Preparative and Catalytic Properties of MoVI Mononuclear and Metallosupramolecular Coordination Assemblies Bearing Hydrazonato Ligands

A series of polynuclear, dinuclear, and mononuclear Mo(VI) complexes were synthesized with the hydrazonato ligands derived from 5-methoxysalicylaldehyde and the corresponding hydrazides (isonicotinic hydrazide (H2L1), nicotinic hydrazide (H2L2), 2-aminobenzhydrazide (H2L3), or 4-aminobenzhydrazide (H2L4)). The metallosupramolecular compounds obtained from non-coordinating solvents, [MoO2(L1,2)]n (1 and 2) and [MoO2(L3,4)]2 (3 and 4), formed infinite structures and metallacycles, respectively. By blocking two coordination sites with cis-dioxo ligands, the molybdenum centers have three coordination sites occupied by the ONO donor atoms from the rigid hydrazone ligands and one by the N atom of pyridyl or amine-functionalized ligand subcomponents from the neighboring Mo building units. The reaction in methanol afforded the mononuclear analogs [MoO2(L1-4)(MeOH)] (1a–4a) with additional monodentate MeOH ligands. All isolated complexes were tested as catalysts for cyclooctene epoxidation using tert-butyl hydroperoxide (TBHP) as an oxidant in water. The impact of the structure and ligand lability on the catalytic efficiency in homogeneous cyclooctene epoxidation was elucidated based on theoretical considerations. Thus, dinuclear assemblies exhibited better catalytic activity than mononuclear or polynuclear complexes.


Introduction
Metallosupramolecular compounds are species formed from metal ions and organic ligands in coordination-driven self-assembly processes [1][2][3][4][5].Aesthetically fascinating architectures can be self-assembled from the same building blocks depending on metal coordination geometry, ligand structure, metal-to-ligand ratio, solvent, concentration, and presence of guests.The formation of zero-dimensional, polyhedral, or infinite structures reflects a complex interplay between kinetic and thermodynamic contributions [6][7][8][9].
The interest in this class of compounds comes not only from the challenges related to the development of reliable synthetic strategies and rational design but also from their wide-ranging potential.Thus, those assemblies can find application in catalysis [10,11], optics [12][13][14], magnetism [15][16][17], gas storage separation [18,19], and as sensors [20,21].Considering the importance of integrating green chemistry principles and sustainability into research, the development of novel catalysts is essential, particularly those that can exhibit high activity and selectivity at mild operating conditions, which is of high interest.
The catalytic properties of molybdenum coordination compounds [22][23][24][25] have led to an increased interest in their metallosupramolecular chemistry in years.As a result, the investigation has grown from essential design principles enabling the formation of the first coordination polymers and metallocycles [26][27][28][29] to exploration of their structural and compositional fine-tuning in order to obtain catalysts with excellent properties [30][31][32][33][34][35].Accordingly, constrained cyclic Mo(VI) metallosupramolecular complexes were recently found to be among the most effective catalysts for olefin epoxidation [20,21].
There is still a promising pool of catalysts based on metallosupramolecular or simple mononuclear (dioxo)Mo(VI) coordination assemblies.Such materials can utilize the advantages of organic components for fine structural tuning and, consequently, their properties.In particular, the compounds with hemilabile coordination bonds are of interest as they could serve as high-efficiency catalysts.For example, the high performance of hexacoordinated Mo(VI) (pre)catalysts closely correlates with their possibility of converting into the pentacoordinated counterparts [34,36].
To explore the (dioxo)Mo(VI) complexes and their catalytic properties, four hydrazone ligands have been selected (Scheme 1) where the subunit derived from 5-methoxysalicylaldehyde remained unchanged.It was expected that the catalytic activity could be modulated by the hydrazide moiety.Self-assembly of {MoO 2 } 2+ units with ligands bearing (iso)nicotinoyl or aminobenzoyl moieties in dichloromethane yielded metallosupramolecular polymers and dimers, respectively, based on metal-nitrogen linkages.The reaction in methanol gave rise to mononuclear coordination compounds containing methanol (MeOH) ancillary ligands with metal-oxygen bonds.Crystal and molecular structures of the two hydrazones, one dinuclear, and three mononuclear coordination compounds were determined using the single-crystal X-ray diffraction method.All compounds were additionally characterized using elemental analysis, powered X-ray diffraction, thermogravimetric, and spectroscopic methods.
The catalytic properties of molybdenum coordination compounds [22][23][24][25] have led to an increased interest in their metallosupramolecular chemistry in years.As a result, the investigation has grown from essential design principles enabling the formation of the first coordination polymers and metallocycles [26][27][28][29] to exploration of their structural and compositional fine-tuning in order to obtain catalysts with excellent properties [30][31][32][33][34][35].Accordingly, constrained cyclic Mo(VI) metallosupramolecular complexes were recently found to be among the most effective catalysts for olefin epoxidation [20,21].
There is still a promising pool of catalysts based on metallosupramolecular or simple mononuclear (dioxo)Mo(VI) coordination assemblies.Such materials can utilize the advantages of organic components for fine structural tuning and, consequently, their properties.In particular, the compounds with hemilabile coordination bonds are of interest as they could serve as high-efficiency catalysts.For example, the high performance of hexacoordinated Mo(VI) (pre)catalysts closely correlates with their possibility of converting into the pentacoordinated counterparts [34,36].
To explore the (dioxo)Mo(VI) complexes and their catalytic properties, four hydrazone ligands have been selected (Scheme 1) where the subunit derived from 5-methoxysalicylaldehyde remained unchanged.It was expected that the catalytic activity could be modulated by the hydrazide moiety.Self-assembly of {MoO2} 2+ units with ligands bearing (iso)nicotinoyl or aminobenzoyl moieties in dichloromethane yielded metallosupramolecular polymers and dimers, respectively, based on metal-nitrogen linkages.The reaction in methanol gave rise to mononuclear coordination compounds containing methanol (MeOH) ancillary ligands with metal-oxygen bonds.Crystal and molecular structures of the two hydrazones, one dinuclear, and three mononuclear coordination compounds were determined using the single-crystal X-ray diffraction method.All compounds were additionally characterized using elemental analysis, powered X-ray diffraction, thermogravimetric, and spectroscopic methods.
The molybdenum-catalyzed epoxidation of alkenes is an important process as it allows the conversion of simple hydrocarbon substrates into high-value fine chemicals.We, therefore, explored the effects of Mo-N (N being a pyridyl or amine nitrogen atom) or Mo-O (O being a MeOH oxygen atom) bond liability on the catalytic performances of materials obtained in the oxidation of cyclooctene.As in our previous works, we focused on catalytic reactions under environmentally friendly conditions without organic solvents.The catalytic efficiency of the Mo-coordination entities was evaluated using tertbutyl hydroperoxide as the oxidant in water.Catalytic results are discussed in relation to the complex nuclearity, coordination environment, and ligand structure.The obtained Scheme 1. Hydrazones derived from 5-methoxysalicylaldehyde and the corresponding hydrazides: isonicotinic hydrazide (H 2 L 1 ), nicotinic hydrazide (H 2 L 2 ), 2-aminobenzhydrazide (H 2 L 3 ), or 4-aminobenzhydrazide (H 2 L 4 ).Previously published ligands derived from 3-or 4methoxysalicylaldehyde are denoted as H 2 L ′ and H 2 L ′′ , respectively.
The molybdenum-catalyzed epoxidation of alkenes is an important process as it allows the conversion of simple hydrocarbon substrates into high-value fine chemicals.We, therefore, explored the effects of Mo-N (N being a pyridyl or amine nitrogen atom) or Mo-O (O being a MeOH oxygen atom) bond liability on the catalytic performances of materials obtained in the oxidation of cyclooctene.As in our previous works, we focused on catalytic reactions under environmentally friendly conditions without organic solvents.The catalytic efficiency of the Mo-coordination entities was evaluated using tert-butyl hydroperoxide as the oxidant in water.Catalytic results are discussed in relation to the complex nuclearity, coordination environment, and ligand structure.The obtained results were compared to the previously published results utilizing corresponding H 2 L ′ and H 2 L ′′ ligands, respectively (Scheme 1).

Hydrazones
Hydrazones H 2 L 1 -H 2 L 4 (Scheme 1) were afforded by reacting equimolar amounts of 5-methoxysalicylaldehyde with the corresponding hydrazides (isonicotinic hydrazide, nicotinic hydrazide, 2-aminobenzhydrazide, or 4-aminobenzhydrazide) in a methanolic solution.In the case of H 2 L 1 and H 2 L 2 , the reagents were heated under reflux.The synthesis of H 2 L 3 and H 2 L 4 required lowering the reaction temperature primarily to avoid the secondary condensation reaction of the aldehyde with the amine functional group.All compounds, except H 2 L 4 , remained stable in the solid state for a longer time at room temperature.However, upon exposure to the air, hydrazone H 2 L 4 transformed over time to the more stable hydrate The crystal structures of H 2 L 1 and H 2 L 1 •H 2 O were reported previously [37,38].Ligands H 2 L 3 and H 2 L 4 crystallize in orthorhombic Pna21 and triclinic P-1 space groups, respectively (Figure 1a, Supplementary Materials, Figure S1, Table S1).The crystal structure of H 2 L 3 contains one molecule of ligand in the asymmetric unit, which packs in an alternating zig-zag fashion (Supplementary Materials, Figure S2), as dictated by a singular intermolecular hydrogen bond between hydrazide N-H donor and aryl hydroxy group acceptor (Supplementary Materials, Figure S3a).Meanwhile, in the crystal structure of H 2 L 4 , one can find two symmetrically inequivalent ligand molecules, which form a complex hydrogenbonded network via interactions between hydrazide and aryl amine donors, and amide keto and aryl hydroxy group acceptors (Supplementary Materials, Figure S3b).However, both ligands crystallize in keto-amino tautomeric form (Supplementary Materials, Scheme S1), as evidenced by C2-O2, C2-N2, and N2-N1 bond lengths (Supplementary Materials, Table S2), which correspond to lengths expected for double, single and double bond, respectively.Additionally, both ligands are in E conformation (Supplementary Materials, Table S3).Hydrogen bond parameters in the crystal structure of H 2 L 3 and H 2 L 4 are given in Table S4 (Supplementary Materials).
results were compared to the previously published results utilizing corresponding H2L′ and H2L′′ ligands, respectively (Scheme 1).

Synthesis and Structural Characteriation
2.1.1.Hydrazones Hydrazones H2L 1 -H2L 4 (Scheme 1) were afforded by reacting equimolar amounts of 5-methoxysalicylaldehyde with the corresponding hydrazides (isonicotinic hydrazide, nicotinic hydrazide, 2-aminobenzhydrazide, or 4-aminobenzhydrazide) in a methanolic solution.In the case of H2L 1 and H2L 2 , the reagents were heated under reflux.The synthesis of H2L 3 and H2L 4 required lowering the reaction temperature primarily to avoid the secondary condensation reaction of the aldehyde with the amine functional group.All compounds, except H2L 4 , remained stable in the solid state for a longer time at room temperature.However, upon exposure to the air, hydrazone H2L 4 transformed over time to the more stable hydrate H2L 4 •H2O.
The crystal structures of H2L 1 and H2L 1 •H2O were reported previously [37,38].Ligands H2L 3 and H2L 4 crystallize in orthorhombic Pna21 and triclinic P-1 space groups, respectively (Figure 1a, Supplementary Materials, Figure S1, Table S1).The crystal structure of H2L 3 contains one molecule of ligand in the asymmetric unit, which packs in an alternating zig-zag fashion (Supplementary Materials, Figure S2), as dictated by a singular intermolecular hydrogen bond between hydrazide N-H donor and aryl hydroxy group acceptor (Supplementary Materials, Figure S3a).Meanwhile, in the crystal structure of H2L 4 , one can find two symmetrically inequivalent ligand molecules, which form a complex hydrogen-bonded network via interactions between hydrazide and aryl amine donors, and amide keto and aryl hydroxy group acceptors (Supplementary Materials, Figure S3b).However, both ligands crystallize in keto-amino tautomeric form (Supplementary Materials, Scheme S1), as evidenced by C2-O2, C2-N2, and N2-N1 bond lengths (Supplementary Materials, Table S2), which correspond to lengths expected for double, single and double bond, respectively.Additionally, both ligands are in E conformation (Supplementary Materials, Table S3).Hydrogen bond parameters in the crystal structure of H2L 3 and H2L 4 are given in Table S4 (Supplementary Materials).

Metallosupramolecular Coordination Assemblies
The Mo(VI) metallosupramolecular complexes 1-4 were synthesized via the reaction of [MoO2(acac)2] (acac = acetylacetonate) with H2L 1 -H2L 4 , respectively, in CH2Cl2 at room temperature.The formulae of [MoO2(L 1,2 )]n (1 and 2) and [MoO2(L 3,4 )]2 (3 and 4), containing the corresponding ligand in a doubly deprotonated form, are consistent with elemental, thermogravimetric analysis, spectroscopic, and X-ray diffraction data.Their formation is in accordance with the energy-related principle of "maximum site occupancy", according to which species with all the coordination sites occupied are more stable than those with vacant sites [39].
Synthesis with ligands H2L 1 or H2L 2 bearing nitrogen-containing heterocyclic ring resulted in the coordination polymers, comprising Mo coordinated by one hydrazone ligand through ONO donor atoms (phenolic-oxygen, azomethine-nitrogen, and hydrazidic-oxygen) and the second ligand through the isonicotinoyl N atom from the adjacent Mo complex unit.Although polymer 1 has been previously reported (CSD refcode ZILVIB) [40], it was prepared under different conditions.The polymer forms infinite chains, the topology of which is dictated by the rigid geometry of the isonicotinoyl fragment and {MoO2} 2+ core.The only degree of freedom in this chain is the Mo-Npyridyl bond length, which is slightly longer than expected for similar complexes with ONO ligand coordinated on the {MoO2} 2+ core and pyridyl fragment coordinated on the free axial site.Particularly, the bond length in 1 is found to be 2.455 Å as opposed to an average of 2.378 ± 0.082 Å for 102 structures in the CSD database [41].This indicates significant steric impedance in the assembly of 1 polymer.
Ligands modified by incorporating the amino functionality, H2L 3 and H2L 4 , gave rise to the cyclic dimers [MoO2(L) 3,4 ]2 (3 and 4).The molecular structure of 3 is shown in Figure 1c.Selected geometric data are given in Supplementary Materials, Table S2.The formation of the dimeric compound [MoO2(L 4 )]2 (4) was proposed based on literature data [34].Additionally, the favored formation of dimers over polymers is related to the preferred formation of smaller molecules containing fewer building blocks [42].
The structure of 3 (Supplementary Materials, Figures S1, S2, and S4) exhibits a distorted octahedral geometry around the Mo(VI) metal center (Supplementary Materials, Tables S6 and S7).The ONO ligand chelates in a meridional fashion occupying three coordination sites.Two cis-oriented oxido ligands and the amino N atom from the adjacent Mo complex unit complete the coordination sphere.Similar to 1, one can analyze the oligomerization "propensity" based on the observed Mo-Namine length.In 3, that length is 2.536(2) Å, whereas for the 11 known examples in the CSD database, the mean length is 2.534 ± 0.098 Å.One must observe that complexes derived from H2L 3 and H2L 4 have greater conformational flexibility when considering the coordination of the amine nitro-

Metallosupramolecular Coordination Assemblies
The Mo(VI) metallosupramolecular complexes 1-4 were synthesized via the reaction of [MoO 2 (acac) 2 ] (acac = acetylacetonate) with H 2 L 1 -H 2 L 4 , respectively, in CH 2 Cl 2 at room temperature.The formulae of [MoO 2 (L 1,2 )] n (1 and 2) and [MoO 2 (L 3,4 )] 2 (3 and 4), containing the corresponding ligand in a doubly deprotonated form, are consistent with elemental, thermogravimetric analysis, spectroscopic, and X-ray diffraction data.Their formation is in accordance with the energy-related principle of "maximum site occupancy", according to which species with all the coordination sites occupied are more stable than those with vacant sites [39].
Synthesis with ligands H 2 L 1 or H 2 L 2 bearing nitrogen-containing heterocyclic ring resulted in the coordination polymers, comprising Mo coordinated by one hydrazone ligand through ONO donor atoms (phenolic-oxygen, azomethine-nitrogen, and hydrazidicoxygen) and the second ligand through the isonicotinoyl N atom from the adjacent Mo complex unit.Although polymer 1 has been previously reported (CSD refcode ZILVIB) [40], it was prepared under different conditions.The polymer forms infinite chains, the topology of which is dictated by the rigid geometry of the isonicotinoyl fragment and {MoO 2 } 2+ core.The only degree of freedom in this chain is the Mo-N pyridyl bond length, which is slightly longer than expected for similar complexes with ONO ligand coordinated on the {MoO 2 } 2+ core and pyridyl fragment coordinated on the free axial site.Particularly, the bond length in 1 is found to be 2.455 Å as opposed to an average of 2.378 ± 0.082 Å for 102 structures in the CSD database [41].This indicates significant steric impedance in the assembly of 1 polymer.
Ligands modified by incorporating the amino functionality, H 2 L 3 and H 2 L 4 , gave rise to the cyclic dimers [MoO 2 (L) 3,4 ] 2 (3 and 4).The molecular structure of 3 is shown in Figure 1c.Selected geometric data are given in Supplementary Materials, Table S2.The formation of the dimeric compound [MoO 2 (L 4 )] 2 (4) was proposed based on literature data [34].Additionally, the favored formation of dimers over polymers is related to the preferred formation of smaller molecules containing fewer building blocks [42].
The structure of 3 (Supplementary Materials, Figures S1, S2 and S4) exhibits a distorted octahedral geometry around the Mo(VI) metal center (Supplementary Materials, Tables S6 and S7).The ONO ligand chelates in a meridional fashion occupying three coordination sites.Two cis-oriented oxido ligands and the amino N atom from the adjacent Mo complex unit complete the coordination sphere.Similar to 1, one can analyze the oligomerization "propensity" based on the observed Mo-N amine length.In 3, that length is 2.536(2) Å, whereas for the 11 known examples in the CSD database, the mean length is 2.534 ± 0.098 Å.One must observe that complexes derived from H 2 L 3 and H 2 L 4 have greater conformational flexibility when considering the coordination of the amine nitrogen atom, compared to the coordination of the pyridyl nitrogen atom in 1 and 2. The PXRD patterns of 1 and 3 were in good agreement with the simulated patterns of the corresponding crystal structures (Supplementary Materials, Figure S5).

Mononuclear Coordination Assemblies
In the reaction between the [MoO 2 (acac) 2 ] complex and hydrazone ligands in methanol, the coordination of the pyridyl/amine moiety was not observed.Instead, the [MoO 2 (L 1-4 )(MeOH)] (1a-4a) compounds crystallized the additional monodentate MeOH ligand (Figure 1b, Supplementary Materials, Figure S1).Complexes 1a-4a were obtained in higher yield (64-89%) than the metallosupramolecular species 1-4 (31-38%).A single-crystal X-ray study of 1a, 2a, and 3a confirmed the structures to be mononuclear.Experimental powder diffractograms of all obtained samples matched well with the calculated ones (Supplementary Materials, Figure S6).Complexes 1a-4a were stable in the solid state at room temperature.However, their quantitative conversion to polynuclear 1 and 2 or dinuclear forms 3 and 4 was observed upon heating in acetone or acetonitrile.
The crystal structures of 1a, 2a, and 3a (Supplementary Materials, Figures S1 and S2) show features already well established for the majority of [MoO 2 (L)(MeOH)] complexes.The {MoO 2 } 2+ core has the same coordinative environment as in polymers (see Section 2.1.2),except for the axial coordination site occupied with oxygen atoms from methanol molecule.Sterically undemanding methanol ligand has a well-defined coordinative bond length of 2.343 ± 0.035 Å based on 206 examples in the CSD database, and complexes 1a, 2a and 3a have this length in the same ballpark, albeit a bit shorter (2.313(2), 2.295(7), and 2.287(4) Å, respectively).Selected bond lengths and angles are given in the Supplementary Materials, Tables S6 and S7.
Crystal packing (Supplementary Materials, Figure S2) and supramolecular interactions in prepared methanol complexes are dictated by a small number of hydrogen bond donors (namely, methanol OH and amine NH 2 moieties), forming relatively simple supramolecular chains.However, due to the presence of an amino group donor, supramolecular chain in 3a is built upon supramolecular dimers, akin to coordinative dimers of 3, whereas supramolecular chains in 1a and 2a are made of monomers interacting solely through methanol OH•••N pyridyl hydrogen bond (Supplementary Materials, Figure S4, Table S8).

Spectroscopic Characterization
The infrared spectra confirmed the presence of the amide tautomer, Supplementary Materials, Figure S7.The intense absorptions in the range of 1695-1653 cm −1 and 1654-1626 cm −1 are attributed to the C=O and C=N imine bond vibrations, respectively.Hydrazones H 2 L 1 and H 2 L 2 exhibited bands at approximately 3260 cm −1 and 3220 cm −1 assigned to -O-H and =N-NH, whereas these bands appeared at approximately 3500 cm −1 and 3400 cm -1 for H 2 L 3 and H 2 L 4 .The medium intensity band observed at 1163 cm −1 -1115 cm −1 was assigned to N-N stretching vibrations [43].
The stretch bands of C=O, -O-H, and =N-NH that were characteristic of free hydrazones disappeared upon metal complexation.The presence of bands in 1-4 belonging to C-O hydrazone (at 1320-1345 cm -1 ), C-O phenolic (at 1260-1275 cm -1 ), and C=N (at 1600-1610 cm -1 ) indicated hydrazone functionality tautomerization N-NH-C=O → N-N=C-OH, deprotonation, and the molybdenum atom coordination through hydrazone ONO donor atoms.Strong absorption bands were characteristic of the {MoO 2 } 2+ core in the range of 935-908 cm -1 and weaker intensity bands for Mo-N bonds below 900 cm -1 corroborated the metallosupramolecular complex formation [34].The methanol coordination in 1a-4a was supported by a new absorption band typical of C-O MeOH at around 1030 cm −1 .Additionally, the mononuclear complexes exhibited stretching frequencies O=Mo-O MeOH around 900 cm −1 .The -NH 2 bond vibrations of the mononuclear complexes present as two medium-intensity bands (3356-3263 cm -1 ) are higher in energy than those of the dinuclear complexes (3334-3227 cm -1 ), Supplementary Materials, Figures S8 and S9.A similar position of these bands was found in the spectra of the previously published complexes [MoO 2 (L 4′ )] 2 and [MoO 2 (L 4′′ )] 2 [34].
The proton and carbon nuclear magnetic resonance (NMR) chemical shifts of H 2 L 1 -H 2 L 4 (Supplementary Materials, Tables S9-S12, Schemes S2-S5, Figures S10-S13) were assigned by 1 H, 13 C{ 1 H} attached proton test (APT), 1 H- 13 C heteronuclear multiple quantum coherence (HMQC), and heteronuclear multiple bond correlation (HMBC) experiments in dimethyl sulfoxide-d 6 (DMSO-d 6 ).Downfield shifts in the range of 11.02-10.66ppm and 12.37-11.88ppm suggested the presence of -OH and =N-NH groups, respectively, thus indicating the amide form =N-NH-(C=O)-in solution.Singlets at 6.54 and 5.89 confirmed the presence of -NH 2 protons in the spectra of H 2 L 3 and H 2 L 4 , respectively [44].The OH, N=NH, and NH 2 signals were broadened to some extent due to their involvement in hydrogen bonding.Singlets at 8.75-8.63ppm and 3.77-3.73ppm were assigned to the CH=N and OCH 3 groups, respectively.The aromatic moieties gave signals in the range of 9.19-6.65 ppm [45].In the 13  Due to the donor properties of the solvent DMSO-d 6 , the metallosupramolecular nature of complexes 1-4 and the presence of ancillary ligand in the mononuclear complexes 1a-4a were disrupted, and complexes [(MoO 2 ) 2 (L 1-4 )(DMSO)] appeared instead.The signals for the (iso)nicotinoyl and aminobenzoyl protons did not exhibit any appreciable change in the chemical shift.In the 1 H NMR spectra, the absence of -OH and =N-NH protons in the low field region indicated their deprotonation upon complexation.Multiple groups of signals in the range of 6.57-9.15ppm were due to aromatic ring protons.The most significant coordination-induced difference between signals was noticed for imine CH=N, which was up to 0.3 ppm.In the 13 C NMR spectra, the complexation-induced differences in the chemical shifts were the most pronounced for C1 and C4 carbons, which experience deshielding effects up to 9.28 ppm and 6.74 ppm, respectively.The chemical shifts for phenolic C12 carbons differed less noticeably, which was up to 2.6 ppm.

Thermal Analysis
The first mass loss in the thermogravimetric (TG) curves of H 2 L 4 •H 2 O is related to the removal of a water molecule from a crystalline hydrate (starting at 105 • C).The data from the differential scanning calorimetry (DSC) measurement shows a broad endothermic peak due to the evaporation of the water during the dehydration process.From DSC measurements, the melting onset point temperature for anhydrous hydrazone H 2 L 1  The thermal behavior of the complexes was studied in the temperature region from 25 to 600 • C in an oxygen atmosphere.The polynuclear and dinuclear complexes 1-4 demonstrated thermal stability of up to 300 • C. Afterwards, the decomposition of complexes occurred in one step.The TG curves of 1a-4a showed two distinct mass losses: the first one due to MeOH removal at temperatures around 140 • C, and the second one due to decomposition (with onset temperatures of 333 • C, 317 • C, 286 • C, and 281 • C, respectively).

Cyclooctene Epoxidation with TBHP
Cyclooctene epoxidation reactions were performed with tert-butylhydroperoxide (TBHP) in water as an oxidizing agent.The idea was to test the prepared molybdenum complexes as catalysts, following the protocol previously established, to compare the activity and selectivity parameters with the results already published.For that reason, the green chemistry protocol was applied.Even though TBHP is commercially available in decane and water, and the catalytic results with the assistance of TBHP in decane are often much better, the choice of TBHP in water justifies the green catalytic concept.All the complexes were tested as catalysts with 0.25 % catalyst loading.The obtained results are shown in Table 1 and the catalytic profiles are presented in Figure 2.
Due to the distinct organic and aqueous phases in the catalytic reaction, the analysis focused solely on the organic phase using gas chromatography (GC).The presence of epoxide in the organic phase was identified, with the selectivity parameter ranging from 80 to 98%.Consistent with previous reports, the GC analysis did not extend to the aqueous layer.Drawing from our earlier studies, it has been established that the epoxide may undergo opening to form the corresponding diol, contingent on the presence of the diol in the water phase.
Int. J. Mol.Sci.2023, 24, x FOR PEER REVIEW 7 of 17 much better, the choice of TBHP in water justifies the green catalytic concept.All the complexes were tested as catalysts with 0.25 % catalyst loading.The obtained results are shown in Table 1 and the catalytic profiles are presented in Figure 2. Due to the distinct organic and aqueous phases in the catalytic reaction, the analysis focused solely on the organic phase using gas chromatography (GC).The presence of epoxide in the organic phase was identified, with the selectivity parameter ranging from 80 to 98%.Consistent with previous reports, the GC analysis did not extend to the aqueous layer.Drawing from our earlier studies, it has been established that the epoxide may undergo opening to form the corresponding diol, contingent on the presence of the diol in the water phase.All the tested catalysts showed very good activity, with cyclooctene conversion being more than 85 %, with the exception of the dinuclear catalyst 4 (67 %).TOF20 min parameter is the highest for dinuclear catalyst 3, meaning that it converts to the catalytically active species in the shortest time, and it is the lowest for catalyst 2. It can also be noticed that the catalytic profiles of catalysts 1a and 2a had a very similar pattern, directly implying that the position of the N atom on the hydrazide part of the ligand (third or fourth atom) had almost no effect on the catalytic behavior.Furthermore, a similar trend was also noticed for catalysts 3a and 4a, indicating that the position of the amino group (on the second or fourth C atom) did not influence the final activity outcome.What is quite interesting is  All the tested catalysts showed very good activity, with cyclooctene conversion being more than 85 %, with the exception of the dinuclear catalyst 4 (67 %).TOF 20 min parameter is the highest for dinuclear catalyst 3, meaning that it converts to the catalytically active species in the shortest time, and it is the lowest for catalyst 2. It can also be noticed that the catalytic profiles of catalysts 1a and 2a had a very similar pattern, directly implying that the position of the N atom on the hydrazide part of the ligand (third or fourth atom) had almost no effect on the catalytic behavior.Furthermore, a similar trend was also noticed for catalysts 3a and 4a, indicating that the position of the amino group (on the second or fourth C atom) did not influence the final activity outcome.What is quite interesting is the comparison of the TOF parameter.It is obvious that the dinuclear compounds 3 and 4, with the TOF 20 min values of 415 and 279, respectively, are more easily converted to the catalytically active species than the polynuclear compounds 1 and 2. On the contrary, the mononuclear complexes 1a and 2a have very similar TOF 20 min values (around 350), almost twice higher than the TOF 20 min values for 3a and 4a.
All the obtained results can be compared to the similar published results compiled in Table 2.The hydrazone derived from 3-or 4-methoxysalicylaldehyde with the corresponding hydrazides (nicotinic hydrazide, 2-aminobenzhydrazide or 4-aminobenzhydrazide) used for the coordination to the molybdenum center are presented in Scheme 1.As noted, complex 2 can be compared to the complexes Furthermore, the selectivity parameter followed the same trend, meaning that the position of the -OMe group on the aldehyde part of the ligands has a strong impact on the catalytic performances, favoring the fifth position on the benzene ring.
On the other side, the influence of the -OMe group on the aldehyde part of the ligands was not so pronounced when comparing catalysts 3 and 3a with [MoO 2 (L 3 ′ )] 2 , [MoO 2 (L 3 ′ ′ )(MeOH)] and [MoO 2 (L 3 ′ ′ )] 2 .For the TOF and activation energy, there is a bigger difference between 3 (the highest TOF) and 3a (the lowest one) compared to the three other complexes with relatively similar TOF for both dimers and slightly higher for the methanol-coordinated species.This behavior seems to indicate that the possibility of the decoordination and formation of the pentacoordinated active species has a strong impact on the mechanisms.
However, the -OMe position impacts the activity of the complexes 4 and 4a when cor- have lower conversion and TOF than catalyst 4a, while the cyclooctene conversion for dinuclear catalysts has an unusual trend [MoO 2 (L 4 ′ ′ )] 2 > [MoO 2 (L 4 ′ )] 2 > 4 but with very close TOF values.In this case, we can conclude that there is an effect of methanol that has to be investigated in order to explain the mechanism.However, in general, it can be concluded that the catalysts bearing -OMe group on the fifth C atom of the benzene ring show better catalytic performances, except in the case when NH 2 is in the fourth position on its aromatic ring, complex 4.

Density Functional Theory (DFT) Calculations
It is interesting to see if the experimental results can be validated by theoretical calculations to understand the mechanisms and the effect of the ligand on the catalytic activity.The main points responsible for such activity and the correlation of the catalytic activity with the structure are possible to be answered when the ligand functionalization is very different.However, subtle changes, as in the presented research, are sometimes difficult to analyze using DFT calculations.
The mechanism proposed several years ago [36] postulated that [MoO 2 (L)(D)]or [MoO 2 (L)] n species turned into a pentacoordinated species [MoO 2 (L)] before reacting with TBHP to create the transient species [MoO 2 (L)(TBHP)], the latter permitting an easier epoxide formation through the O β of the epoxide.It has been shown that the presence of substituents with strong electronic effects on the ligands had dramatic effects on the catalysis [36].These results were linked to different electronic effects that could even be assessed using the geometrical features of the calculated species.
In the present case, the differences between the four ligands lie in the hydrazide nature: N atom is within the aromatic ring (in the case of complexes with L 1-and L 2-ligands) or bonded to the aromatic ring as NH 2 (in the case of complexes with L 3-and L 4 ligands).In the case of coordination polymers 1 and 2, an insight into decoordination through DFT would be too complicated.Therefore, we limited the discussion about the decoordination to complexes 3 and 4. The crystal structure of compound 3 was determined as a dimer using the X-ray diffraction method, and therefore, the calculations for compound 4 were adopted according to the known structure.Further calculations included mononuclear complexes 1a-4a.
We limited the calculations to the gas phase since trends can be observed from these conditions [47], and the data are shown in Table 3.All other calculation steps are shown in Scheme 2. The optimized geometries are listed in Supplementary Materials, Table S14.
very different.However, subtle changes, as in the presented research, are sometimes difficult to analyze using DFT calculations.
The mechanism proposed several years ago [36] postulated that [MoO2(L)(D)]or [MoO2(L)]n species turned into a pentacoordinated species [MoO2(L)] before reacting with TBHP to create the transient species [MoO2(L)(TBHP)], the latter permitting an easier epoxide formation through the Oβ of the epoxide.It has been shown that the presence of substituents with strong electronic effects on the ligands had dramatic effects on the catalysis [36].These results were linked to different electronic effects that could even be assessed using the geometrical features of the calculated species.
In the present case, the differences between the four ligands lie in the hydrazide nature: N atom is within the aromatic ring (in the case of complexes with L 1-and L 2-ligands) or bonded to the aromatic ring as NH2 (in the case of complexes with L 3-and L 4 ligands).In the case of coordination polymers 1 and 2, an insight into decoordination through DFT would be too complicated.Therefore, we limited the discussion about the decoordination to complexes 3 and 4. The crystal structure of compound 3 was determined as a dimer using the X-ray diffraction method, and therefore, the calculations for compound 4 were adopted according to the known structure.Further calculations included mononuclear complexes 1a-4a.
We limited the calculations to the gas phase since trends can be observed from these conditions [47], and the data are shown in Table 3.All other calculation steps are shown in Scheme 2. The optimized geometries are listed in Supplementary Materials, Table S14.(L 4 )(MeOH)] (4a).It has to be considered that, once the methanol decoordinated, the resulting pentacoordinated complex [MoO 2 (L)] can have several options: (i) turning back into the mononuclear complex [MoO 2 (L)(D)], (ii) oligomerize or (iii) interact with TBHP.From the absolute enthalpy values obtained by calculations for each step (Table 3), the decoordination of the methanol (step 0) lies in the range of 10.3 to 11.1 kcal and these values do not provide a clear conclusion.In an identical manner, in the case of the stabilization of the formed pentacoordinate species [MoO 2 (L)] with TBHP on the Mo center (step I), the values are also very close for all four ligands.On the other side, TS enthalpies (step II) were ligand-dependent and the trend observed was TS 2 < TS 1 < TS 3 < TS 4 .The lower the TS, the more easily the active species are formed, and consequently, TOF 20 min should be higher.This is in relative accordance with the experimentally obtained TOF values for the complexes 1a-4a, TOF 2a ≈ TOF 1a > TOF 3a >TOF 4a .
In contrast, for the dinuclear complexes 3 and 4, the order of the experimental TOFs, TOF 4 < TOF 3 cannot follow the calculated TS values, TS 3 > TS 4 .It is experimentally expected that dimers 3 and 4 are transformed more easily than polymers 1 and 2.Even if it is not simple to draw conclusions between dimers and polymers, a comparison between dimers and methanol-coordinated monomers 3 and 3a (and 4 and 4a) can be performed.At this point, the experimental TOF values can be linked to the TS shift considering the enthalpy of the starting molecule.The enthalpy to maintain those dimers is smaller than that for the methanol-coordinated ones, and this difference provokes a relative TS around 5 kcal lower for the polymers.This seems to be relevant enough to give some ideas about the TOF difference in this case.
It is interesting to see whether these activities can be deduced from the calculated geometric data.The relevant distances and dihedral angles in absolute values (related to ligand planarity) around the molybdenum atom are listed in Table 4.The experimental results showed that the [MoO 2 (L 3 )] 2 (3) species seemed to be the best catalyst.Thus, as shown in the energy diagram, the Mo-O β interatomic distance parameter mainly influenced the TS path (step II).As explained previously, a lower TS value (higher activity) is linked to a shorter distance.This is the case if we compare the analog complexes with L 3 and L 4 ligands.Interesting facts appeared with the N1 bonded to Mo, such as the shorter Mo-N 1 distances in the TS were for the least active compounds.
Ligand distortion seemed to have an influence.While the [MoO 2 (L)(MeOH)] dihedral angle lies between 9.5 and 11.3 • , the removal of the MeOH molecule favors the planarity (5.6 to 6.3) addition of TBHP increases the dihedral angle and, in TS, distortion is the highest.

Preparative Part
Elemental analyses were provided by the Analytical Services Laboratory of the Ruđer Bošković Institute, Zagreb.NMR data are given in Tables S9-S12 and Figures S5-S8 in the Supplementary Materials.All reagents and solvents were commercially available (Alpha-Aesar (Karlsruhe, Germany), Sigma-Aldrich (Chemie GmbH, München, Germany)) and were used without further purification.[MoO2(acac)2] was prepared according to the published procedure [48].

Synthesis of H2L 1 and H2L 2
A mixture of 5-methoxysalicylaldehyde (0.76 g, 5 mmol) and isonicotinic or nicotinic hydrazide (0.69, 5 mmol) in 100 mL of methanol was heated at reflux temperature with continuous stirring for 3 h.The resulting solution was concentrated under reduced pressure to one-quarter of its volume and left at room temperature for several days.The obtained product H2L 1 or H2L 2 was filtered and dried to a constant weight.Yield: 1.05 g, 77% (H2L 1 ); 1.14 g, 84% (H2L 2 ).

Synthesis of H 2 L 1 and H 2 L 2
A mixture of 5-methoxysalicylaldehyde (0.76 g, 5 mmol) and isonicotinic or nicotinic hydrazide (0.69, 5 mmol) in 100 mL of methanol was heated at reflux temperature with continuous stirring for 3 h.The resulting solution was concentrated under reduced pressure The GC parameters were assessed using authentic samples of reactants and products.The conversion of olefins and the formation of the corresponding epoxides were determined based on calibration curves, with acetophenone as the internal standard.

Theoretical Calculations
The geometries of the investigated species were optimized using the DFT approach with Gaussian 09 rev.D01 program suite [54] using the B3LYP three-parameter functional [55][56][57] in conjunction with the 6-31G* basis set [58][59][60][61] for light atoms (O, N, C, H) and the CEP-31G set for Mo atoms [62,63].The geometries of the complexes and intermediates were optimized using geometries inspired by X-ray diffraction.All coordinates are listed in the Supplementary Materials (Table S14).The frequency analysis confirmed that the optimized geometries were local minima.The full optimization of the transition states was guided by a preliminary scan of the relevant internal coordinate.By frequency analyses, all optimized geometries were confirmed to be stationary points and local minima for stable molecules or reaction intermediates and first-order saddle points for TSs.For TSs, the analysis of the imaginary frequency confirmed the expected motion along the chosen reaction coordinate.These values and related schemes are shown in Table S13.The calculated frequencies were also used to obtain the thermochemical parameters at 298 K according to the standard approximations (ideal gas, rigid rotor, and harmonic oscillator).

Conclusions
The structural formations of molybdenum complexes can vary based on the specific hydrazide employed.Depending on this hydrazide moiety (aminobenzohydrazide vs. (iso)nicotinic hydrazide), structural assemblies such as dinuclear or polynuclear molybdenum complexes can be attained.Furthermore, the selection of the solvent plays a pivotal role as it determines whether a mononuclear (from methanol) or dinuclear/polynuclear (from dichloromethane) complex can be isolated.The catalytic performances of the synthesized complexes were evaluated under environmentally friendly green chemistry conditions.All the examined compounds exhibited remarkable catalytic capabilities for cyclooctene conversion and selectivity toward epoxide production.Moreover, dinuclear compounds demonstrated a superior rate of conversion to catalytically active species compared to polynuclear catalysts.Conversely, mononuclear catalysts derived from (iso)nicotinic hydrazide displayed a higher propensity for conversion into the pentacoordinated active species compared to complexes derived from aminobenzohydrazides. Future investigations will focus on the catalytic oxidations of bio-derived substrates, particularly emphasizing the exploration of dinuclear catalysts.

Figure 1 .
Figure 1.Molecular structure of (a) H 2 L 3 , (b) 3a, (c) 3, with atoms represented as thermal ellipsoids at 50% probability level.Structural overlay of mononuclear complex 3a (green) and complex dimer 3 (purple) is shown in (d), showing the similarity of ligand conformation in both cases.

Figure 2 .
Figure 2. Kinetic profiles of cyclooctene conversion with catalysts (a) 1-4 and (b) 1a-4a.The blue curve presents the reaction with catalysts 1 (left) and 1a (right), the orange curve presents the reaction with catalysts 2 (left) and 2a (right), the grey curve presents the reaction with catalysts 3 (left) and 3a (right), and the yellow curve presents the reaction with catalysts 4 (left) and 4a (right).TBHP in water was used as an oxidizing agent.

Figure 2 .
Figure 2. Kinetic profiles of cyclooctene conversion with catalysts (a) 1-4 and (b) 1a-4a.The blue curve presents the reaction with catalysts 1 (left) and 1a (right), the orange curve presents the reaction with catalysts 2 (left) and 2a (right), the grey curve presents the reaction with catalysts 3 (left) and 3a (right), and the yellow curve presents the reaction with catalysts 4 (left) and 4a (right).TBHP in water was used as an oxidizing agent.

Scheme 2 .
Scheme 2. Presentation of the postulated mechanism for the epoxidation olefin.The ONO atoms (blue) and lines symbolize the ligands.The different steps (I-IV) are indicated by numbers close to the arrows.

Scheme 2 .
Scheme 2. Presentation of the postulated mechanism for the epoxidation olefin.The ONO atoms (blue) and lines symbolize the ligands.The different steps (I-IV) are indicated by numbers close to the arrows.

Scheme 3 .
Scheme 3. Representation of the MoO 2 L complex with a coordination sphere, highlighting the considered atoms taken into consideration for bond lengths (bold black) and for dihedral angles (red).The O β is shown in Scheme 2.
Elemental analyses were provided by the Analytical Services Laboratory of the Ru der Bošković Institute, Zagreb.NMR data are given in Tables S9-S12 and FiguresS5-S8in the Supplementary Materials.All reagents and solvents were commercially available (Alpha-Aesar (Karlsruhe, Germany), Sigma-Aldrich (Chemie GmbH, München, Germany)) and were used without further purification.[MoO 2 (acac) 2 ] was prepared according to the published procedure[48].
C NMR spectra of H 2 L 1 -H 2 L 4 , the py and phenyl hydrazone moiety signals C5-C10 were observed in the range of 151.90-121.06ppm and 152.03-112.09ppm, respectively.The signals in the range of 164.59-160.94ppm and 148.07-145.95ppm were attributed to C4 of the C=O group and C1 carbon of the CH=N group, respectively.

Table 2 .
Comparison of the catalytic parameters (conversion, selectivity, TOF 20min , and TON) or the previously published catalysts.
Scheme 3. Representation of the MoO2L complex with a coordination sphere, highlighting the considered atoms taken into consideration for bond lengths (bold black) and for dihedral angles (red).The Oβ is shown in Scheme 2.