Amide functionalized aminobisphenolato MoO 2 and WO 2 complexes: Synthesis, characterization, and alkene epoxidation catalysis

The use of dioxidomolybdenum( VI ) and -tungsten( VI ) complexes supported by a variety of structurally different tri-and tetradentate aminobisphenolato ligands as pre-catalysts in the epoxidation of alkenes is well established. However, under the widely used standard 1 mol% catalyst loadings these types of complexes generally show modest activity only. Recently, amide functionalities in the ligand design of various aminomonophenolato MoO 2 complexes have been shown to lead to heightened catalytic activity in alkene epoxidation. In this paper we show that similar ligand amide functionalization can lead to significant enhancement in the alkene epoxidation activity of aminobisphenolato MoO 2 complexes. Although the W variants showed much lower performance in comparison, the epoxidation activity of the Mo congeners is generally ca. two orders of magnitude higher than previously reported for structurally related aminobisphenolato complexes. An interesting phenomenon dubbed as “ dilution effect ” was discovered, wherein pre-catalyst loadings as low as 0.01 mol% may be realized without significantly reduced impact in activity. Moreover, the [pre-catalyst]:[oxidant] molar ratio – an often overlooked reaction parameter in the literature – was found to be critical for optimal catalytic performance.


Introduction
High-valent oxidomolybdenum(VI) and -tungsten(VI) centers are found in a variety of metalloenzymes that take part in oxygen atom transfer (OAT) reactions, in which oxygen is transferred from or to a suitable donor/acceptor molecule [1,2].In light of the oxygen activating nature of the M(VI)=O (M = Mo, W) moieties, many metal-organic model compounds have been studied in oxidation catalysis relevant to fine and bulk chemical industries, and the research is on-going.Epoxidation is a particularly attractive oxidative transformation due to availability of starting materials and high degree of functionalizability of the products.Oxidation of alkenes is one of the primary approaches to obtaining epoxides and, on one hand, allows capitalization of the wide variety of readily available olefinic feedstock chemicals.On the other hand, the synthetic versatility and thus importance of the epoxide functional group is highlighted by the fact that they readily undergo stereospecific ring-opening reactions in the presence of various nucleophiles, providing 1,2-difunctional compounds useful as intermediates in the synthesis of fine chemicals [3,4].
Epoxides are obtained from alkenes in a number of ways.In a laboratory setting, organic peracids such as m-chloroperbenzoic acid are still frequently used as stoichiometric epoxidation reagents [4].Transition metal-catalyzed epoxidation in the presence of organic hydroperoxides is in turn often utilized industrially.For example, molybdenum complexes such as [Mo(CO) 6 ] have been exploited as catalysts in the industrial Halcon/ARCO hydroperoxide process, which provides chemically valuable propylene oxide [5].To attain high reactivity in various oxidation reactions such as epoxidation, the electronic and steric features of ligands are generally used to modulate the reactivity of metal centers.In this regard, aminebisphenols are a large family of versatile multidentate proligands, because they generally form air and moisture stable complexes with virtually all transition metals and are Abbreviations: ARCO, Atlantic Richfield Company; Dipp, 2,6-di-isopropylphenyl; GC-MS, gas chromatography mass spectrometry; HB, hydrogen bond or hydrogen bonding-; IR, infrared (spectroscopy); NMR, nuclear magnetic resonance (spectroscopy); TBHP, tert-butyl hydroperoxide; TBHP(aq), 80 w-% (ca.8.0 M) aqueous tert-butyl hydroperoxide; TBHP(dec), ca.5.5 M tert-butyl hydroperoxide in decane; TOF, turnover frequency TON = turnover number.relatively easy to synthesize and functionalize [6].Specifically, aminobisphenolato ligands featuring an O 3 N/O 2 N 2 donor sets are interesting in a bioinorganic sense, as they structurally resemble some biological ligands found e.g., in vanadium-dependent haloperoxidases [7,8].
Recently, dioxidomolybdenum(VI) complexes 1-7 (Scheme 1) supported by hydrogen-bond (H-bond, HB) donating capable amide functionalized aminomonophenolato ligands have been reported to be exceedingly active epoxidation pre-catalysts, reaching TONs up to 110,000 at 0.5 ppm catalyst loading relative to some substrates [9].It has been suggested that HB donors such as amides in the ligand backbone may help to stabilize reaction intermediates during epoxidation, manifesting as enhanced activity of the catalysts [9].Inspired by these results, we were interested to learn whether or not amide functionalization of aminobisphenolato ligands leads to similar enhancement in epoxidation activity in respective dioxidomolybdenum(VI) complexes, due to obvious structural similarities between these ligand platforms (Scheme 1).To this end, five new aminobisphenolato ligands featuring amide pendant arms and their corresponding dioxidomolybdenum(VI) and -tungsten(VI) complexes Mo1-Mo5 and W1-W4 were designed, synthesized, characterized, and evaluated as pre-catalysts in alkene epoxidation.
The results from these investigations reveal that HB donating amide pendant arms endow Mo1-Mo5 and W1-W4 with multiple attractive properties.Similarly to 1-7, the catalyst loading of Mo1-Mo4 (Mo5 could not be evaluated in catalysis due to poor solubility) may be lowered to 0.01 mol-% relative to substrate, with no significant detrimental impact to their activity in the epoxidation of cis-cyclooctene.However, reduced performance was observed with more challenging alkene substrates, and the W-variants W1-W4 generally performed markedly worse compared to their Mo counterparts across all tested alkene substates.The molar ratio between Mo pre-catalysts and oxidant ([Mo]: [TBHP], where TBHP = tert-butylhydroperoxide) -a frequently overlooked reaction parameter in the literaturewas systematically investigated and found to be crucial for optimal catalytic performance.Additionally, all complexes readily crystallize, with single-crystals suitable for XRD obtained as fast as in 15 min after the introduction of ligand and metal precursors in methanol.This synthetic utility is attributed to the bulky, amphiphilic nature of the ligands, as well as to the propensity of the complexes to participate in intermolecular Hbonding via the amide pendant arms, as observed in the crystal structures (Supporting information).It is anticipated that these properties may be readily exploited when synthesizing complexes of other (transition) metals as well using the ligands described herein.

Catalytic experiments
The preliminary alkene epoxidation reactions of all complexes were performed using S1 as the benchmark substrate in standard conditions i. e., reactions were performed in chloroform with a 24 h reaction time, 50 • C reaction temperature and using two equivalents of aqueous TBHP (80% TBHP(aq), ca.8.0 M solution in 3:2 di-tert-butylperoxide:water) relative to substrate.The conversion of the reactions was evaluated by 1 H NMR spectroscopy after the 24-hour period.For 1 and 5 mol-% experiments 1-6 mg of catalyst was dissolved in 1 mL CDCl 3 and treated with appropriate amount of substrate and TBHP(aq).For lower loadings S1 solution containing 20 µL mL -1 , corresponding to a molar concentration of ca.0.153 M was prepared in CDCl 3 .Similarly, catalyst concentrations of 1.53 × 10 -4 to 1.53 × 10 -6 M i.e., 0.1-0.001mol-% relative to substrate were used.The catalyst solutions were prepared in CHCl 3 .In a typical run, 500 µL substrate solution and 500 µL catalyst solutions were combined in a screw-capped scintillation vial.The solutions were then treated with ca.19.2 µL TBHP(aq) solutions and then maintained at 50 • C for 24 h.
For reaction progress monitoring, S1 oxidation was followed by insitu 1 H NMR spectroscopy at 50 • C. The procedure for these experiments was similar than described above for the 24-hour reactions, with the exception that the reactant solutions were combined directly in mm o.d.NMR tubes, and that the 0.153 M substrate CDCl 3 solutions additionally contained 0.078 M 1,2-dichloroethane (1,2-DCE) as the internal standard.Due to the very low solubility of the W complexes, their catalytic activity could not be reliably assessed using similar stock solution approach.For S1 oxidation catalyzed by W1-W4, as well as S2-S5 oxidation catalyzed by Mo1-Mo4, a Heidolph Parallel Synthesizer 1 was used.In a typical experiment, the respective amount of catalyst (usually 2-3 mg of complex for 1 mol-% catalyst loading) was suspended in 0.5 mL of the respective solvent (CHCl 3 or 1,2-DCE) in a mL reactor equipped with a magnetic stir-bar and mixed with the substrate, 50 µL of mesitylene (internal standard) and heated to the respective reaction temperature (50 or 80 • C).Then the oxidant (usually 2 or 3 equiv.TBHP(dec), ca.5.5 M n-decane solution, or H 2 O 2 with respect to substrate) was added.Aliquots for GC-MS (20 μL) were withdrawn at given time intervals, quenched with MnO 2 and diluted with ethyl acetate.The reaction products were analyzed by GC-MS (Agilent Technologies 7890 GC System), and the epoxide produced from each reaction mixture was quantified vs. mesitylene as the internal standard (uncertainty is ±5%).

Alkene epoxidation catalysis
For complexes 1-7, excellent catalytic results were obtained with complexes featuring tert-butyl substituents in the phenolato moiety and either N-tert-butyl-or N-phenylacetamide pendant arms [9].As such, N-phenyl-and N-tert-butylacetamide pendant arms were opted for all studied complexes (Mo1, W1 and Mo4, W4, respectively).Additionally, N-benzylacetamide and N-methyl-N-phenylacetamide substituted complexes (Mo3, W3 and Mo2, W2, respectively) were also designed.In Mo2/ W2, which act as control pre-catalysts, any HB donating effects are blocked by the methylated amide group, whereas the benzyl group in Mo3/W3 acts as a "dangling" group, possibly interfering with any HB donating effects.Mo5 featuring a severely hindered 2,6-di-isopropylphenylacetamide pendant arm was also synthesized but could not be evaluated in catalysis due to poor solubility (Supporting Information).
The catalytic properties of all complexes were evaluated in the epoxidation of S1-S5 (Scheme 2).First, the activity of the complexes was assessed in the epoxidation of the benchmark substrate cis-cyclooctene (S1) [10].Under standard conditions all complexes are active at 1 mol-% catalyst loadings (Fig. 1).While Mo1-Mo4 reach practically full conversion and selectivity, complexes W1-W4 are markedly inferior, reaching only modest conversions with rather poor selectivity for the epoxide, a phenomenon frequently observed in other WO 2 systems including organometallic cylopentadienyl complexes [11][12][13][14][15] as well as octahedral dioxido complexes [16][17][18][19].It should be noted that some W complexes, such as [WCl 2 (O) 2 (OPMePh 2 ) 2 ], reported by Luck and Jimtaisong, have been reported to be more active in epoxidation compared to the Mo analog when H 2 O 2 is used as the terminal oxidant [17].In fact, some W complexes show superior activity relative to their Mo congeners in reactions utilizing H 2 O 2 as the terminal oxidant, whereas Mo complexes are often superior when TBHP is used as the oxidant [14].In a study by Herrmann and co-workers, dioxido Mo/W complexes based on chiral 2 ′ -pyridinyl alcoholate ligands show identical enantiocontrol in asymmetric epoxidation, although the W variants are slightly slower than their Mo analogs [18].This study highlights a notion that W complexes are inferior epoxidation catalysts relative to Mo due to kinetic factors primarily, but operate via a similar catalytic mechanism [18].
Mo1-Mo4 have a very high initial reaction rate after a variable induction period lasting from about 15 min to two hours, reaching full conversion at the 16 h mark (Fig. 1).The W-variants show similarly rapid initial activity, attaining conversions of 20-35% during the initial three hours of reaction (Fig. 1).However, unlike with Mo1-Mo4, the epoxide yield is lowered afterwards, most likely due to over oxidation and hydrolysis to diol [20], or associated ring-opening reactions [11,21].From the Fig. 1 it may be concluded that Mo1-Mo4 as well as W1-W4 collectively show rather similar activity to one another, and that no significant SAR (structure-activity relationship) effects can be obviously drawn at 1 mol-% catalyst loadings.
Amide functionalized dioxidomolybdenum(VI) complexes 1-7 retained their epoxidation activity at extremely low pre-catalyst loadings approaching 5 ppm (0.0005 mol-%) relative to S1 [9].Other, more challenging substrates S3-S5 were also efficiently converted to the corresponding epoxide at low 0.001 mol-% pre-catalyst loadings.Moreover, the complexes remained active in protic solvents such as alcohols and when using H 2 O 2 as the terminal oxidant.As such, due to obvious structural similarities between 1, 7 and Mo1-Mo4/W1-W4, low catalyst loadings were tested for Mo1-Mo4 and W1-W4 as well (Fig. 2).Remarkably, for Mo1-Mo4 catalyst loadings as low as 0.01 mol-% can be realized in S1 oxidation, with moderate activity still present.In terms of TONs and TOFs, best results are obtained at 0.01 mol-% catalyst loadings: TONs of 5300, 1500, 3500 and 3700, and maximal TOFs 920, 210, 600 and 240 h -1 are obtained, in order, for Mo1-Mo4, respectively.These figures are one to two orders of magnitude higher than generally reported for aminobisphenolato MoO 2 complexes [22][23][24][25][26][27][28].At 0.1 mol-% loading Mo1 and Mo4 reach nearly quantitative conversion in 16 h, whereas Mo2 and Mo3 reach 54% and 68% conversions, respectively.Even down to 0.01 mol-%, Mo1, Mo3 and Mo4 still reach moderate conversions up to 52%.At 0.1 mol-% the kinetic curves of the reactions follow typical 1st order-like behavior, whereas at 0.01 mol-%, except for Mo1, the kinetics start to become linear, conceivably due to saturation of the catalysts.It is noteworthy that with Mo1 and Mo4, the conversion vs. time profiles are affected only barely by lowering the catalyst loadings from 1 to 0.1 mol-%.Interestingly, Mo2, having a blocked HB donating capability, shows least activity from all tested Mo complexes, although not significantly so.This may be interpreted as the HB donating effect having a beneficial effect on catalysis.In contrast, W1-W4 are completely inactive below 1 mol-% loadings (data not shown).For this reason, follow-up catalytic investigations were performed for Mo1-Mo4 only.
To evaluate the wider epoxidation potential of Mo1-Mo4 beyond S1, more challenging substrates styrene (S2), 1-octene (S3), racemic limonene (S4) and α-terpineol (S5) were tested (Table 1 entries 1-4).S2, although typically relatively easily oxidized, may give a number of products aside from styrene oxide such as over oxidation products phenylacetaldehyde, styrene glycol, benzaldehyde as well as benzoic acid [29].Thus, S2 is a good indicator of selectivity of a catalyst, although reaction conditions may also play a role in this regard.Linear terminal alkenes such as S3 are good indicators of overall epoxidation power, as they are much more difficult to epoxidize in comparison to internal alkenes, and especially cyclic internal alkenes such as S1, for which ring strain is also a factor [30,31].Terpene substrates such as S4 and S5 are also difficult to epoxidize [10].While both offer similar difficulty in terms of oxidizability, S4, having two olefinic sites, gives indications of the regioselectivity of a catalyst, while S5 gives a measure of chemoselectivity, as well as functional group tolerance of catalysts.
Catalytic data for Mo1-Mo4 are given in Table 1 entries 1-4.For substrate S2 as well as S5, complexes Mo1-Mo4 are unselective towards the corresponding epoxide.In the case of S2, a nearly full conversion is obtained, with benzaldehyde being formed solely in all cases, indicating Scheme 2. General reaction scheme of epoxidations catalyzed by Mo1-Mo4 and W1-W4 tested with the substrates S1-S5 as shown.

P. Salonen et al.
over-oxidation of the epoxide [29].In contrast, with S5, there is no conversion at all with any of the studied complexes, highlighting the challenging nature of this substrate.It is also possible that the pre-catalysts are poisoned by the alcohol functional group in S5, since other, simpler alcohols, were observed to readily inhibit epoxidation of S1 (see Table S4).Oxidation of S4, on the other hand, affords the corresponding limonene oxide in high yield (66-89%) and decent to moderate selectivity (36-78%), although Mo4 seems to be inactive.All complexes show comparable activity in the epoxidation of S3, reaching 27-70% conversion and 83-95% selectivity.Higher conversion of S3 is obtained by lengthening the reaction time to 48 h.Interestingly, and as observed solely in the case of S1 so far, the catalytic potential of Mo1-Mo4 is seemingly increased with successive dilution: At 0.1 mol-% S3 is still epoxidized by Mo1, Mo3 and, notably, with Mo4, which showed no activity at 1 mol-%.No conversion could be observed at 0.01 mol-% catalyst loadings with any of the pre-catalysts, however.
To summarize: Mo1-Mo4 show very high catalytic activity for epoxidation of benchmark S1 within 24 h in chloroform, especially below the widely used literature standard 1 mol-% catalyst loading.However, catalytic activity is quenched in alcohols and MeCN, whereas less polar and non-coordinating DCM and 1,2-DCE afford lower, but still comparable results to the optimal solvent, chloroform.Non-polar solvents n-hexane and toluene generally afford poor results.The tungsten variants W1-W4 had a significantly inferior performance in all epoxidations.In general, the epoxidations followed the trends reported in the literature [9,25,26,28,[32][33][34][35] namely that S1 gives the oxide in high yield and selectively, whereas the oxidation of S2 gives a good conversion with poor selectivity to epoxide, and oxidation of terminal alkene S3 is selective but slow (poor conversion).However, the oxidation of S4 was achieved with moderate conversion and selectivity using Mo1-Mo3, but not Mo4.S5 did not react with any of the pre-catalysts, indicating that the OH group of the substrate may deactivate the complexes.
All reactions worked equally well with TBHP(aq) and TBHP(dec), indicating that the small water content in TBHP(aq), which however becomes significant relative to active catalysts at very low pre-catalyst loadings, did not seem to have a detrimental effect on the catalysis overall.A possible explanation for this behavior is the highly lipophilic coordination environment near the metal centers, imparted by the di-tert-butyl phenolato moieties, which may hinder or prevent the approach and subsequent coordination of water to the metal centers.
Similarly, the coordination of H 2 O 2 is equally hindered, which manifests as inactivity of the pre-catalysts when H 2 O 2 is used as the oxidant.However, alcohols and TBHP have comparatively more amphiphilic character than H 2 O/H 2 O 2 .Thus, TBHP acts as a competent oxidant in non-coordinating solvents, whereas alcohols such as MeOH and EtOH effectively poison the pre-catalysts, regardless of used oxidant.Moreover, in reactions that are performed in chloroform and in the presence of TBHP(aq), a biphasic system will form to a certain extent, because water is not miscible with chloroform.In other words, the poisoning effects of water towards catalysis in chloroform are mitigated because it remains trapped within the aqueous phase, whereas all complexes, substrates and TBHP are effectively confined in the lipophilic phase.This also explains why TBHP(dec) shows similar performance relative to TBHP(aq) as an oxidant: As far as the pre-catalysts, oxidant and substrate are concerned, the lipophilic reaction environment is likely similar using both oxidants.By extension, poor performance of aqueous H 2 O 2 may be explained by considering that the opposite is also true: H 2 O 2 is more soluble in water than in chloroform, and thus unable to act as competent oxidant.In other tested reaction solvents that show amphiphilic character, such as alcohols, MeCN, etc.; H 2 O 2 and TBHP are both effectively outcompeted by the relatively large excess of coordinating solvents.The sensitivity of the catalytic activity to solvent medium delivering the oxidant molecule to the metal center has been noted earlier e.g., for dinuclear Mo/W complexes of the type [(Cp*) 2 M 2 O 5 ] (Cp* = pentamethylcyclopentadienyl, M = Mo, W) as well [14].In this report by Poli and co-workers, the W variants showed superior activity over the Mo congeners when aqueous H 2 O 2 was used as the oxidant.In contrast, the Mo version showed best performance using TBHP(dec), whereas the water content in TBHP(aq) hindered their catalytic performance [14].

Dilution effect
It has been demonstrated that all epoxidation experiments conducted using Mo1-Mo4 at a first glance show a rather unusual phenomenon we have dubbed as the "dilution effect" wherein the performance (both relatively and absolutely speaking) of the pre-catalysts is seemingly enhanced, in terms of better TOF values, with the lowering of the catalyst loading (Fig. 3).This behavior is not anticipated for completely stable and soluble pre-catalysts, for which the epoxidation rate is expected to be 1st order in catalyst concentration [36].The phenomenon would also suggest that only a minor proportion of the complexes are active at any given time during catalysis, as also suggested by the groups of Kühn and Poli in their studies involving complexes based on cyclopentadienyl and Schiff base ligands [11][12][13]34,37,38].
The effect of the [Mo]:[TBHP] relationship was tested by conducting the S1 oxidations with Mo1-Mo4 using a variable molar ratio between [Mo]:[TBHP], additionally by using an increased 5 mol-% catalyst loadings relative to S1.It was expected that an increase in catalyst loading to 5 mol-% would lead to reduced catalytic performance if TBHP loadings were not changed accordingly.Indeed, raising the catalyst loading to 5 mol-% relative to S1 without appropriately modifying [TBHP] molar ratio to 5:100:1000, thus affording a [Mo]:[TBHP] molar ratio of 1:200, i.e. the same as in the standard 1 mol-% experiments, almost identical behavior between the 1 and 5 mol-% experiments is observed (Fig. 3).Although the [substrate]:[oxidant] molar ratio in these experiments is changed to 1:10, this was found to have a negligible    effect to the non-catalytic autoxidation of S1 compared to the standard 1:2 used in the other experiments (Fig. 3).Similar results were observed for the other pre-catalysts Mo2-Mo4 (Supporting Information Figs.S84-S87).These experiments reveal that the molar ratio between the precatalysts and TBHP is extremely important.Using a 5 mol-% precatalyst loading relative to substrate, while not intrinsically detrimental to catalysis, must be accompanied by an appropriate increase in oxidant loading so that a [Mo]:[TBHP] molar ratio of at least 1:200 is reached.Similarly, the seemingly elevated activity of all pre-catalysts at 0.1 and 0.01 mol-% catalyst loadings simply originates from the very large excess of TBHP relative to Mo pre-catalysts ([Mo]:[TBHP] molar ratio as low as 1:20,000) in these experiments, which strongly drives the formation of the active catalysts.In this way, thanks to the large difference in [Mo]:[TBHP] molar ratios, the absolute amount of activated catalyst is higher in the 0.1 mol-% experiments than in the 1 or 5 mol-% experiments.

Induction period
In addition to the dilution effect, a variable induction period prior to reaction onset was observed in many of the epoxidation experiments conducted using Mo1-Mo4 (cf.Figs. 1 and 2).The behavior of Mo1 and Mo4 i.e., the pre-catalysts showing the longest induction periods, was further examined in the presence of excess TBHP(aq) (Figs. 4 and S92).As highlighted by Mo1 (Fig. 4) these experiments demonstrate that the length of the induction period may be removed entirely if the precatalysts are incubated in the presence of excess TBHP(aq) prior to addition S1.If, on the other hand, the pre-catalysts are incubated in the presence of S1 for 3-6 h prior to addition of TBHP(aq), the induction period persists.The experiments above also demonstrate that the incubation of the pre-catalysts with TBHP has a tangible effect on catalysis, even when the 1 H NMR spectra of the complexes during catalysis remain  unchanged.This behavior may be indicative of a reversible equilibrium between the catalytic resting state (the species observed by NMR, corresponding to the pre-catalyst themselves) and the active forms, as suggested by Kühn [39] and Poli [34,37].

Comparison of catalysis
In the present study it has been established that Mo1-Mo4 invariably show high catalytic activity in the epoxidation of S1 at sub 1 mol-% catalyst loadings, akin to 1-7 that inspired us to prepare Mo1-Mo4 in the first place.The slightly inferior performance of the HB donor blocked Mo2 relative to the other tested complexes could hint that Mo1, Mo3 and Mo4 may benefit from the presence of the amide NH moiety.However, these SAR effects are only clearly visible in the epoxidation of S1, and in fact, Mo1-Mo4 collectively behave rather similarly across S2-S5 (Table 1 entries 1-4).
The catalytic prowess of Mo1-Mo4 also quite closely match that of 12-14, MoO 2 aminobisphenolato complexes featuring 2-methoxyethyl, N,N-dimethylethyl and methylpyridyl pendant arms, respectively, reported by us earlier [25].These pre-catalysts show rather similar activity with S3-S5 than Mo1-Mo4 (Table 1 entries [12][13][14].For instance, S3 and S4 are epoxidized with comparable conversions and selectivities, and 12-14 are deactivated with S5, similarly to Mo1-Mo4.However, Mo1-Mo4 are much more active in the epoxidation of S1, whereas S2 is more efficiently epoxidized with 12-14.It should be emphasized that 12-14 have not been evaluated in alkene epoxidation at a catalyst loading below 1 mol-%, and so comparisons at these loadings cannot adequately be done. In another study, S2 was catalytically epoxidized in the presence of MoO 2 aminobisphenolato complexes featuring benzimidazole (11) and ethylpyridyl (15) based pendant arms, respectively [22,24].There are two interesting features regarding these reported catalytic experiments.Firstly, the catalytic reactions do not benefit at all by raising the catalyst loading from 2.5 mol-% to 5.0 mol-% in terms of conversion (Table 1 entries 11, 15 and 16) [22,24].This phenomenon was observed across all studied pre-catalysts, and it has been attributed to poor solubility of the catalysts [22].However, it seems more likely that 15 is instead subject to the same "dilution effect" we encountered with Mo1-Mo4 in this study.Namely, the catalytic experiments performed using 15 and other pre-catalysts in the study have a [Mo]:[TBHP] molar ratio difference of 1:50 at the lowest, and not exceeding 1:125.Given our findings in this report, increasing the molar ratio difference to a factor of no less than 1:200, and perhaps optimally 1:2000, might induce a significant enhancement in activity of the complexes, without affecting the autoxidation of S2.Unfortunately, this was not tested with 15 [22].Similar behavior was observed for MoO 2 complexes supported by aminobisphenolato ligands featuring methylpyridyl pendant arms in the same paper [22].
In any event, Mo1-Mo4 perform significantly weaker when compared to 1-7 (Table 1 entries 5-7).For the monomeric 1-3, the catalyst loading may be lowered to 0.001 mol-% with respect to S1, S2, S4 and S5, with highest TONs in the range of ca.75,000 (for S1 and 2), 28,000 (for S2 and 3), 47,000 (for S4 and 3) and 25,000 (for S5 and 2) [9].Thus, 1-7 may be subject to the same dilution effect as Mo1-Mo4, but remain active even with challenging substrates at a very low catalyst loading, unlike Mo1-Mo4.Hydrogen bonding capable motifs in the ligand design have been implicated earlier to be able to enhance the epoxidation activities of various dioxidomolybdenum(VI) complexes bearing ligands featuring alcoholate pendant arms [28,40] as well as amide pendant arms [9].It is suggested that such moieties are able to stabilize incoming oxidant molecules, such as TBHP or H 2 O 2 , via hydrogen bonding, and/or aid in hydrogen atom shuttling, activating the oxidant and effecting oxygen transfer to the substrate [9,26].These reports essentially invoke a mechanism wherein oxidant activation is at least partly ligand-assisted, unlike the generally accepted mechanism(s), where oxidant activation and eventual substrate oxidation are effected by H-bonding and shuttling interactions between oxidant molecules and terminal metal oxido ligands [34,37,39,[41][42][43].However, the extremely high catalytic activity of 1-7 may not entirely be governed by the presence of an NH containing amide functionality.In fact, John and co-workers have recently evaluated several MoO 2 complexes featuring tri-and tetradentate aminobisphenolato ligands in deoxydehydration of styrene glycol [44].Their studies reveal that heightened catalytic activity is observed with coordinatively unsaturated complexes featuring tridentate ligands, as opposed to coordinatively saturated complexes containing tetradentate ones [44].Similarly, enhanced activity may be observed by removing a phenolato arm from the ligand backbone, effectively leading to dinuclear complexes essentially similar to 4-7 [44].Thus, Mo1-Mo4 may show reduced performance relative to 1-7 due to steric factors imparted by the bulky tetradentate aminobisphenolato ligands.

Summary and conclusions
Nine benchtop stable dioxidomolybdenum(VI) and -tungsten(VI) complexes Mo1-Mo5 and W1-W4 supported by new dianionic tetradentate amide functionalized aminobisphenolato ligands were designed, synthesized, and characterized.All complexes were found to crystallize very easily in methanol, possibly thanks to intermolecular HB effects from the neutral amide moieties.These properties may be readily exploited when synthesizing complexes based on other transition metals as well.
All complexes except Mo5 were evaluated in the epoxidation of selected alkene substrates.The catalytic activity of Mo1-Mo4 in the epoxidation of S1 was found to be very high, and catalyst loadings as low as 0.01 mol-% could be realized without significant loss of activity, similar to 1-7, reported in the literature.Dubbed the "dilution effect", the high activity of Mo1-Mo4 below 1 mol-% loadings was experimentally shown to be result of very high [Mo]:[TBHP] molar ratio differences (as large as 1:20,000), facilitating formation of active catalysts by the action of TBHP.Interaction of the pre-catalysts with TBHP was also shown to account for induction periods seen for some of the tested complexes.Mo1-Mo4 are one to two orders of magnitude more active in the epoxidation of S1 than other MoO 2 aminobisphenolato supported complexes reported in the literature.However, the activity of Mo1-Mo4 with more challenging alkene substrates S2-S5 is similar based on comparison.Likewise, the tungsten derivatives W1-W4 showed typically poor behavior in all epoxidations, in accordance with the literature.Aside from any possible benefits of having an amide NH moiety in the ligand design, this study also reveals that ligand sterics might be more important in determining the catalytic activity of MoO 2 aminobisphenolato complexes in epoxidation.This proposition is also in line with the work of Poli et al. who suspect that five-coordinate systems show elevated catalytic activity relative to six-coordinate systems [34].
Work has already begun in our laboratories to further probe the P. Salonen et al.
effects of the "dilution effect", and to see whether other aminobisphenolato supported MoO 2 complexes reported by us earlier also display heightened catalytic activity at very low [Mo]:[TBHP] ratios.Specifically, the impact of any HB effects and/or ligand sterics in tri-and tetradentate aminobisphenolato ligands will be focused upon and studied experimentally and computationally in the future.

Table 1
Comparison of substrate S1-S5 conversion (selectivity for epoxide in parentheses) catalyzed by Mo1-Mo4 and various other aminophenolato supported MoO 2 complexes reported in the literature.The coordinating atoms and ligands are colored.