coexistence of severalspecies in solution, it is very difficult to connect these structural features to the reactivity of such mixtures. The focus on the environment of the very basic alkyl group leads to al ess complicated picture. In this approach, the role of the alkoxy groups is more or less reduced to the role of ac hemically inert structural support or solubilizing co-reagent. In many alkyl lithium compoundsa nd in the neopentyl mixed aggregates 1, 2,a nd 3,t he alkyl group is found in a m 3 bridging positionb etween three metal atoms. A m 4 position was found only fort he oxygen atom of alkoxy groups. This results in four possible environments for the metalated a-carbona tom:L i 3 (I), Li 2 K( II), LiK 2 (III), and K 3 (IV)( Scheme 13).
Motif I is found in numerous alkyl compounds of lithium. [3] Motifs II, III,a nd IV are very rare in structurally characterized alkyl potassium compounds. With the exception of ill-defined methyl sodium with lithium atoms statistically replacing sodium atoms, [13] there is no information availablea bout mixed alkali metal alkyl compounds. While there are numerous examples of combinations of alkali metals with less electropositive metals such as magnesium or zinc, [59] the structuralm otif of alkyl groups coordinated by am ixed alkali metal environment is so far restricted to compounds 1 (motif III), [52] 2 (motifs III + IV), [52] and 3 (motif II) [54] with their mixed alkyl/alkoxy arrangement. Motif IV with aK 3 environmenti sa lso present in pure alkyl potassium compounds such as polymeric methyl potassium [63] or tetrameric trimethylsilylmethyl potassium coordinated by TMEDA. [51] Due to the small number of available relevant structures and the positional disorder of the involved neopentyl groups,i ti s difficult to obtain ar eliablep icture of the steric and electronic bondings ituation between an alkyl group and am ixed metallic environment. However,i nt he case of motif III and IV,t he findings are backed by computational modelso fc ompounds 1 and 2. [52] The metalated a-carbon of ap rimary alkyl group is bonded to at rimetallic lithium platform via af our-center twoelectron bond, [64] if an substantial covalent contribution to the interaction is assumed. Based on the positions of the attached alkyl group and the (less reliable) positions of theh ydrogen atoms, it is possible to speculate about the position of the electron pair of the sp 3 hybridized carbon interacting with the metal atoms. In case of alkyllithium compounds, the electron pair points towardst he space between the three lithium atoms. Here, the orientation of the attached alkyl group and the two hydrogen atoms also depends on steric interactions such as b-CH 2 lithium attractions. In some basic lithium compounds, this leads to an eclipsed conformation [55] according to the three lithium atoms (motif I,S cheme13). If one or more potassium atoms are present in the trimetallic platform, ad ifferent structuralm otif appears: the potassium atom, the acarbon,a nd the carbon of the alkyl group (here:n eopentyl) form an angle larger than % 1608.This approximately linear formation places the two protons of the neopentyl-CH 2 unit in close proximity of the potassium atom, comparable to an agostici nteraction. In motif II,w hich is present in the (barely disordered) structure of compound 3,t he position of the tertbutyl group (as wella st he two protons) suggests that the electron pair of the carbon-metal interaction points in the middle of the two lithium atoms (Figure 12). This would be in accordance with am ore covalent three-centert wo-electron bond between carbon and lithium and am ore electrostatic interaction between carbon and potassium. Motif III also exhibits the linear K-alkyl arrangementa nd ad irect interaction of lithium with the a-carbon atom. However,s tructural and computational data suggest that the electron pair of this interaction is slightly displaced towards the second potassium atom ( Figure 12).
Similar to the tri-lithium motif I,t he a-carbon atom is situated rather symmetrically over the K 3 triangleinmotif IV.However,t he linear K-C a -C arrangementc auses the "free" electron pair at the a-carbon pair to point towards the area between the two other potassium atoms.
Scheme13. Possible homometallic and heterometallic environmentsfor alkali metal/alkyl interactions in mixed lithium/potassium compounds. The tri-metallic platform includes the motifsL i 3 (I), Li 2 K( II), LiK 2 (III), and K 3 (IV). The dashed linesrepresent topologicalmetal-metald istances and agostic type potassium/CH 2 -interactions. Regarding the structures of compounds 1, 2,a nd 3 and the intrinsic structural motifs, it is an interesting point that the acarbon atom of the neopentyl group is alwayss ituatedo ver a trimetallic platform possessing as many potassium atoms as possible. However,t he explanation for this behavior lies not in ac onceivable affinity of the alkyl group towards potassium. The ability of lithium to form (polar) covalent bonds to carbon in comparison to more ionic or electrostatic potassium carbon interactions (Li-C:c ovalentb ut polar;K -C:m ainly electrostatic; see Scheme13, Figure 12) also contradicts this purely topological fact. The reason is found in the optimal interaction between lithium as hard Lewis acid and alkoxide as hard Lewis base, forcing potassium atoms and alkyl groups into the same structuralc orner.T he structural OtBu/Np mimicry prevents the expulsion and self-aggregation of the soft Lewis acid/base pair KNp from the complexa nd endingu pi na ni nsoluble compound. Accordingly,t he Np group is tolerated in the soluble mixed aggregate, although finding itselfi nasterically exposed situation promoting its reactivity.T he same situation is conceivable for other alkyl groups, although in lower concentration or in solid state. In this enforced mismatch of potassium and alkyl group,w hich results in at empestuous chemicalr elationship, lays an important reason for the singular reactivity of these mixed aggregates.

Structural Motifs and Superbasicity
The essential part of every organometallic superbase is an alkyl group bonded to one or more electropositive metals.I nt he case of mixed aggregate alkali metal superbases, this central alkyl/alkali metal arrangementi ss upported by further alkali metal alkoxide units. Ad irect involvement of the alkoxy groups as intermediate base cannot be ruled out so far.S imilar two-step reactions were observed in mixed metal alkyl/amide bases, [65] but here, the basicity of the involved amido groups is considerably higher than the basicityo ftert-alkoxides. [66] It is much more plausible that the alkyl group alone is acting as basic group in corresponding transition states.
Considering the vital role of the alkyl group, the bonded alkali metal atoms will have am ore dominant effect on its reactivity than their coordination by alkoxide anions. Then the main tasks of the alkoxides would be to provide the architecture for the tri-metallic platform supporting the alkyl group, and to enhance the stability and solubility of such aggregates.
The constitution of the tri-metallic platform will have considerable influence of the reactivity of the alkyl group. First, the electro-positivity and the size of the involved alkali metals will have al arge effect on the metal carbon bond polarity.I no ther words, the stabilization of the negative chargeo ft he carbanionic alkyl group will also affect the basicity.S econd, the Lewis acidic nature of the alkali metal (Li:h ard, K: soft) offers organics ubstrates ad ockings itei na dvance of metalation. [67] Functional groups in the substrate with donor atoms such as nitrogen or oxygen will prefer lithium as 'Lewis acid, while softer p-electron system of aromatic compounds will interact preferably with potassium atoms. [68] Mixed metal species can offer both coordination modesi nt he same time. In the course of am etalation,i ti sa lso important to consider the reaction path including the transition state and the formed products (Scheme14). The structuralm otifs I-IV will have ac onsiderable influence of the energyofthe transition state, which will determine the kinetics of the reaction and therefore the regio-selectivity of the outcome. The energetic stabilization of the final products is also of similar importance.
It is obvious that each organic substrate will show adifferent behavior towards the metalation platform, depending on the number of involved lithium and potassium atoms, respectively. This appliest ot he pre-complexation, the transition state, and the ultimate product.
An actual study of reactions, which will shed light on the reactivity of these structuralm otifs, will be complicated by a number of problems. These include the co-existence of different structuralm otifs in the same molecule, interchanging equilibria between different species, and the "morphological evolution" during the reaction. Theoretical calculations, which allow the study of well-defineds pecies, are an alternative.

Assessment of the Actual BasicityofA lkali Metal Superbases
The practical determination of the absolute basicity of superbasic alkali metal compounds is not an easyt ask, because of their high reactivity and poor solubility.T his is furtherc omplicated by the coexistence of different speciesw ith severalp otentially basic sites and ongoing interchanging equilibriai ns olution.T he poor solubility of the products formed in the courseo fametalationm akes it impossible in mostc ases to study the positiono ft he equilibrium (thermodynamic basicity) or the speed of the reaction (kinetic basicity). [7a] Another possibility is to find the metalation threshold [35] of such bases by checking which hydrocarbon can be metalated andw hich not. [15] However,ifo nly aliphatic or cyclic hydrocarbons such as pentaneo rc yclohexane escapem etalation by Lochmann-Schlosserb ases, then there is little space left for further differentiationoft he basicity of such bases.
One characteristic, which is restricted to alkali metal superbases, is the possibility of polymetalations of arenes. Dimetalations of benzene [69] and naphthalene [69b, 70] were achieved exclusively by bases containing the heavier alkali metals sodium or potassium.
Severalp oly-metalations of ferrocene were achieved by compounds including heavier alkali metals, [71] even tetra-metalations by mixed metal (synergic) sodium magnesiates and very recently also by sodium zincates are reported. [72] But also alkyllithium or donor-activated alkyllithium are able to perform polymetalations. [73] Metalation of Ferrocene with Potassium Alkyl/ Alkoxy Aggregates Compound 4 was successfully used in at etra-metalation of ferrocene. [54] Ferrocene is as uitable test-substrate for metalation for severalr easons. It is air-stable,s olid and can easily be added in smallq uantities;t he yellow-orange color of ferrocene changes to red when metalation occurs;f inally,m etalated ferrocene is reasonably stable at ambient temperature and can be reactedw ith ar ange of electrophilicr eagents. However, like many other examples of poly-metalated ferrocene, the red metalated product of the reaction of ferrocene with af ive-fold excess of 4 formed in situ in n-hexane is completely insoluble in inert solvents and highly reactive. This prevented the spectroscopic and structuralc haracterization of the metalated solid. Despite the solubility of 4 in n-hexane, the product of this reaction shares the same fate as countless other aromatic compounds metalated by Lochmann-Schlosser superbases:t heir composition and structures remainamystery.Adestructive hydrolysis in this case proved the presence of ferrocenea nd alkoxide, and absence of neopentane in the metalated product. The reaction with excessC O 2 led to the formation of 1,1',3,3'ferrocenetetracarboxylica cid in yields close to 80 %b esides smaller amounts of di-and tri-substituted ferrocenes (Scheme 15).
It is reasonable to assumet hat the observed yields of the CO 2 -trapped product represent the metalated species;t here are no hints towards substantial hydrolysis on one side or post-quench metalation processes on the other.I nt his respect, compound 4 demonstratedamore efficient tetra-metalationo f ferrocenei nc omparison to other examples using Lochmann-Schlosser superbases. Thiso utstanding reactivity is not restricted to neopentyl compounds. n-Butyllithium with an excesso f KOtAm used under the same conditions also achieves the tetra-functionalization of ferrocene with yields close to 60 %. [74] Conclusions from Metalation Reactions Using Potassium Mixed Aggregates Assumingasimilar basicityo fn-butyl and neopentyl, [75] the differencei nt he respective yields in these otherwise similar reactions arises from changes in solubility.W hen n-butyllithium is added to al argeq uantity/amount of KOtAm, [54] the absence of an immediate precipitation suggestst he formation of am ixed aggregate such as K 4 (nBu)(OtAm) 3 .T his finding is in contrast to ar eport by Lochmann,[61] where the reaction was achieved conversely.
In addition, the successful tetra-metalation of ferroceneb y compound 4 demonstrates that homo-metallic potassium aggregates are equal to other superbases in terms of basicity.I n this case, the reactivity can be connected directly to structural motifs F (Scheme 9) and IV (Scheme 13). There was no substantiald ifference in the obtained yield of substituted ferrocene, when 4 was produced in situ by mixing LiNp with KOtAm or by mixingK Np with three equivalents of KOtAm. In the former case, this indicates that lithium is trapped in inactive compounds such as Li 4 K 4 (OtAm) 8 (1 0 )o rL i 4 K(OtAm) 5 (3 0 ). [47] The reactiono fm etalated speciesp roduced by Lochmann-Schlosser superbases with electrophiles involves some obstacles (Scheme 16). The metalated speciesa re more or less insoluble, which does not significantly affect the reaction if the electrophile is reactive enougha nd/ors oluble itself. Of more importance is the fact that the alkali metal alkoxide, which is present in excessa nd in higher concentrations, also acts as a nucleophile. This is no problem when the electrophile, such as CO 2 or I 2 ,c an be used in excess, or if the metalate speciesc an be separated from excess alkoxide by filtration. Finally,t he metalated speciesi tself is ap otassium compound with as ubstantialb asicity.E lectrophilesw ith acidic hydrogen atoms (such as benzylic or allylic groups) are at risk to be metalated themselves before as uccessful functionalization of the metalated carbona tom can be achieved. Metal-metal exchange reactions [76] (such as potassium-zinc exchange), whichc ould lift the restriction to proton-free electrophiles, introducen ew synthetic problems.
To conclude, relevant pieces of information can be drawn from the first metalation reactions using compound 4.F irst, mixed alkali metal mixed aggregates do not lose their singular reactivity in the absence of lithium. Ap ossible mixed metal synergyist herefore not essential. Second, the differenceinsolubility in mixtures using neopentyl or n-butylg roups reflect itself in the yields of the final product, but not in overall outcome such as changingd egrees of polymetalation or regioselectivity.The main features of the neopentyl group in reference to the n-butyl group are the absence of missing metal-b-CH 2 interactions, which is also causing the inability of b-elimination, and the "mimicry" of tert-butoxy groups. Due to its greater stability and solubility,t his makes the neopentyl group af easible test case for the much more populara nd commercially available n-butyl compounds used in Lochmann-Schlosser superbases.

Summary and Outlook
For many years after their discovery, the composition of Lochmann-Schlosser superbases could not be determined conclusively based on mixtures of alkyllithium and potassium alkoxides. The use of the neopentyl group in such mixtures leads to products soluble in n-hexane or other alkanes. This way,i tw as possible to perform NMR studies in solution andt oo btain crystalsf or structurals tudies. The solid-state structures of these compounds simultaneously revealed genuinem ixed aggregatesc ontaining lithium, potassium, alkyl, and alkoxy groups.D epending on the amounto fi nitial materials,t he ratio of the components varied, the mixed aggregates showeda n excess of lithium and/or alkoxy groups. In solution, it is possible to identify alkoxy-rich Li 4 K 4 -hetero-octamer by NMR spectroscopy. Increasingt he alkyl content leads to the equilibrium of lithium-richer aggregates, which are undistinguishable by NMR spectroscopy.W hen alkyllithium is combinedw ith an excess of potassium alkoxide it is possible to isolate ap otassium alkyl/alkoxy aggregate. The basicityo ft his compound could be demonstrated by as ynthetically useful tetra-metalation of ferrocene. An umber of structuralm otifs were identified in structurally known mixed aggregates. These motifs can be derived from the involved startingm aterials and products or from connectivity of the alkyl groups to the alkali metals.S uperbasicity of mixed aggregates can be observed in the absence of lithium,but the presence of potassium or other heavier alkali metals is mandatory.A lkali metal alkoxide provide a solubilizing framework for otherwise insoluble and rather unstable alkylpotassium, which has ap ositive effect on the reactivity of such aggregates.O ne has to bear in mind, that alkyl alkali metal compounds represent Lewis acid/base complexes, where the Lewis acidic needs of the alkali metal atoms are hardly met by the Lewisb asicity of the alkyl groups. Addition of alkali metal alkoxide introduces new Lewis basic groups but also Lewis acidic metal atoms in the same time. This results in ap redominantly Lewis acidic behavior,w hich makes the mixed aggregates susceptible for all kinds of Lewis basic molecules, even more so in non-donating solvents such as n-hexane.
In consideration of the fact that all reactions with alkali metal superbases are carriedo ut in solution, it is important to gain more information about the behavior of these compoundsi nt he solution phase. By introduction of NMR-active isotopes, such as 6 Li, 2 H, 133 Cs, or 13 C, and the use of sophisticated DOSY NMR techniques, it will be possible to identify the present speciess pectroscopically andt os tudy the kinetica nd thermodynamic properties by NMR spectroscopy.A dditional solid-state structureso fr elevanta lkali metal compounds will fill important gaps. The existence of alkyl/alkoxy mixed aggregates other than compounds 1, 2, 3,a nd 4 is quite feasible. This would include similar sodium, rubidium, and cesium compounds. Still unknowns tructures of alkylpotassium or alkylsodium compounds,b ut also structures of mixed alkali metal alkyl compounds, such as Li x K y R z ,a re of great interest. Furthermore, understanding the reactivity of alkali metal superbases would benefits ubstantially if more could be learneda bout the nature of the metalated substrates. New synthetic strategies may lead to soluble products, allowing their characterization. In the same time, it could be possible to perform trans-metalation reactions, openingu pn ew synthetic routes with aw ide range of organic nucleophilese nabling cross-couplingr eactions.