New Cu 4 Na 4 - and Cu 5 -Based Phenylsilsesquioxanes. Synthesis via Complexation with 1,10-Phenanthroline, Structures and High Catalytic Activity in Alkane Oxidations with Peroxides in Acetonitrile

: Self-assembly of copper(II)phenylsilsesquioxane assisted by the use of 1,10-phenanthroline (phen) results in isolation of two unusual cage-like compounds: (PhSiO 1,5 ) 12 (CuO) 4 (NaO 0.5 ) 4 (phen) 4 1 and (PhSiO 1,5 ) 6 (PhSiO 1,5 ) 7 (HO 0.5 ) 2 (CuO) 5 (O 0.25 ) 2 (phen) 3 2. X-Ray di ﬀ raction study revealed extraordinaire molecular architectures of both products. Namely, complex 1 includes single cyclic (PhSiO 1,5 ) 12 silsesquioxane ligand. Four sodium ions of 1 are additionally ligated by 1,10-phenan throlines. In turn, “sodium-less” complex 2 represents coordination of 1,10-phenanthrolines to copper ions. Two silsesquioxane ligands of 2 are: (i) noncondensed cubane of a rare Si 6 -type and (ii) unprecedented Si 7 -based ligand including two HOSiO 1.5 fragments. These silanol units were formed due to removal of phenyl groups from silicon atoms, observed in mild conditions. The presence of phenanthroline ligands in products 1 and 2 favored the π – π stacking interactions between neighboring cages. Noticeable that in the case of 1 all four phenanthrolines participated in such supramolecular organization, unlike to complex 2 where one of the three phenanthrolines is not “supramolecularly active”. Complexes 1 and 2 were found to be very e ﬃ cient precatalysts in oxidations with hydroperoxides. A new method for the determination of the participation of hydroxyl radicals has been developed. H O 2 oxidations of


Oxidation of Alkanes and Alcohols
The oxidation reactions were typically carried out in air in thermostated Pyrex cylindrical vessels with vigorous stirring; total volume of the reaction solution was 2.5 mL (Caution: The combination of air or molecular oxygen and H 2 O 2 with organic compounds at elevated temperatures may be explosive). Initially, a portion of 50% aqueous solution of hydrogen peroxide or 70% aqueous tert-butyl hydroperoxide was added to the solution of the catalyst and substrate in acetonitrile. The reaction solutions were analyzed by GC (Gas Chromatograph) (the instrument 3700, fused silica capillary column FFAP/OV-101 20/80 w/w, 30 m × 0.2 mm × 0.3 µm; helium as a carrier gas. Attribution of peaks Catalysts 2019, 9, 701 3 of 20 was made by comparison with chromatograms of authentic samples). Usually samples were analyzed twice, i.e., before and after the addition of the excess of solid PPh 3 . This method was developed and used previously by Shul'pin [45][46][47][48][49][50][51]. Alkyl hydroperoxides are transformed in the GC injector into a mixture of the corresponding ketone and alcohol. Due to this we quantitatively reduced the reaction samples with PPh 3 to obtain the corresponding alcohol. This method allows us to calculate the real concentrations not only of the hydroperoxide but also of the alcohols and ketones present in the solution at a given moment.
Catalysts 2019, 9, x FOR PEER REVIEW 3 of 20 some of us very recently [39], represents another type of Si12-siloxanolate ligand, namely Ph12Si12O12(OH)(O − )11. It points out the significant flexibility of CLMSs' self-assembly, taking in mind that both 1 and "Bicycle Helmet" [39] were isolated as THF-solvated complexes with N, N-ligands. Nevertheless, two of these products differed not only in the composition of central siloxanolate ligand but also in other features of composition. Compare, for example, (i) types of nuclearity (Cu4Na4 for 1 vs. Cu5Na1 8b ), and (ii) N, N-ligand/cage' ratio (4/1 for 1 vs. 3/1 [39]). Mentioned above non-rigidity of siloxanolate ligands in the procedure of CLMS' assembly found additional evidence during the study of the synthesis of Cu-silsesquioxane in dimethylformamide (DMF) media (Scheme 1, complex 2). This reaction allows unprecedented pentanuclear CLMS 2 (PhSiO 1,5 ) 6 (PhSiO 1,5 ) 7 (HO 0.5 ) 2 (CuO) 5 (O 0.25 ) 2 (phen) 3 (Figure 2, top), which is very unlike pentacopper (PhSiO 1.5 ) 10 (CuO) 5 -CLMSs, reported earlier by us [54] and the Y. Kononevich team [52]. Most likely, this difference was due to the participation of phenanthroline ligands in the reaction reporting here. Note, three phenanthroline ligands coordinated three copper ions of 2, while the remaining two copper centers were O-ligated only. This ligation was realized via participation of two types of polycyclic siloxanolate ligands. The first type represented incompletely condensed cubane silsesquioxane of Si 6 composition ( Figure 2, center). Unlike widespread CLMSs, obtained using Si 7 -triols [2,4,6,[18][19][20], cage metallasilsesquioxanes including Si 6 -tetraol as a structural unit [60,61] are the rarest representatives of the CLMSs family. Noteworthy is that Si 6 -fragment in 2 was formed in situ during the self-assembly of complex 2, while Refs. [60,61] describe the use of corresponding Si 6 tetraol as a synthon. The second type of silsesquioxane skeleton in 2, to our opinion, was absolutely a unique one. First of all, this Si 7 -unit ( Figure 2, bottom) did not contain condensed cycles, characteristic for Si 7 -triols, mentioned above [2,4,6,[18][19][20]. Instead of that, Si 7 ligand in 2 involved the Si 4 -based cycle, connected to the "Si 3 -based tail". To a certain extent, this structural composition resembled the composition of the Ursa Major constellation. Secondly, this ligand contained two non-silsesquioxane fragments of the HOSiO 1.5 composition. Taking in mind that only phenyltriethoxysilane PhSi(OEt) 3 was used in the synthesis, we should witness the mild conditioned removal of organic groups from silicon (followed by the formation of silanol fragments). To the best of our knowledge, this process is a truly unprecedented one. Compare it, for example, to earlier reported process of calcination of metallasilsesquioxanes, both oligomeric [21] and cage-like ones [62][63][64][65], where cleavage of Si-R fragments required a temperature no less than 130 • C. It seems logical that mild temperature of synthesis of 2 allowed us to retain most of the phenyl groups in the silicon atoms. Mentioned above non-rigidity of siloxanolate ligands in the procedure of CLMS' assembly found additional evidence during the study of the synthesis of Cu-silsesquioxane in dimethylformamide (DMF) media (Scheme 1, complex 2). This reaction allows unprecedented pentanuclear CLMS 2 (PhSiO1,5)6(PhSiO1,5)7(HO0.5)2(CuO)5(O0.25)2(phen)3 ( Figure 2, top), which is very unlike pentacopper (PhSiO1.5)10(CuO)5-CLMSs, reported earlier by us [54] and the Y. Kononevich team [52]. Most likely, this difference was due to the participation of phenanthroline ligands in the reaction reporting here. Note, three phenanthroline ligands coordinated three copper ions of 2, while the remaining two copper centers were O-ligated only. This ligation was realized via participation of two types of polycyclic siloxanolate ligands. The first type represented incompletely condensed cubane silsesquioxane of Si6 composition (Figure 2, center). Unlike widespread CLMSs, obtained using Si7triols [2,4,6,[18][19][20], cage metallasilsesquioxanes including Si6-tetraol as a structural unit [60,61] are the rarest representatives of the CLMSs family. Noteworthy is that Si6-fragment in 2 was formed in situ during the self-assembly of complex 2, while Refs. [60,61] describe the use of corresponding Si6 tetraol as a synthon. The second type of silsesquioxane skeleton in 2, to our opinion, was absolutely a unique one. First of all, this Si7-unit (Figure 2, bottom) did not contain condensed cycles, characteristic for Si7-triols, mentioned above [2,4,6,[18][19][20]. Instead of that, Si7 ligand in 2 involved the Si4-based cycle, connected to the "Si3-based tail". To a certain extent, this structural composition resembled the composition of the Ursa Major constellation. Secondly, this ligand contained two non-silsesquioxane fragments of the HOSiO1.5 composition. Taking in mind that only phenyltriethoxysilane PhSi(OEt)3 was used in the synthesis, we should witness the mild conditioned removal of organic groups from silicon (followed by the formation of silanol fragments). To the best of our knowledge, this process is a truly unprecedented one. Compare it, for example, to earlier reported process of calcination of metallasilsesquioxanes, both oligomeric [21] and cage-like ones [62][63][64][65], where cleavage of Si-R fragments required a temperature no less than 130 °C. It seems logical that mild temperature of synthesis of 2 allowed us to retain most of the phenyl groups in the silicon atoms. Taking into account the novelty of siloxanolate ligands obtained, chemical connectivity of copper atoms within metal-containing cluster was of interest. It was analyzed following the procedure described by Kostakis and Powell [66] to obtain the graph of the Cu-containing skeleton keeping the cluster connectivity as implemented into the ToposPro package [67]. The procedure includes analysis of metal atoms connected via bridge atoms only (here, oxygen atoms of siloxanolate anions, Figure 3). We found that metal atoms in the cages of 1 and 2 belonged to discrete clusters with, respectively, 1,2M4-1, and 1,2,3M5-1 topology ( Figure 3). Notation of clusters is given as NDkm, where N is equal to the coordination number of topologically non-equivalent nodes, M denotes a discrete cluster (in contrast with 1-, 2-and 3-periodic architectures), k is the number of metal atoms in the cluster and m-a classification number to distinguish topologically distinct clusters with equal NDk parameters. The majority of globular bi-and trimetallic silsesquioxanes [13,59] clusters realize the 1,2M4-1 topology of the metal core, while the 1,2,3M5-1 topology was not met in metallasilsesquioxanes before. Moreover, previously published databases of Co [68], Mn [69] and Nicontaining [70] high-nuclearity cluster also lack such architectures. Taking into account the novelty of siloxanolate ligands obtained, chemical connectivity of copper atoms within metal-containing cluster was of interest. It was analyzed following the procedure described by Kostakis and Powell [66] to obtain the graph of the Cu-containing skeleton keeping the cluster connectivity as implemented into the ToposPro package [67]. The procedure includes analysis of metal atoms connected via bridge atoms only (here, oxygen atoms of siloxanolate anions, Figure 3). We found that metal atoms in the cages of 1 and 2 belonged to discrete clusters with, respectively, 1,2M4-1, and 1,2,3M5-1 topology ( Figure 3). Notation of clusters is given as NDk-m, where N is equal to the coordination number of topologically non-equivalent nodes, M denotes a discrete cluster (in contrast with 1-, 2-and 3-periodic architectures), k is the number of metal atoms in the cluster and m-a classification number to distinguish topologically distinct clusters with equal NDk parameters. The majority of globular bi-and trimetallic silsesquioxanes [13,59] clusters realize the 1,2M4-1 topology of the metal core, while the 1,2,3M5-1 topology was not met in metallasilsesquioxanes before. Moreover, previously published databases of Co [68], Mn [69] and Ni-containing [70] high-nuclearity cluster also lack such architectures. Both complex 1 and 2 represent specific supramolecular organization due to the π-π stacking interactions of phenanthroline ligands. In the case of 1, contacts involved all four phenanthrolines leading to the formation of an extended supramolecular structure (Figure 4, top). To the best of our knowledge, analogues cases (heterometallic complexes, exhibiting packing through Na-phen units) are not widespread [71][72][73][74][75] and reported only for praseodymium/sodium [71], copper/sodium [72,73], iron/sodium [74], bismuth/sodium [75] and cerium/sodium [76] systems but never for CLMSs. Interestingly one of the phenanthroline ligands in the composition of complex 2 were involved in ππ stacking interactions with phenyl groups of the same cage while two others provide formation of an extended structure due to contacts with phenanthrolines of neighboring cages (Figure 4, bottom). Both complex 1 and 2 represent specific supramolecular organization due to the π-π stacking interactions of phenanthroline ligands. In the case of 1, contacts involved all four phenanthrolines leading to the formation of an extended supramolecular structure (Figure 4, top). To the best of our knowledge, analogues cases (heterometallic complexes, exhibiting packing through Na-phen units) are not widespread [71][72][73][74][75] and reported only for praseodymium/sodium [71], copper/sodium [72,73], iron/sodium [74], bismuth/sodium [75] and cerium/sodium [76] systems but never for CLMSs. Interestingly one of the phenanthroline ligands in the composition of complex 2 were involved in π-π stacking interactions with phenyl groups of the same cage while two others provide formation of an extended structure due to contacts with phenanthrolines of neighboring cages (Figure 4, bottom).  Both complex 1 and 2 represent specific supramolecular organization due to the π-π stacking interactions of phenanthroline ligands. In the case of 1, contacts involved all four phenanthrolines leading to the formation of an extended supramolecular structure (Figure 4, top). To the best of our knowledge, analogues cases (heterometallic complexes, exhibiting packing through Na-phen units) are not widespread [71][72][73][74][75] and reported only for praseodymium/sodium [71], copper/sodium [72,73], iron/sodium [74], bismuth/sodium [75] and cerium/sodium [76] systems but never for CLMSs. Interestingly one of the phenanthroline ligands in the composition of complex 2 were involved in ππ stacking interactions with phenyl groups of the same cage while two others provide formation of an extended structure due to contacts with phenanthrolines of neighboring cages (Figure 4, bottom).

Catalytic Oxidations of Alkanes and Alcohols
Transition metal complexes bearing voluminous organic ligands are widely used as catalysts in various organic reactions including CH-functionalization (see reviews and original works [77][78][79][80][81][82][83][84][85]): "Over the past 50 years organic chemists have invented an incredible number of synthetically useful transition metal catalyzed reaction types. In the most cases, the transition metals are complexed to sterically and electronically tunable organic ligands, which govern activity and selectivity" [80]. "Capitalizing on a well-defined second coordination sphere provided by the protein around the metal cofactor, researchers have modified (artificial) metalloenzymes to react with exquisite levels of selectivity including enantio-, diastereo-, chemo-, and regioselectivity. Such catalyst control has been made possible by combining the tools of chemistry and biology" [81].
Di Stefano and co-workers [86] published a list of the following tools to distinguish between metal-based and free-radical oxidation mechanisms: (1) Alcohol/ketone ratio (A/K); (2) kinetic isotope effects (KIE); (3) reaction under argon/air; (4) the Shul'pin test for alkyl hydroperoxides; (5) regioselectivity; (6) epimerization; (7) chirality transfer; (8) labeling studies; (9) use of radical traps and (10) cyclohexene oxidation. In the present work, we applied the Shul'pin test (adding PPh3), regioselectivity in oxidation of linear and branched alkanes as well as epimerization of 1,2dimethylcyclohexane. Previously, we studied the catalytic systems based on the data on the selectivity of alkane oxidation and we concluded that the oxidizing species are hydroxyl radicals. The synthesized complexes have been also used in oxidation of some alcohols with tert-butyl hydroperoxide in acetonitrile.
We found that compounds 1 and 2 catalyze the oxidation of alkanes with H2O2 in acetonitrile. We investigated oxidations of alkanes and alcohols with hydroperoxides catalyzed by specially synthesized in this work (see above) compounds 1 and 2. The results of our investigations of some oxidation reactions catalyzed by complexes 1 and 2 under various conditions (temperatures 20-50 °C) are summarized in Figures 5-9. In some cases HNO3 was added to the reaction solution. It could be clearly seen that after addition of PPh3 to the reaction sample concentrations of cyclohexanone decreased while concentration of cyclohexanol after the same time period noticeably rose. This may

Catalytic Oxidations of Alkanes and Alcohols
Transition metal complexes bearing voluminous organic ligands are widely used as catalysts in various organic reactions including CH-functionalization (see reviews and original works [77][78][79][80][81][82][83][84][85]): "Over the past 50 years organic chemists have invented an incredible number of synthetically useful transition metal catalyzed reaction types. In the most cases, the transition metals are complexed to sterically and electronically tunable organic ligands, which govern activity and selectivity" [80]. "Capitalizing on a well-defined second coordination sphere provided by the protein around the metal cofactor, researchers have modified (artificial) metalloenzymes to react with exquisite levels of selectivity including enantio-, diastereo-, chemo-, and regioselectivity. Such catalyst control has been made possible by combining the tools of chemistry and biology" [81].
Di Stefano and co-workers [86] published a list of the following tools to distinguish between metal-based and free-radical oxidation mechanisms: (1) Alcohol/ketone ratio (A/K); (2) kinetic isotope effects (KIE); (3) reaction under argon/air; (4) the Shul'pin test for alkyl hydroperoxides; (5) regioselectivity; (6) epimerization; (7) chirality transfer; (8) labeling studies; (9) use of radical traps and (10) cyclohexene oxidation. In the present work, we applied the Shul'pin test (adding PPh 3 ), regioselectivity in oxidation of linear and branched alkanes as well as epimerization of 1,2-dimethylcyclohexane. Previously, we studied the catalytic systems based on the data on the selectivity of alkane oxidation and we concluded that the oxidizing species are hydroxyl radicals. The synthesized complexes have been also used in oxidation of some alcohols with tert-butyl hydroperoxide in acetonitrile.
We found that compounds 1 and 2 catalyze the oxidation of alkanes with H 2 O 2 in acetonitrile. We investigated oxidations of alkanes and alcohols with hydroperoxides catalyzed by specially synthesized in this work (see above) compounds 1 and 2. The results of our investigations of some oxidation reactions catalyzed by complexes 1 and 2 under various conditions (temperatures 20-50 • C) are summarized in Figures 5-9. In some cases HNO 3 was added to the reaction solution. It could be clearly seen that after addition of PPh 3 to the reaction sample concentrations of cyclohexanone decreased while concentration of cyclohexanol after the same time period noticeably rose. This may be explained      In this work, we used a kinetic method developed by some of us [82]. To determine the kinetic characteristics of the oxidizing species generated by two catalytic systems studied in this work, the dependence of the cyclohexane oxidation rate on its initial concentration was studied (Figures 8 and  9). The nature of the dependence corresponds to the assumption of the competitive interaction of the oxidizing species X with acetonitrile (stage 1) and cyclohexane RH (stage 2): H2O2 + catalyst → X rate of this stage is Wi, Here, Equation (i) is the effective reaction of generating oxidizing species X at a rate of Wi; Equations (1) and (2) are the transformations of CH3CN and RH with the formation, in particular, various products from acetonitrile and of cyclohexyl hydroperoxide ROOH, induced by interactions that limit their rate with X. The rate constants of these interactions are k1 and k2, respectively. Assuming that the concentration of species X is quasi-stationary, we obtained the following expression for the initial rate of formation of ROOH: In this work, we used a kinetic method developed by some of us [82]. To determine the kinetic characteristics of the oxidizing species generated by two catalytic systems studied in this work, the dependence of the cyclohexane oxidation rate on its initial concentration was studied (Figures 8 and 9). The nature of the dependence corresponds to the assumption of the competitive interaction of the oxidizing species X with acetonitrile (stage 1) and cyclohexane RH (stage 2): Here, Equation (i) is the effective reaction of generating oxidizing species X at a rate of W i ; Equations (1) and (2) are the transformations of CH 3 CN and RH with the formation, in particular, various products from acetonitrile and of cyclohexyl hydroperoxide ROOH, induced by interactions that limit their rate with X. The rate constants of these interactions are k 1 and k 2 , respectively. Assuming that the concentration of species X is quasi-stationary, we obtained the following expression for the initial rate of formation of ROOH: Let us present Equation (3) in the form of Equation (4), convenient for the analysis of experimental data: It follows from Figures 8B and 9B, the experimental data shown in Figures 8A and 9A, satisfy Equation (4), and this indicates the validity of the model of the competitive interaction of species X with CH 3 CN and RH. Analysis of the dependence W on [CyH] for both catalysts in the frames of the model given by Equations (i)-(2) for the concurrent interaction of oxidizing species with acetonitrile and cyclohexane allowed us using Equation (4) to determine values k 1 [CH 3 CN]/k 2 : 0.44 M for 1 and 0.66 M for 2. These parameters were sufficiently higher than characteristics known for the reactions of cyclohexane with hydroxyl radicals: Interval 0.04-0.1 M. It is important to emphasize here that the regioselectivity parameters collected in Table 1 indicated that oxidizing species in the cases of both catalysts 1 and 2. The contradiction between conclusions made on the basis of the kinetic study and investigation using selectivity parameters can be understood if we assumed that the approach of reaction centre to acetonitrile and cyclohexane molecules was different due to their different special characteristics. Concentrations of CH 3 CN and CyH close to the reaction center and in solution balk volume were different. Due to this the constant ratios k 1 /k 2 in the reactions catalyzed by 1 or 2 differ from that found for the interactions with the participation of free hydroxyl radicals.   4) is the relative normalized (calculated taking into account the number of hydrogen atoms at each carbon) reactivities of hydrogen atoms at carbons 1, 2, 3 and of the chain of n-heptane. Parameter 1 • :2 • :3 • is the relative normalized reactivities of hydrogen atoms at the primary, secondary and tertiary carbons of methylcyclohexane and isooctane. Parameter trans/cis is the ratio of isomers of tert-alcohols with mutual transand cis-orientation of the methyl groups formed in the oxidation of cisand trans-1,2-dimethylcyclohexane. All parameters were measured after the reduction of the reaction mixtures with triphenylphosphine before GC analysis and calculated based on the ratios of isomeric alcohols. b PCA is pyrazine-2-carboxylic acid. For this system, see Ref. [83]. c L is trimethyltriazacyclononane. For this system, see Ref. [84].
As shown above, the low selectivity of the oxidative effect of X species, which is close to the selectivity found for reactions involving hydroxyl radicals, and this indicates that hydroxyl radicals are generated in the catalytic systems studied in the present work. Similar results (a decrease in the relative reactivity of the oxidizing species for cyclohexane in comparison with acetonitrile) were obtained earlier by us in clarifying the nature of the oxidizing species during the catalysis of the decomposition of hydrogen peroxide on the complex palanquin-like Cu 4 Na 4 silsesquioxane [41] as well as using aluminum nitrate [85].
We found that the additives of HNO 3 reduced the catalytic activity of both the 1 complex and the 2 complex (see Figures 8A and 9B). This difference could be explained either by a change in the structure of the initial complex 1 or 2 in a solution containing HNO 3 , or by a change in the rate of interaction of complexes 1 and 2 with H 2 O 2 . The acid probably broke up the initial complex into many smaller particles, which were less active in catalysis. The data on the oxidation of linear n-heptane (see Table 1) showed that the oxidative effect of the species generated by the catalytic decomposition of H 2 O 2 in the presence of compounds 1 or 2, both in the absence of HNO 3 and in its presence, was close to the regioselectivity characteristic of the reaction with hydroxyl radical. On the other hand, analysis of the dependence of the initial oxidation rate of cyclohexane on its initial concentration during catalysis of the H 2 O 2 decomposition on complex 1 in the presence of HNO 3 in the frames of a homogeneous model (i)-(2) of the competitive interaction of the oxidizing species with cyclohexane and acetonitrile (similar to the one given above for the reaction in the absence of HNO 3 ) showed that the reactivity of the oxidizing particle corresponded to the participation of the hydroxyl radical: The value of k 1 [CH 3 CN]/k 2 = 0.11 M was close to the values typical for the hydroxyl radical parameters 0.04-0.1 M. While for complex 1 in the absence of HNO 3 , the value of k 1 [CH 3 CN]/k 2 = 0.44 M differed significantly from that which were typical for reactions involving hydroxyl radicals.
The combination of our results shows that in the presence of nitric acid the structure of the initial complex was likely changed, namely the complex structures of multicore complexes were transformed into simpler compounds, the reaction centers of which were equally accessible to both acetonitrile and cyclohexane. This led to a difference in the reactivity of complexes 1 and 2 in the presence of HNO 3 from that in its absence (when the reaction centers were more accessible for interaction with acetonitrile than with cyclohexane).
The complexes 1 and 2 were also used in the oxidation of some alcohols with tert-butyl hydroperoxide in acetonitrile. 1-Phenylethanol could be transformed into acetophenone in almost the quantitative yield (Figures 10 and 11). Cyclohexanol and heptanol-2 were oxidized to the corresponding ketones less efficiently. of H2O2 in the presence of compounds 1 or 2, both in the absence of HNO3 and in its presence, was close to the regioselectivity characteristic of the reaction with hydroxyl radical. On the other hand, analysis of the dependence of the initial oxidation rate of cyclohexane on its initial concentration during catalysis of the H2O2 decomposition on complex 1 in the presence of HNO3 in the frames of a homogeneous model (i)-(2) of the competitive interaction of the oxidizing species with cyclohexane and acetonitrile (similar to the one given above for the reaction in the absence of HNO3) showed that the reactivity of the oxidizing particle corresponded to the participation of the hydroxyl radical: The combination of our results shows that in the presence of nitric acid the structure of the initial complex was likely changed, namely the complex structures of multicore complexes were transformed into simpler compounds, the reaction centers of which were equally accessible to both acetonitrile and cyclohexane. This led to a difference in the reactivity of complexes 1 and 2 in the presence of HNO3 from that in its absence (when the reaction centers were more accessible for interaction with acetonitrile than with cyclohexane).
The complexes 1 and 2 were also used in the oxidation of some alcohols with tert-butyl hydroperoxide in acetonitrile. 1-Phenylethanol could be transformed into acetophenone in almost the quantitative yield (Figures 10 and 11). Cyclohexanol and heptanol-2 were oxidized to the corresponding ketones less efficiently.

Conclusions
We presented here non-trivial Cu4Na4 1 and Cu5 2 phenylsilsesquioxanesas complexes with 1,10phenanthroline ligands. The cage structure of complex 1 was characterized by the presence of the Si12-based ligand, favoring the contacts of external sodium ions to phenanthroline molecules. Complex 2 involved two different types of silsesquioxane fragments, namely the Si6-non-condensed cubane and Si7-"Ursa Major" ligand. The latter represents a very unusual removal of phenyl groups from silicon atoms with simultaneous formation of Si-OH groups. It points to a complicated way of self-assembly of metallasilsesquioxane cages especially in the presence of additional organic ligands. Stacking interactions of phenanthrolines formed extended supramolecular assembly of both compounds 1 and 2 in crystal packing. Catalytic tests of 1 and 2 showed their high activity in oxidations of alkanes and alcohols with peroxides. Further investigation of the "sesquioxane Oligand plus N,N-ligand" approach to isolation of catalytically prospective Cu(II)-based structures is currently on way in our team. A new method for the determination of the participation of hydroxyl radicals was developed.

Conclusions
We presented here non-trivial Cu 4 Na 4 1 and Cu 5 2 phenylsilsesquioxanesas complexes with 1,10-phenanthroline ligands. The cage structure of complex 1 was characterized by the presence of the Si 12 -based ligand, favoring the contacts of external sodium ions to phenanthroline molecules. Complex 2 involved two different types of silsesquioxane fragments, namely the Si 6 -non-condensed cubane and Si 7 -"Ursa Major" ligand. The latter represents a very unusual removal of phenyl groups from silicon atoms with simultaneous formation of Si-OH groups. It points to a complicated way of self-assembly of metallasilsesquioxane cages especially in the presence of additional organic ligands. Stacking interactions of phenanthrolines formed extended supramolecular assembly of both compounds 1 and 2 in crystal packing. Catalytic tests of 1 and 2 showed their high activity in oxidations of alkanes and alcohols with peroxides. Further investigation of the "sesquioxane O-ligand plus N,N-ligand" approach to isolation of catalytically prospective Cu(II)-based structures is currently on way in our team. A new method for the determination of the participation of hydroxyl radicals was developed.