Alkaline Earth Metal Template (Cross‐)Coupling Reactions with Hybrid Disila‐Crown Ether Analogues

Abstract Alkaline earth metal iodides were used as templates for the synthesis of novel silicon‐based ligands. Siloxane moieties were (cross‐)coupled and ion‐specific, silicon‐rich crown ether analogues were obtained. The reaction of 1,2,7,8‐tetrasila[12]crown‐4 (I) and 1,2‐disila[9]crown‐3 (II) with MgI2 yielded exclusively [Mg(1,2,7,8‐tetrasila[12]crown‐4)I2] (1). The larger Ca2+ ion was then employed for cross‐coupling of I and II and yielded the complex [Ca(1,2,7,8‐tetrasila[15]crown‐5)I2] (2). Cross‐coupling of I and 1,2,4,5‐tetrasila[9]crown‐3 (III) with SrI2 enables the synthesis of the silicon‐dominant 1,2,4,5,10,11‐hexasila[15]crown‐5 ether complex of SrI2 (3). Further, the compounds [Sr(1,2,10,11‐tetrasila[18]crown‐6)I2] (4), [Sr(1,2,13,14‐tetrasila[24]crown‐8)I2] (5), and [Sr(1,2,13,14‐tetrasila‐dibenzo[24]crown‐8)I2] (6) were obtained by coupling I, 1,2‐disila[12]crown‐4 (IV) or 1,2‐disila‐benzo[12]crown‐4 (V), respectively. Using various anions, the (cross‐)coupled ligands were also observed in an X‐ray structure within the mentioned complexes. These template‐assisted (cross‐)couplings of various ligands are the first of their kind and a novel method to obtain macrocycles and/or their metal complexes to be established. Further, the Si−O bond activations presented herein might be of importance for silane or even organic functionalization.


Introduction
It is well known for crowne thers to form coordination compounds with metal centers across aw ide range of the periodic table. These complexes are generally very stable and these polyethers as well as relateds ystemss uch as cryptands, or in general,m ultidentate ligands, gained many fields of applications since their discovery in the mid-sixties. The synthesis of such polyethers, however,isoften not trivial because the molecules of the starting materials need to be broughti ntoas uitable conformation for the formation of as pecific (most often cyclic) product. The synthesis of these cyclic ethers is therefore mostly metal-template assisted. [1] Te mplates are known to provide as uitable coordination sphere for the starting materials, whichf avors the linkaget o ring closure and allows obtaining cyclic ligand systemsi n much highery ields. [1] Somep rominente xamples are summarized in Scheme 1. Over the years, also silicon-based polyethers, their syntheses as well as their coordination behavior towards different Lewis acids were studied. [5] For al ong time, templateassisted syntheses of cyclosiloxanes were only observed (accidently)b yu sing highly reactive starting materials in the presence of silicon grease (Scheme 2). [6][7][8][9][10] Ring transformations in the presence of metal cationsw ere described later and the generation of sila-polyethers by metal-templated ring-opening polymerization was presented as an interesting possibility to obtain novel silicon-based ring systems. [5,11,12] The coordination behavior of silicon-basedm acrocycles, however,i ss ignificantly different to that of organic ethers, whichi sa ccompaniedb y the characteristics of the SiÀOb ond. Different explanations for the reduced capability of binding Lewis acids shown by siloxanes wereprovided in the literature.
Most recently, both concepts-covalency and ionicity-were considered as harmonious models to understand the basicity of siloxanes. Regardingacovalent model,negative hyperconjugation interactions are described in the case p(O)!s*(SiÀ C). [14,15] This strengthening of the SiÀOb ond then competes with the shift of electron density towards Lewis acids. Concerning the ionic model, the highly polarized SiÀOb ond features spatially diffuse electron pairs around the O-atom and repulsive interaction in between Si d+ + and metal n+ + disturbs silyl ether bonding. [16][17][18] Both, covalency and ionicity gain simultaneously in importance when the Si-O-Sia ngle increases and basicity lowers. [19][20][21] Given that the understanding of the SiÀOb ond was established in this case, research in the field of coordination chemistry with cyclosiloxanes andr elatedl igand systems showed clearlyt hat these effects have to be taken into account. Nonetheless, the coordination chemistry turned out to be very suitable for early Group 1a nd 2m etal ions, especially within disila-ligands, which providem ore suitable bite angles than cyclosiloxanes of D n type (D = Me 2 SiO, n = 6, 7). [22][23][24][25][26] The nature of ac ation, which shall be complexed within siloxane moieties, is preferably hard. [13,[27][28][29][30][31] In the case of the early alkaline earth-metal ions, even the commerciallya vailablec yclosiloxanes are able to dissolve salts characterized by ah igh lattice energy and stable complexes wereo btained. [32] Soft cations like Rb + + and Cs + + so far relucteds ilyl ether coordination. [33,34] Overall,s ila-ligands are suitable ligands fore arly s-block-metal coordination chemistry and opened an ew chapter in hostguest chemistry over the years. In this contribution, we report the synthesis of novel hybrid disila-crown ether moieties, which are accessible by alkaline earth metal-template synthesis. First outlookso fatargeted template synthesis of hybrid disila-ligands were given in past works. Scheme 2r epresents sila-ligands formed by metal templates.
The fivefold, coplanarc oordination by sila-crown ligandst owards the larger Ca 2+ + ion was described earlier and, as can be seen from variouse xamples, seems to match perfectly. [26,32] Hence, we repeatedt he reaction of I and II with CaI 2 .A fter workup procedures two new signals in the 29 Si{ 1 H} NMR are observed at 19.6 and 19.9 ppm. Crystallization of the reaction product 2 however,w as unsuccessful from differents olvents and temperaturesa lso due to poor solubility.W et herefore expanded the anion I À to I 3 À upon iodine addition. Thec ompound [Ca(1,2,7,8-tetrasila [15]crown-5)(I 3 ) 2 ]( 2a)t hen was crystallized and the molecular structure was determinedb ym eans of SC-XRD( Figure 2). Compound 2a contains the new ligand 1,2,7,8-tetrasila [15]crown-5, which was formed by ac ross-coupling reaction of the ligands I and II mediated by the Ca 2+ + ions. The cross-coupling of sila-ligands is an elegant way to obtain novel macrocycles. The synthesis of these macrocycles cannoty et be realized by conventional silane chemistry and thus, the alkaline earth metal template opens up new synthetic pathways of such new ligands. The O Si ÀCa 2+ + distances measure 239.3(7) and 245.2(9) pm and compare well to the complex [Ca(1,2-disila [18]crown-6)OTf 2 ]( OTf = CF 3 SO 3 À ). [26] The synthesis of [Sr(1,2,4,5,10,11-hexasila [15]crown-5)I 2 ]( 3)i s another example of ac ross-coupling reactioni nw hich an alkaline earth metal cation is used. 3 represents the first hybrid disila-crowne ther bearing more disilane than ethylene units. The even larger Sr 2+ + cation can be used as ac onvex template to cross-couple I and 1,2,4,5-tetrasila [9]crown-3 (III), which is accessible applying reaction conditions reported before. [27] The reactiony ields as ilicon-based [15]crown-5 ether merging half an equivalent of I and one equivalent of III.D ue to three disilane units within the ligand framework, the Sr 2+ + cation fits well Scheme3.Alkalinee arth metal-template cross-coupling reactions (left) and coupling reactions (right) of the silicon-based crown ethers 1-6.   19.44, 19.27, and 13.32 ppm, three resonances are observed in the 29 Si{ 1 H} NMR spectrum, all of which compare well to those of O Si ···Sr 2+ + coordination compounds characterized before. [26,27] Ac rystal structure of 3 could not be obtained,e ven after iodine addition. Hence, the Lewis acidic salt GaI 3 was added as an acceptorf or the iodide anion.I ti sn otable, that also in 3a,t he SiSiÀOÀSiSi fragment of the siloxane framework provides as mall Si-O-Si angle, which measures 121.1(6)8.T he O SiÀOÀSi ···Sr 2+ + distance might be the longestO ···Sr 2+ + distance observed in 3a,b ut the high basicity of siloxanes towards Group 2i ons is clearly emphasized.
After several months, af ew single crystals of [Sr(1,2,4,5,10,11-hexasila [15]crown-5)(GaI 4 ) 2 ]( 3a)w ere obtained, which were analyzed through SC-XRD ( Figure 3). The O Si ···Sr 2+ + distances (concerning SiÀOÀCa sw ell as SiÀOÀSi donor groups)c ompare well to various complexes of Sr 2+ + and sila ligands,w hich were reported before. [26,27] Using the Sr 2+ + cation as at emplate, we were able to observe reactions of other disila-crowne thers as well. The smallc avity of II does not allow for 1:1c omplexation of Sr 2+ + and thus, at otal of two equivalents of II react to form the 1,2,10,11-tetrasila [18]crown-6e ther.[ Sr(1,2,10,11-tetrasila [18]crown-6)I 2 ]( 4)i so btained as a colorless powder and in solution, ar esonance at 17.5 ppm in the 29 Si{ 1 H} NMR spectrum was observed. Given that we were also experiencing problems with the crystallization of this compound, the X-Ray structure could only be determined upon iodine addition. According to ar educed niggli formula, compound 1 1 [Sr(1,2,10,11-tetrasila [18]crown-6I 2/2 ]I 3 (4a)w as obtained as brownp latelets which were investigated through SC-XRD. As shown by the molecular structure in the crystal, the ligand does not perfectly match with the Sr 2+ + cation because a coplanar arrangemento ft he donor atoms is not observed (see Figure 4). However,a ll oxygen atoms of the crown ether still manage to participate in the coordination.U pon iodine addition, one I 3 À anion is formed,w hich does not coordinate the central ion in the solid state. This enables the formation of [Sr(1,2,10,11-tetrasila [18]crown-6)I] + + fragments to build infinite chains along [001] whichi sm ost likely the driving force for crystallization. The O Si ···Sr 2+ + distances in 4a compare well to those in compound 3a and [Sr(1,2-disila [18]crown6)I 2 ]. [24] At this point, it is clear that (cross-)coupling reactions are possible with different alkaline earth metal iodides. We tried to understand how the reaction works and the mechanism behindt he formation of theseu nprecedented macrocycles, but unfortunately,w ew ere unable to characterize any intermediate products by means of NMR spectroscopy or SC-XRD. We can assure that heating at reflux is neededt oc leave the smaller rings. Further, we can assure that the reactiond oes only occur if an iodide salt is employed. Other alkaline earth metal halidesd on ot form these ligands. This lets us conclude that one key step in the formation of the (cross-)coupling product has to be an ucleophilic substitution reaction. Furthermore,a se vident from different works, we could convincingly show that the SiÀOb ond is significantly weakened due to coordinating an alkaline earth metal ion. Thus, the SiÀOb ond instead of the CÀObond is cleaved, whichisa lso represented by the herein obtained macrocycles. Ap roposed mechanismf or the formation of the cross coupling products is depicted in Scheme4 for 4 as an example. As drawn here, the sila-ligand is polarizedb yt he Lewis acid first, which makesi te lectrophilic and thusa ccessible for nucleophilic attack of I À .A fterwards, the SiÀOb ond is cleaved. Subsequently,t he intermediate then has to cleave the second sila-crown. At this point of the reac-tion, the alkaline earth metal ion acts as ac onvex template, in which the open-chained ligand speciesi sb roughti nto as uitable conformation for ring closure, which is the final step of the presentedr eactions. We want to emphasize, that such am echanism wasalso postulated by Harder in apast work. [31] Further reactions of SrI 2 with the ligands1 ,2-disila [12]crown-4( IV)a nd 1,2-disila-benzo [12]crown-4 (V)w erep erformed. Both ligands are too small for Sr 2+ + and thus, template-assisted ring opening yieldedn ovel species by intermolecular coupling of the respective crown ether.T he reaction of IV with SrI 2 results in the formation of the first disilanyl-bearing [24]crown-8 ether.[ Sr(1,2,13,14-tetrasila [24]crown-8)I 2 ]( 5)w as obtained by coupling two equivalents of IV ( Figure 5). The molecular structure in the crystal reveals that six out of eight oxygen atoms, as well as the two iodide anions, are coordinating. The overall eightfold coordinated Sr 2+ + cation is therefore embedded in a pseudo-1,2-disila [18]crown-6 moiety.G iven that the two oxygen atoms of the second disilanef ragment do not coordinate, at wisting of the crown ether is observed. The driving force of this reaction is the arrangemento fs ix of the coordinating crowne ther oxygen atoms forming an [18]crown-6-like coordination sphere similar to that observed in 4.T he O Si ···Sr 2+ + distances in 5 are, however,s lightly shorter than those in [Sr (1,2-disila[18]crown6)I 2 ]. [26] As ingle 29 Si{ 1 H} NMR resonance signalf or 5 is observed at 15.2 ppm. As plit of this resonance at low temperature of 190 Kwas not observed.
Thus, the exchange between the coordinating disilane units is too fast on the NMR time scale and results in the described equivalency.A tt his point, it should be noted that the spectroscopic investigation of the compound is challenging. Compound 5 decomposesr eadily with only traces of moisture forming [Sr (11,12-disila-EO7)I]I (EO7 = heptaethylene glycol). [13] The twistingo ft he crown ether is most likely as uitable preorganization for the formation of the open-chained EO7 ligand   Figure 6). After a 2:1r eaction of V with SrI 2 , V performs intermolecularc oupling as well. Here, the dibenzo crown ether helicallye ncapsulates the central ion under replacement of one iodide anion.A ll eight oxygen atoms do now participate in the coordination of Sr 2+ + .E ven though the ligand in 6 is more rigid than that of 5, shielding of the centrali on is observed. This might be ar esult of slightly reduced basicity of benzo crown ethers in general. The replacement of iodide by oxygen donor groupsi sf avored over metal-anioni nteractions and is most likely also present in solution because the 29 Si{ 1 H} NMR chemical shift is with 19.8 ppm very distinct for Sr 2+ + coordination.
One of the remaining questions is of course whether it is possible to remove the metal center for possible use of such new ligand in coordination chemistry.W ea re currently investigating this subject,b ut are still at an early stage. To give an outlook, we performed exchange reactions. As an example, an excess of 6 was reacted with [222]cryptand and indeed we were able to obtain free ligand 1,2,13,14-tetrasila-dibenzo [24]crown-8 (7)a fter workup. The metal-free ligand speciesw as characterized by means of HR-ESI MS as well as NMR spectroscopy.I nH R-ESI + + MS, m/z relations of 625.2502[ 7+ +H] + + (100) as well as 647.2320 [7+ +Na] + + (100) were found. In addition, 29 Si{ 1 H} NMR spectroscopy revealed highfield chemical shift and as ingle resonancea t1 1.5 ppm is found. This value compares well to the free ligand V and also to related metal-free disila-ligand systems. [22,24,26,28,33] The problem so far is that a significant amount of the cryptand remains in the oily residue after conversion, so no furtherc oordinationc hemistry could be performedy et. To which extent this will be possible, not only for the ligands presented herein, is of currenti nterest and will be subjectoffurtherresearch.
So far,w ea lso tried (cross)-coupling reactions with BaI 2, or other alkaline earth metal halidess uch as chlorides and bromides but variousa ttempts failed. The methodo btaining the novel macrocycles herein is for this reason restricted to iodide salts.

Conclusions
In this study alkaline earth metal iodides were successfully used for the coupling reaction of different silicon-based crown ether analogues to form novel ligand environments. The work presentsS i ÀOb ond-cleavage reactions driven by am acrocyclic effect due to metal-template reaction of the respective siliconbased ligands. Given that the Mg 2+ + ion is too small to crosscouple small silicon-basedl igands, ac ross-coupling reaction was established with the larger Ca 2+ + and Sr 2+ + ions. Furthermore, the Sr 2+ + ion was used for coupling reactions of three differentd isila-crown ethers. Barium as at emplate for coupling reaction, however,t urned out to be unsuccessful. Although the organic [ 12]crown-4 is too smallf or the Mg 2+ + ion, the cavity of the disilane-bearing analogue 1,2,7,8-tetrasila [12]   tion of the small rings I and II for 2,a nd I and III for 3.T he ligand in 3 represents the first crowne ther with more disilane than ethylene units between the donor atoms. Finally, [Sr(1,2,10,crown-6)I 2 ]( 4), [Sr(1,2,13,14-tetrasila [24]crown-8)I 2 ]( 5)a nd [Sr(1,2,13,14-tetrasila-dibenzo [24]crown-8)I]I (6)w ere characterizeda nd obtained by templatedriven dimerization of II, IV,o rV with SrI 2 .T he ligand moieties observed in 2-6 cannot yet be synthesized by conventional silane chemistry andf or this reason, the template-assisted (cross-)coupling of variousl igandsi sa ne legant wayt oo btain novel macrocyclesa nd/or their metal complexes.O verall, this work gives an outlooko nG roup 2i on-catalyzed silane syntheses, especially due to the factt hat first attempts showedt hat the metal ion can, in principle, be removed from an ovel silaether.W ea lso hope that there will be ab roader application of the presented SiÀOb ond activations as am ean of molecule functionalization.

Experimental Section
General All manipulations were carried out with rigorous exclusion of oxygen and moisture using basic Schlenk techniques establishing an inert-gas atmosphere with av acuum line. All solvents were dried and freshly distilled before use. The alkaline earth metal salts MgI 2 (Alfa Aesar, 9 9,996 %), CaI 2 and SrI 2 (Alfa Aesar,9 9%), BaI 2 (abcr, 9 9.995 %) and GaI 3 (abcr,9 9%)w ere finely ground and stored in aMbraun glovebox under Ar.N MR spectra were recorded on aB ruker AV III HD 300 MHz or AV III 500 MHz spectrometer,r espectively.I nfrared (IR) spectra of the respective samples were measured using attenuated total reflectance (ATR) mode on the Bruker-type spectrometer Alpha FT-IR. ESI mass spectra were acquired with aL TQ-FT Ultra mass spectrometer (Thermo Fischer Scientific) and LIFDI mass spectra were acquired on aJ EOL AccuTOF-GCv device. The resolution was set to 100.000. Elemental analysis was carried out on aV ario MicroCube. In case of complex 2 and 3, we were not able to obtain an accurate elemental analysis. This is probably due to the wax-like, greasy characteristics of the compounds and/or formation of SiC during the measurement. The crown-ethers I, II, IV,a nd V were synthesized according to literature-known procedures. [22,28,33] Compound 7 can only be obtained with major impurities of [222]crypt.

crown-5)I 2 ](3)
Compounds I (0.100 g, 0.29 mmol) and III (0.170 g, 0.55 mmol) were dissolved in a,a,a-trifluorotoluene (10 mL). Subsequently,S rI 2 (0.188 g, 0.56 mmol) was added. Heating the mixture at reflux for 90 min resulted in aw hite suspension. Removing the solvent under reduced pressure yielded aw hite precipitate which was extracted with DCM (10 mL), followed by filtration. After washing with two portions of n-pentane (4 mL each) and drying in vacuo, 3 was obtained as ac olorless, greasy solid (0.185 g, 40 %). For singlecrystal growth, product 3 (0.030 g, 0.04 mmol) and GaI 3 (0.018 g, 0.04 mmol) were dissolved in a,a,a-trifluorotoluene (5 mL). The suspension was stirred 30 min and gently warmed to 60 8Cf or 5min. The mixture was then freed of the solvent, extracted with DCM (3 mL) and filtered. The filtrate was then concentrated until Compound II (0.258 g, 1.17 mmol) was dissolved in a,a,a-trifluorotoluene (10 mL). Subsequently SrI 2 (0.200 g, 0.59 mmol) was added. Heating the mixture at reflux for 60 min. resulted in aw hite suspension. Removing the solvent under reduced pressure gave a white precipitate which was extracted with DCM (20 mL) followed by filtration. Upon washing with n-pentane (5 mL) and removal of the solvent, 4 was obtained as ap ale white powder (0.259 g, 56 %). For single crystal growth, product 4 (30 mg 0.04 mmol) and I 2 (10 mg, 0.04 mmol) were dissolved in DCM (4 mL). Layering the solution with n-pentane (20 mL) yielded single crystals of 4a as brown blocks after three days.  Compound IV (0.209 g, 0.79 mmol) was dissolved in a,a,a-trifluorotoluene (15 mL). Subsequently SrI 2 (0.135 g, 0.40 mmol) was added. Heating the mixture at reflux for three hours resulted in aw hite suspension. Removing the solvent under reduced pressure yielded aw hite precipitate which was extracted with DCM (20 mL), followed by filtration. The product 5 was then obtained as ac olorless powder after removal of the solvent (0.252 g, 74 %). For single-crystal growth, the powder was dissolved in DCM (4 mL) and layered with n-pentane (20 mL). Single crystals were obtained overnight as colorless needles.  Compound V (0.200 g, 0.64 mmol) was dissolved in a,a,a-trifluorotoluene (15 mL). Subsequently,S rI 2 (0.109 g, 0.32 mmol) was added. Heating the mixture at reflux for three hours resulted in a white suspension. Removing the solvent under reduced pressure yielded aw hite precipitate which was extracted with DCM (20 mL) followed by filtration. The product 6 was then obtained as ac olorless powder after removal of the solvent (0.200 g, 65 %). For singlecrystal growth, the powder was dissolved in DCM (3 mL) and layered with n-pentane (20 mL). Single crystals were obtained overnight as colorless platelets. Synthesis of 1,2,13,14-tetrasila-dibenzo [24]crown-8 (7) Compound 6 (0.190 g, 0.19 mmol, excess) was dissolved in DCM (10 mL) and [222]cryptand (0.050 mg, 0.13 mmol, 0.7 equiv) was added. The resulting suspension was stirred overnight to give a clear solution. The solvent was removed under reduced pressure to give an oily,g reasy residue. Extracting with n-pentane (10 mL) and subsequent filtering of the suspension yielded ac lear solution. Removing the solvent under reduced pressure gave ag reasy crown ether-[222]cryptand mixture (0.13 g).