Coordinatively Unsaturated Metallates of Cobalt(II), Nickel(II), and Zinc(II) Guarded by a Rigid and Narrow Void

Both natural enzymatic systems and synthetic porous material catalysts utilize well-defined and uniform channels to dictate reaction selectivities on the basis of size or shape. Mimicry of this design element in homogeneous systems is generally difficult owing to the flexibility inherent in most small molecular species. Herein, we report the synthesis of a tripodal ligand scaffold that orients a narrow and rigid cavity atop accessible metal coordination space. The permanent void is formed through a macrocyclization reaction whereby the 3,5-dihydroxyphenyl arms are covalently linked through methylene bridges. Deprotonative metallation leads to anionic and coordinatively unsaturated complexes of divalent cobalt, nickel, and zinc. An analogous series of trigonal monopyramidal complexes bearing a nonmacrocyclized variant of the tripodal ligand are also reported. Physical characterization of the coordination complexes has been carried out using multiple spectroscopic techniques (NMR, EPR, and UV–vis), cyclic voltammetry, and X-ray diffraction. Complexes of the macrocyclized [LOCH2O]3– ligand retain a rigid cavity upon metallation, with this cavity guarding the entrance to the open axial coordination site. Through a combination of spectroscopic and computational studies, it is shown that acetonitrile entry into the void is sterically precluded, disrupting anticipated coordination at the intracavity site.


■ INTRODUCTION
Synthetic inorganic chemistry has witnessed an increasing focus on the secondary coordination sphere as a means to control reaction outcomes at ligated metal centers. 1−5 Specific efforts are often guided by the finely tuned microenvironments designed by nature to govern substrate binding and product formation at metallocofactors. 3,6 Indeed, carefully tailored ligands incorporating H-bond donors/acceptors, 7−15 Lewis acids, 16−21 or electrostatic charges 22−25 are now often used to stabilize transition states or products as a means to alter reaction outcomes of metal complexes with small molecules.
In order to regulate the arrival of exogenous species and control substrate conformation, nature frequently buries metalloenzymes within protein structures, with access for exogenous substrates available only through well-defined channels. Mimicry of this property in synthetic complexes can be approached by housing metal-binding chelators within channels defined by macrocyclic motifs. 26−29 Provided that the channel serves as the sole point of ingress for substrates, metalbased reactivity becomes dependent on the kinetic feasibility of passage through the channel. Similar considerations guard reactivity within crystalline porous materials, as substrates must diffuse into a crystallite to access non-surface active sites. 30,31 Industrial zeolite catalysis often exploits this approach on enormous scales to realize selectivities on the basis of size/ shape. 32 In principle, such a design strategy could effect interesting selectivities in synthetic homogeneous catalytic schemes, although realization has often proven elusive using known systems. 33 Pioneering work from Breslow demonstrated that cyclodextrins appended with chelating motifs could localize a hydrophobic channel adjacent to a metal center. 26,34 Subsequent approaches in ligand design have expanded the diversity of macrocyclic systems−most frequently employing calixarenes and resorcinarenes−and have focused on spatially controlling the macrocycle's orientation with respect to the metal center (Figure 1a). 35−41 Macrocycles bearing judiciously designed chelating groups can result in the macrocycle interior serving as the sole access point for the enshrouded metal center (Figure 1b). However, synthetic routes to these systems can be laborious as the macrocycle must be synthesized, chemically modified, and then covalently linked to the metal binding motif. As an alternative approach, the groups of Lu 42 and Schrock 43 have demonstrated macrocyclization reactions upon tris(2-aminoethyl)amine (TREN)-based scaffolds to orient either a 45-membered ring or an arene-capped organic cavity above the metal binding pocket (Figure 1c). These procedures obviate the separate synthesis of a macrocyclic species that is subsequently decorated with chelating motifs and theoretically could enable the construction of a library of proligands with varying cavity dimensions, although no followup reports have appeared for either system.
We endeavored to develop a short and modular synthetic route to a ligand that orients a narrow and rigid channel above an accessible metal coordination site. Starting with a TRENderived tris-amide platform bearing 3,5-dihydroxyphenyl arms, we demonstrate the construction of a rigid and narrow macrocycle-defined channel that is oriented directly atop the metal-binding pocket (Figure 1d). This three-step procedure employs macrocyclization as the final step and evokes the rigidification of resorcinarenes to form cavitands by linking adjacent phenol groups through a methylene bridge. 44,45 Deprotonation and metallation with cobalt(II), nickel(II), and zinc(II) sources leads to anionic four-coordinate complexes containing an open coordination site that is housed inside the rigid cavity. An additional series of complexes bearing a ligand variant that lacks a rigid cavity is also presented for comparison. It is shown that the narrow channel profile in the macrocyclized ligand precludes acetonitrile binding to the open site in the cobalt(II) congener, while nitrile coordination readily occurs in the nonmacrocyclized cobalt(II) complex. Computational investigations reveal that intracavity acetonitrile binding is precluded by the narrow dimensions of the void.

■ RESULTS AND DISCUSSION
A tripodal tris-amide bearing 3,5-dihydroxyphenyl arms was targeted as a precursor to a macrocyclized proligand ( Figure  2a). Benzyl protection of the phenolic groups in 3,5dihydroxybenzoic acid allowed for smooth amide bond formation with TREN using 1,1′-carbonyldiimidazole (CDI) as a coupling mediator. 46 Deprotection of H 3 L OBn with 1,4-  cyclohexadiene (1,4-CHD) and catalytic Pd/C resulted in the formation of the desired H 3 L OH as a hygroscopic colorless solid. Taking inspiration from the rigidification of resorcinarenes to form cavitands, 44,45 it was endeavored to covalently tether adjacent aromatic rings to one another through methylene bridges. A variety of dihalomethane reagents (CH 2 BrCl, CH 2 Br 2 , CH 2 BrI, and CH 2 I 2 ) were found capable of producing the desired macrocyclized species H 3 L OCH2O upon heating with H 3 L OH in the presence of a Brønsted base, as assayed by 1 H NMR spectroscopy and electrospray ionization mass spectrometry ( Figure 2). It should be noted that macrocyclization must compete with oligomerization processes which link multiple tris-amide species together through a methylene unit. 47 Given that organic cavitand formation typically employs carbonate bases and N,N-dimethylformamide (DMF) as the solvent, 48,49 our optimization efforts surveyed a library of carbonate salts (Li 2 CO 3 , Na 2 CO 3 , K 2 CO 3 , Cs 2 CO 3 , Ag 2 CO 3 , and (NH 4 ) 2 CO 3 ) while also varying the dihalomethane reagent (see above) and its rate of addition, reaction temperature, and dilution. It was found that the slow addition of CH 2 BrI to H 3 L OH (12 mM) and K 2 CO 3 in DMF at room temperature, with subsequent heating to 60°C, produces H 3 L OCH2O in the highest consistent yields (Figure 2). By the removal of aliquots of known volume from reaction mixtures and analysis by 1 H NMR with an internal standard, it was determined that H 3 L OCH2O was produced in approximately 10% yield. Isolation of macrocyclized H 3 L OCH2O from these crude products is accomplished via Soxhlet extraction with CHCl 3 . Serial triturations of the obtained extractant with water, methanol, and hexanes yields spectroscopically pure H 3 L OCH2O as a colorless solid in average isolated yields of approximately 8%.
Single crystals of H 3 L OCH2O were obtained from a dilute methanol solution. Solid-state structure determination by X-ray diffraction confirms the formation of an 18-membered ring that defines the crest of a narrow cavity, which is bound by the faces of three aromatic rings (Figure 2b). The methylene bridges give rise to a pair of doublets in the 1 H NMR spectrum that correspond to diastereotopic H atoms bearing an endo or exo relationship to the cavity ( Figure S5). Although the solidstate structure depicts two distinct methylene orientations (two point upward atop the crest while the third is rotated downward), these bridges hinge rapidly and display equivalence by 1 H NMR spectroscopy at temperatures of as low as −80°C ( Figure S15). The covalent linkages defining the macrocycle endow H 3 L OCH2O with excellent thermal stability, as heating to 100°C in dry DMSO-d 6 for 1 week leaves H 3 L OCH2O unchanged. Notably, exposing H 3 L OCH2O to wet DMSO-d 6 at this temperature results in substantial decomposition over the course of several days ( Figure S16 Figure  S42). The primary coordination spheres of both crystallographically independent nickel complexes show a distortion As is common for TREN-derived LX 3 -type ligands, 58 ). These events remain electrochemically irreversible at scan rates of as fast as 10 V/s and do not appear in the voltammograms of [ZnL OCH2O ] − , substantiating their assignments as metal-based events ( Figure S29). The onset potentials of these waves compare well with those seen for the corresponding cobalt(II) and nickel(II) dimetallocryptands reported by Cummins and Nocera. 54,60 However, they are shifted anodically when compared to four-coordinate trisamidylamine cobalt(II) and nickel(II) complexes reported by others. 10 (Tables S11 and S14). Importantly, however, the calculated enthalpy and free energy of HCN binding exhibit only very minor differences between  The presence of a gate protecting the cavity interior may result in cavity penetration becoming the rate-limiting event in a bimolecular metal-based reaction. 29 While ligand-binding events at unsaturated metal centers often have small activation barriers associated with coordination sphere rearrangements, 69 the rate-limiting event for intracavity coordination could plausibly be penetration of the void by the incoming ligand. To explore this notion, relaxed surface scans were performed to investigate the energetics of MeCN and HCN ingress and egress through the narrow void (Figure 8, bottom). In both cases, a significant increase in ΔE (i.e., ΔH at 0 K) occurs as the nitrile progresses through the void, with maxima for both ligands occurring at Co−N nitrile distances outside of the van der Waals contact. Indeed, it seems intuitive that these distances would maximize steric pressures given both the slightly convex profile of the arene-defined void and the potential for clashing of the nitrile with the methylene groups lining the cavity crest. The methyl group of an intracavity-ligated MeCN must reside near the narrowest region of the void, and the attendant steric clashing offsets the stabilization afforded by a weak Co−N nitrile coordinative bond. However, HCN is apparently small enough that steric pressures subside at shorter Co−N nitrile distances to yield a net energetically favorable binding event. Note that attempts were made to synthesize [Co(NCH)L OCH2O ] − by exposing [CoL OCH2O ] − to HCN (Caution!Hydrogen cyanide is a highly toxic gas) but were hampered by competitive demetallation to yield H 3 L OCH2O along with unidentified cobalt-containing species.
The inaccessibility of the void within [CoL OCH2O ] − to MeCN likely reflects a high degree of rigidity and a low proclivity for the cavity to undergo guest-responsive conformational changes. In this way, the [L OCH2O ] 3− ligand differs from the calix [6]arene-bearing "funnel complexes" championed by Reinaud, which generally display guest-adaptive cavities. 29,70 Indeed, it is well known that calix [6]arenes display conformational flexibility enabled by rotations about the C sp 3 −C arene bonds. 71,72 In contrast, the macrocyclic motif within the [L OCH2O ] 3− ligand more closely approximates the structure of Cram's cavitands, 44 wherein each arene is linked to its neighboring rings through two points of attachment, imbuing the macrocycle with greatly enhanced rigidity and less conformational adaptability.

■ CONCLUSIONS
We have shown that macrocyclization of a TREN-derived proligand can engender a rigid and narrow void that is oriented on top of the metal binding pocket. Despite its rigidity,  Inorganic Chemistry pubs.acs.org/IC Article clashing of the nitrile with atoms near the cavity rim thermodynamically outweighs the formation of a weak M− N nitrile bond. Additionally, calculations demonstrate that the entry of even small-profile species into the void space must overcome a significant kinetic barrier that is likely to slow substrate ingress. We envision that the localization of reactive metal-bound motifs within the cavity of complexes bearing the [L OCH2O ] 3− ligand will allow for kinetic stabilization and characterization of metal-bound species that have traditionally proven elusive.
■ EXPERIMENTAL SECTION General Considerations. All manipulations were carried out under an atmosphere of purified dinitrogen using standard Schlenk and glovebox techniques. Unless otherwise stated, reagent-grade starting materials were purchased from commercial sources and either used as received or purified by standard procedures. 73 Unless otherwise stated, organic solvents were deoxygenated and dried using a Pure Process Technologies solvent purification system. N,N-Dimethylacetamide (DMA) and dimethyl sulfoxide-d 6 were stored over activated 3 Å molecular sieves, transferred via cannula into a separate flask, sparged with N 2 for 1 h, and stored over fresh 3 Å molecular sieves in the glovebox prior to use. Chloroform-d was stirred with CaH 2 , distilled into a separate flask, degassed via freeze− pump−thaw cycles, and stored over activated 3 Å molecular sieves in the glovebox prior to use. Methanol was neither deoxygenated nor dried and was used as received. Molecular sieves (3 Å) and Celite were separately preactivated in a 180°C oven overnight, transferred into a round-bottomed flask and heated under vacuum (P < 100 mTorr) at a temperature in excess of 200°C for at least 12 h, and then stored in the glovebox. Tetra(n-butylammonium) hexafluorophosphate for electrochemical measurements was recrystallized three times from ethanol and then dried under vacuum with P 2 O 5 at 120°C until the pressure reached 50 mTorr.
Solution 1 H and 13 C{ 1 H} nuclear magnetic resonance (NMR) spectra were recorded on a Bruker DPX-400 or DPX-500 spectrometer locked on the signal of deuterated solvents. 1 H and 13 C{ 1 H} chemical shifts are reported in parts per million relative to SiMe 4 ( 1 H and 13 C δ = 0.0 ppm) with reference to residual solvent resonances. Electronic absorption measurements were recorded using either a PerkinElmer Lambda 35 UV−vis spectrometer (ambienttemperature measurements) or an Agilent Cary 6000i UV−vis−NIR spectrometer (variable-temperature measurements). Samples were prepared in the glovebox and sealed in a quartz cuvette (1 cm or 1 mm path length). X-band electron paramagnetic resonance (EPR) measurements were carried out on a Bruker EMXplus spectrometer (microwave frequency of 9.382 GHz). Samples were prepared as solutions in CH 2 Cl 2 in a glovebox, glassed via flash-cooling in liquid nitrogen, and loaded into the spectrometer. Electrochemical measurements were performed using a CH Instruments 620 D potentiostat with a three-electrode setup, including a Ag/AgNO 3 (1 M) reference electrode (CHI111), a Pt wire counter electrode (CHI115, surface area in solution of 0.14 cm 2 ), and a glassy carbon working electrode (CHI104, 3 mm diameter). Solutions were prepared using dry and degassed CH 2 Cl 2 at a concentration of 1 mM metal complex. The [NBu 4 ]PF 6 electrolyte concentration was 0.1 M. Voltammograms were referenced to the Cp 2 Fe +/0 couple by using ferrocene as an internal standard. Solution-phase effective magnetic moments were determined using the Evans method. A dried solid analyte sample was dissolved in 0.800 mL of 9:1 v/v CH 2 Cl 2 /PhCF 3 . This solution was placed in a borosilicate NMR tube along with a flame-sealed glass capillary containing a 4:1 v/v CH 2 Cl 2 /PhCF 3 internal standard. Elemental analyses were performed on a PerkinElmer 2400 Series II Analyzer at the CENTC Elemental Analysis Facility, University of Rochester.
Preparation of Anhydrous Cobalt(II) Acetate. A flame-dried three-necked round-bottomed flask outfitted with a reflux condenser was evacuated and then backfilled with nitrogen a total of three times.
The flask was charged with acetic anhydride (76.0 mL, 805 mmol) and sparged with N 2 for 1 h. Cobalt(II) acetate tetrahydrate (20.0 g, 79.6 mmol) was then added to the flask under a N 2 purge, and the solution was heated to reflux for 12 h. After cooling, the solution was filtered using a Schlenk frit, and while an inert atmosphere was maintained, the light-pink solid was transferred into the glovebox and washed with 5 × 50 mL portions of Et 2 O. The solid was then dried under vacuum and was confirmed to be anhydrous via FT-IR spectroscopy (KBr pellet). Yield = 13.7 g, 77.4 mmol, 96%.
Preparation of Anhydrous Nickel(II) Acetate. A flame-dried three-necked round-bottomed flask outfitted with a reflux condenser was evacuated and then backfilled with nitrogen a total of three times. The flask was charged with acetic anhydride (76.0 mL, 805 mmol) and sparged with N 2 for 1 h. Nickel(II) acetate tetrahydrate (20.0 g, 79.7 mmol) was then added to the flask under a N 2 purge, and the solution was heated to reflux for 12 h. After cooling, the solution was filtered using a Schlenk frit, and while an inert atmosphere was maintained, the light-green solid was transferred into the glovebox and washed with 5 × 50 mL portions of Et 2 O. The solid was then dried under vacuum and was confirmed to be anhydrous via FT-IR spectroscopy (KBr pellet). Yield = 13.8 g, 78.1 mmol, 97%.
Synthesis of H 3 L OBn . Note: All water used in this preparation was 18.2 MΩ nanopure water. All glassware used in this preparation was soaked in a KOH/iPrOH bath, rinsed with water three times, and dried prior to use. Neglecting these precautions led to reaction mixtures that were darkened in color and resulted in dif f iculties purif ying the subsequently synthesized H 3 L OCH2O . We attribute these observations to the proclivity of TREN-derived tris-amides to act as very strong chelators of halides and small inorganic anions. 74 A 1 L three-necked roundbottom flask was charged with 3,5-di(benzyloxy)benzoic acid (39.0 g, 117 mmol, 3.3 equiv) and 300 mL of CH 2 Cl 2 . Solid 1,1′carbonyldiimidazole (18.9 g, 117 mmol, 3.3 equiv) was added slowly, and the reaction was then stirred for 1 h. After purging the headspace with N 2 to ensure the liberation of evolved CO 2 , tris(2-aminoethyl)amine (5.30 mL, 35.3 mmol) was then added in a single portion via syringe, and the reaction was allowed to stir for 12 h. The reaction mixture was then washed with water five times in a separatory funnel. The organic layer was dried under vacuum (P = 30 mTorr) to afford a colorless powder. Yield = 31.6 g, 29.0 mmol, 82%. 1 13  Synthesis of H 3 L OH . Note: All water used in this preparation was 18.2 MΩ nanopure water. All glassware used in this preparation was soaked in a KOH/iPrOH bath, rinsed with water three times, and dried prior to use. Neglecting these precautions led to reaction mixtures that were darkened in color and resulted in dif ficulties in purif ying the subsequently synthesized H 3 L OCH2O . We attribute these observations to the proclivity of TREN-derived tris-amides to act as very strong chelators of halides and small inorganic anions. 74 In a three-necked roundbottomed flask equipped with a reflux condenser and under an N 2 atmosphere, a solution of H 3 L OBn (22.6 g, 20.6 mmol) in methanol (550 mL) was added under an N 2 purge. Subsequently added under an N 2 purge were Pd/C (10 wt % Pd; 2.20 g, 2.07 mmol Pd, 0.1 equiv) and then 1,4-cyclohexadiene (19.6 mL, 207 mmol, 10 equiv). The reaction was then heated to 60°C for 16 h. The resulting black suspension was filtered, and the filtrate was then dried in vacuo to yield a hygroscopic colorless solid. Yield: 9.79 g, 17.6 mmol, 85%. 1

Synthesis of H 3 L OCH2O .
A three-necked 2 L round-bottomed flask (which had been previously soaked in a KOH/iPrOH bath, washed with nanopure water, and dried) was equipped with a 500 mL addition funnel, purged with N 2 for 30 min, and then charged with H 3 L OH (9.79 g, 17.7 mmol) and K 2 CO 3 (43.9 g, 318 mmol, 18 equiv). To the flask was added 1 L of dry and deoxygenated DMF via cannula transfer. The addition funnel was charged with a solution of CH 2 BrI (5.32 mL, 70.6 mmol, 4 equiv) in dry and deoxygenated DMF (500 mL) using a cannula. The CH 2 BrI solution was then added dropwise to the reaction mixture with stirring over the course of 13 h at 23°C and then stirred for an additional 9 h. Subsequently, the temperature was increased to 60°C and the reaction mixture was stirred for another 24 h. Upon cooling, the reaction mixture was concentrated under reduced pressure. The resulting pink/red residue was placed in a glass coarse fritted filter thimble and extracted with chloroform using a Soxhlet extractor over the course of 2 weeks. The resulting suspension was then dried in vacuo, and the off-white solid was serially triturated at 50°C with water, methanol, and hexanes, yielding Synthesis of H 3 L OMe . A 1 L Schlenk flask was charged with a 500 mL of a CH 2 Cl 2 solution of 3,5-dimethoxybenzoic acid (82.3 g, 452 mmol, 3.3 equiv) and then purged with N 2 . While under the N 2 purge, 1,1′-carbonyldiimidazole (77.7 g, 475 mmol, 3.5 equiv) was added as a solid. The reaction mixture was then allowed to stir for 1 h under N 2 . After the headspace was purged with N 2 to ensure the liberation of evolved CO 2 , tris(2-aminoethyl)amine (20.5 mL, 137 mmol) was then added in a single portion via syringe, and the reaction was allowed to stir for 12 h. After the reaction mixture was washed five times with H 2 O using a separatory funnel, the organic layer was dried in vacuo. The resulting pale-yellow residue was recrystallized from THF to yield H 3 L OMe as colorless crystals. Yield: 81.4 g, 93%. The 1 H NMR spectrum of this compound matched that reported previously. 57

Synthesis of [K(18-crown-6)][CoL OCH2O ].
In the glovebox, a scintillation vial was charged with H 3 L OCH2O (0.100 g, 0.169 mmol) and KH (0.027 g, 0.68 mmol, 4 equiv). DMA (2 mL) was added, and the effervescent solution was stirred for 1 h, during which time bubbling ceased. This solution was filtered through Celite to remove residual KH, added to solid Co(OAc) 2 (0.046 g, 0.25 mmol, 1.5 equiv), and stirred for 12 h. After filtering through Celite and washing the filter cake with additional DMA (4 mL), Et 2 O was added to the filtrate to precipitate the desired complex. This suspension was stirred for 30 min and then filtered through a fine-porosity fritted funnel. The crude complex remaining on the funnel was washed five times with a 5:1 Et 2 O/DMA solution. The resulting greenish residue was then dissolved in a solution of 18-crown-6 (0.054 g, 0.20 mmol, 1.2 equiv) in CH 2 Cl 2 (2 mL) and stirred for several minutes before being filtered through Celite. The filtrate was dried in vacuo and then taken up in a minimum amount of MeCN. Storage in a glovebox freezer at −35°C for 3 days yielded turquoise crystals, which were harvested and washed with Et 2 O to remove a small amount of cocrystallized 18crown-6. The crystals were redissolved in CH 2 Cl 2 and then dried in vacuo at 60°C to yield [K (18- ]. In the glovebox, a scintillation vial was charged with H 3 L OCH2O (0.100 g, 0.169 mmol) and KH (0.027 g, 0.68 mmol, 4 equiv). DMA (2 mL) was added, and the effervescent solution was stirred for 1 h, during which time bubbling ceased. This solution was filtered through Celite to remove residual KH and then added to solid Ni(OAc) 2 (0.045 g, 0.25 mmol, 1.5 equiv) and stirred for 12 h. After filtering through Celite and washing the filter cake with additional DMA (4 mL), Et 2 O was added to the filtrate to precipitate the desired complex. This suspension was stirred for 30 min and then filtered through a fine-porosity fritted funnel. The crude complex remaining on the funnel was washed five times with a 5:1 Et 2 O/DMA solution. The resulting light-orange residue was then dissolved in a solution of 18-crown-6 (0.054 g, 0.20 mmol, 1.2 equiv) in CH 2 Cl 2 (2 mL) and stirred for several minutes before being filtered through Celite. The filtrate was dried in vacuo and then taken up in a minimum amount of MeCN. Storage in a glovebox freezer at −35°C for 3 days yielded orange/pink crystals, which were harvested and washed with Et 2 O to remove a small amount of cocrystallized 18-crown-6. The crystals were redissolved in CH 2 Cl 2 and then dried in vacuo at 60°C to yield [K (18- ]. In the glovebox, a scintillation vial was charged with H 3 L OCH2O (0.100 g, 0.169 mmol) and KH (0.027 g, 0.68 mmol, 4 equiv). DMA (2 mL) was added, and the effervescent solution was stirred for 1 h, during which time bubbling ceased. This solution was filtered through Celite to remove residual KH and then added to solid Zn(OAc) 2 (0.047 g, 0.25 mmol, 1.5 equiv) and stirred for 12 h. After filtering through Celite and washing the filter cake with additional DMA (4 mL), Et 2 O was added to the filtrate to precipitate the desired complex. This suspension was stirred for 30 min and then filtered through a fine-porosity fritted funnel. The crude complex remaining on the funnel was washed five times with a 5:1 Et 2 O/DMA solution. The resulting colorless residue was then dissolved in a solution of 18-crown-6 (0.054 g, 0.20 mmol, 1.2 equiv) in CH 2 Cl 2 (2 mL) and stirred for several minutes before filtering through Celite. The filtrate was dried in vacuo and then taken up in a minimum amount of MeCN. Storage in a glovebox freezer at −35°C for 3 days yielded colorless crystals, which were harvested and washed with Et 2 O to remove a small amount of cocrystallized 18crown-6. The crystals were redissolved in CH 2 Cl 2 and then dried in vacuo at 60°C to yield [K (18-

Synthesis of [K(18-crown-6)][NiL OMe ].
In the glovebox, a scintillation vial was charged with H 3 L OMe (1.0 g, 1.57 mmol) and KH (0.201 g, 5.01 mmol, 4 equiv). DMA (8 mL) was added, and the effervescent solution was stirred for 1 h, during which time bubbling ceased. This solution was filtered through Celite to remove residual KH, and then added to solid Ni(OAc) 2 (0.304 g, 1.72 mmol, 1.5 equiv) and stirred for 3 h. After filtering through Celite, 18-crown-6 (0.455 g, 1.72 mmol, 1.2 equiv) was added to the solution, which was then layered with Et 2 O and allowed to stand overnight, resulting in the crystallization of the desired complex. The pink-orange crystals were harvested and washed with a 5: In the glovebox, a scintillation vial was charged with H 3 L OMe (0.200 g, 0.314 mmol) and KH (0.050 g, 1.2 mmol, 4 equiv). DMA (2 mL) was added, and the effervescent solution was stirred for 1 h, during which time bubbling ceased. This solution was filtered through Celite to remove residual KH and then added to solid Zn(OAc) 2 (0.086 g, 0.47 mmol, 1.5 equiv) and stirred for 3 h. After the mixture was filtered through Celite, Et 2 O was added to precipitate the desired complex. This suspension was stirred for 30 min and then filtered through a fineporosity fritted funnel. The crude complex remaining on the funnel was washed five times with a 5:1 Et 2 O/DMA solution. The resulting colorless residue was then dissolved in a solution of 18-crown-6 (0.099 g, 0.38 mmol, 1.2 equiv) in CH 2 Cl 2 (2 mL) and stirred for several minutes before filtering through Celite. The filtrate was dried in vacuo and then taken up in a minimum amount of MeCN. Storage in a glovebox freezer at −35°C for 3 days yielded colorless crystals which were harvested together and washed with Et 2 O to remove a small amount of cocrystallized 18-crown-6. The crystals were redissolved in CH 2 Cl 2 and then dried in vacuo at 60°C to yield [K (18- The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.inorgchem.3c01335. Nuclear magnetic resonance data for all newly reported compounds; additional UV−visible absorption spectroscopy data; fitting of all electron paramagnetic resonance spectra; additional cyclic voltammetry data; details of crystallographic data collection and refinement; and details of density functional theory calculations (PDF) Cartesian coordinates for all geometrically relaxed structures obtained from density functional theory calculations (XYZ)