Mimicking the Cu Active Site of Lytic Polysaccharide Monooxygenase Using Monoanionic Tridentate N-Donor Ligands

Abstract

In an effort to prepare small molecule mimics of the active site of lytic polysaccharide monooxygenase (LPMO), three monoanionic tridentate N donor ligands comprising a central deprotonated amide group flanked by two neutral donors were prepared, and their coordination chemistry with Cu(I) and Cu(II) was evaluated. With Cu(I), a dimer formed, which was characterized by X-ray crystallography and NMR spectroscopy. A variety of mononuclear and dinuclear Cu(II) species with a range of auxiliary ligands (MeCN, Cl^–, OH^–, OAc^–, OBz^–, CO₃^2–) were prepared and characterized by X-ray diffraction and various spectroscopies (UV–vis, EPR). The complexes exhibit structural similarities to the LPMO active site.

Introduction

In the subject area of challenging substrate oxidations by copper enzymes,¹ lytic polysaccharide monooxygenase (LPMO) is an important biological example of a catalyst capable of activating dioxygen to cleave strong C–H bonds, which in this case are those in cellulose and associated polymers, to yield short-chain oligomers.²⁻⁶ While the structure of LPMO and its active site are well-known,^7,8 the precise mechanism of oxygen activation and subsequent oxidation of the cellulose substrate and how the structural motif of the active site of LPMO supports these processes remains the subject of extensive research efforts.

The active site of LPMO contains a mononuclear copper center with three conserved coordinating groups: two histidine residues and a more unusual third group, the amine terminus of the protein chain (Figure 1). This terminus belongs to one of the coordinating histidine residues, thereby forming a metalacyclic moiety dubbed the “histidine brace”.⁸ The enzyme is unreactive when the Cu center is in the +2 oxidation state, but upon reduction to Cu(I) the active site can react with O₂ to generate an oxidizing species.⁹ The protonation state of the ligands is ambiguous, with neutral or monoanionic environments having been proposed.¹⁰⁻¹² Even though the specific Cu species responsible for the proposed rate-determining step of hydrogen atom transfer (HAT) from saccharide substrates is unknown, several Cu–oxygen cores have been implicated, such as [Cu–OO]⁺, [Cu–OH]²⁺, and [Cu–O]⁺.^8,10,13−20 These cores have been studied in model complexes, in the gas phase, and/or computationally and have been shown to be competent at HAT from various O–H and C–H bonds.

Figure 1.

Proposed active site of lytic polysaccharide monooxygenase (LPMO) enzymes: neutral (left) and monoanionic (right) states showing only the conserved supporting ligands.

In previous work, formally Cu(III) species with [Cu–OH]²⁺ cores using a family of dianionic pyridine dicarboxamide supporting ligands (Figure 2) were prepared and shown to perform proton-coupled electron transfer (PCET) reactions rapidly.²²⁻²⁸ The two strongly electron-donating carboxamide donors of these ligands stabilize such oxidized species, but on the other hand, they hinder formation of Cu(I) complexes. As a result, it is not possible to use biomimetic Cu(I)/O₂ reactions³² directly with these ligands to produce reactive cores such as [Cu–OO]⁺. Instead, this core was accessed by introduction of free superoxide to Cu(II) precursors.^{29−31,33,34} With the goal of examining the effects of changing the supporting ligand charge, we sought to design pincer ligands that placed a carboxamido group in the central position with flanking neutral N donors. We hypothesized this monoanionic donor set would mimic the postulated deprotonated histidine brace of LPMO and sufficiently decrease the overall donor strength, allowing us to access the Cu(I) state and enabling reactions with O₂ to be observed, while potentially retaining the accessibility of the formally Cu(III) state. In support of this hypothesis, many mononuclear Cu–O₂ adducts that are formed by the reaction of Cu(I) complexes with dioxygen feature monoanionic supporting ligands, although these are typically bidentate, such as β-diketiminates or anilido-imines (Figure 2).^21,35,36 These adducts typically exist as transient and/or side-on species.^21,37,38 We speculated that adding a third arm to a similar ligand motif could force the formation of end-on copper–dioxygen species akin to that proposed for LPMO⁹ by blocking the extra coordination site of the Cu center.

Figure 2.

Previously studied ligands used to study Cu–oxygen chemistry.

Thus inspired, we designed three new ligands that provide a N-donor ligand environment analogous to the proposed deprotonated active site of LPMO, employing an anionic amido group within a tridentate framework. These ligands, derived from the proligands HL¹–HL³ (Figure 3), feature a pyridine carboxamido fragment paired with either a pyrazolyl (HL¹ or HL²) or trisubstituted imine group (HL³). This design places the carboxamido unit in the middle position, similar to what is present in LPMO, from where it could, in principle, influence charge delocalization and reactivity of the bound dioxygen moiety. Herein, we report success in synthesizing the target ligands and exploring their copper coordination chemistry. Ultimately, however, our efforts to generate reactive copper–oxygen species were thwarted. Nonetheless, the fundamental advances in coordination chemistry we describe through the development of new ligands and complexes provides a foundation for future work, as literature examples of Cu complexes with monoanionic N,N′,N′′ ligands employing a central anionic group are fairly limited.³⁹⁻⁶⁹ We are hopeful that further modifications of the scaffolds we have studied here could yield future successes with copper–oxygen chemistry.

Figure 3.

Proligands and derived complexes prepared in this work.

Results and Discussion

Ligand Synthesis

The pyridine–carboxamide–pyrazole proligands HL¹ and HL² were synthesized in three steps, in a similar fashion to a previous report,⁷⁰ where a nucleophilic aromatic substitution of 1-fluoro-2-nitrobenzene with NaH and 3,5-subsituted pyrazoles yielded the pyrazole-substituted nitrobenzene compounds, which were reduced to anilines via palladium-catalyzed hydrogenation (Scheme 1). These anilines were converted to the final proligands in overall yields of 59% and 44% for HL¹ and HL², respectively, via reaction with picolinoyl acid chloride, prepared in situ from oxalyl chloride and picolinic acid. The pyridine–carboxamide–imine proligand HL³ was synthesized in two steps by performing an amidation reaction between 2-aminoacetophenone and picolinic acid using tosyl chloride and triethylamine, as previously reported,⁷¹ then assembling the final proligand via a TiCl₄-aided condensation between the newly formed ketone and 2,6-diisopropylaniline in an overall yield of 25–33%. The low yield is mostly due to the latter step, as the imine condensation between these two reactants proved to be problematic, being inaccessible by several other routes and only poorly afforded by the one we utilized. Nonetheless, gram amounts of ligand could be prepared due to the readily available starting materials. The formulations of the intermediates and the three proligands were established by ¹H and ¹³C{¹H} NMR spectroscopy (Figures S1–S13) and elemental analysis.

Scheme 1. Syntheses of the Proligands HL^1-3 Used in This Work.

Cu(I) Metalation

Since the active site of LPMO uses a Cu(I) center to reduce oxygen, we sought to synthesize a mononuclear Cu(I) complex by reacting HL² with K[N(SiMe₃)₂] (KHMDS) and Cu(MeCN)₄OTf in MeCN (Scheme 2). The resulting orange compound was characterized by NMR spectroscopy, UV–visible spectroscopy, cyclic voltammetry, CHN analysis, and X-ray diffraction. The ¹H NMR spectrum (Figure S14) of the new species depicted a loss of the amide N–H resonance and an upfield shift of all resonances relative to the spectrum of HL². In addition, separate peaks for the methyl groups on the iPr substituents were observed in both the ¹H NMR and ¹³C{¹H} NMR spectra, the latter of which contain the full complement of 21 peaks (Figure S14). These data indicate deprotonation of HL² and coordination to a copper ion(s). Such coordination is also supported by the presence of absorption features in the UV–vis spectrum of the orange product (Figure S15) at 426 nm (ε = 16490 M^–1 cm^–1) and 592 nm (ε = 1365 M^–1 cm^–1).

Scheme 2. Metalation of HL² with Cu(I).

The X-ray crystal structure (Figure 4) revealed that instead of being mononuclear the compound is a dimer, (L²Cu)₂, where each ligand bridges the two copper ions in a (κ¹, κ²) fashion. The Cu(I) ions adopt distorted trigonal planar geometries. The Cu–N lengths are typical for similar Cu(I) complexes,^44,45,49,60 and the Cu–Cu distance is 2.7509(7) Å, which is too long for significant bonding interactions between the two metals.⁷²

Figure 4.

X-ray crystal structure of (L²Cu)₂ with the THF molecules and H atoms omitted for clarity. Atoms are shown as 30% thermal ellipsoids. Selected bond distances (Å) and angles (deg): Cu1–N1, 1.909(2); Cu1–N3, 1.935(2); Cu1–N4, 2.151(2); N1–Cu1–N3, 155.79(9); N1–Cu1–N4, 118.13(9); N3–Cu1–N4, 81.90(9).

Although the solid-state structure shows the formation of a dimer for the 1:1 Cu(I)/L² complex, it is not clear if the dimeric structure is retained in solution. Some species, like tris(pyrazolyl)hydroborate Cu(I) complexes, have been shown to crystallize as dimers, but the solution-state data suggests the dimers can dissociate in solution to monomeric units.^73,74 The solution NMR data are consistent with both the hypothesized monomer in Scheme 1 and (L²Cu)₂, as the symmetry of the solid-state dimer structure would result in only one set of ligand resonances.

Even not knowing the nuclearity in solution, and because Cu(I) dimers can activate small molecules,^75,76 we examined the reactivity of the Cu(I) complex in solution with O₂ in an attempt to access and stabilize a 1:1:1 L²/Cu/O₂ adduct. Thus, we bubbled gaseous O₂ into a THF solution of (L²Cu)₂ at −80 °C for 2 min. There were no apparent changes in the spectrum, indicating no reaction with O₂ at this temperature (Figure S15). Additionally, no spectral changes were observed upon warming the solution in the presence of O₂. This unfortunate lack of reactivity of the Cu(I) complex contrasts with the reactivity of other species supported by tridentate monoanionic N-donor ligands,^77,78 the reasons for which are unknown.

Cyclic voltammetry experiments using THF solutions of (L²Cu)₂ revealed an irreversible oxidation event at 170 mV vs Fc/Fc⁺ (Figure S24; scan rate 100 mV s^–1; 0.3 M Bu₄NPF₆). When cycling past this oxidation event, another irreversible reduction occurs on the reducing sweep at −83 mV vs Fc/Fc⁺. However, this reduction event is not observed when performing the reductive sweep without first undergoing oxidation. This finding is suggestive of a highly hysteretic Cu(I/II) redox process whereby oxidation generates a Cu(II) complex that undergoes a chemical transformation to a species that is then reduced at the low potential. We speculate that this species may be a monomeric Cu(II) compound on the basis of observed formation of such complexes in other synthetic reactions described below.

Cu(II)-Cl Complexes Bearing (L¹)^- and (L²)^-

We reacted HL¹ and HL² with multiple Cu(II) sources with the aim of preparing mononuclear copper complexes. Deprotonation of HL¹ and HL² with KHMDS followed by reaction with anhydrous CuCl₂ yielded dark green products that were characterized by UV–vis and EPR spectroscopy, elemental analysis, and X-ray crystallography. Crystalline samples that were suitable for X-ray diffraction were obtained and revealed the products to be L¹CuCl and L²CuCl (Figure 5). In the solid state, L²CuCl exists as a dimer composed of two L²CuCl units where the Cl from one unit has a weak interaction in the axial position of the Cu in the other unit (2.7025(9) and 2.7464(9) Å), a motif seen in other Cu(II)–halide complexes supported by tridentate ligands.^{67,68,79−84} On the other hand, L¹CuCl does not exhibit the same phenomenon, remaining monomeric in the solid state, as seen in other similar Cu(II)–halide complexes.^{40,41,43,85,86} The bond lengths are comparable between the two complexes, but all of the bond lengths were slightly shorter in L¹CuCl, and the geometry index values, τ₄, of the metal centers are 0.37 and 0.31 for L¹CuCl and L²CuCl, respectively, indicating moderate deviation from a square planar geometry. Compared to a known Cu(II)–Cl structure of LPMO,⁷ the Cu–N and Cu–Cl distances are similar (Cu–N(his): 2.0 Å on average in the LPMO structure and Cu–N(pyridine, pyrazole): 1.99 and 2.03 Å on average in L¹CuCl and L²CuCl, respectively; Cu–N(NH₂): 2.2 Å in the LPMO structure and Cu–N(NH): 1.93 and 1.95 in L¹CuCl and L²CuCl, respectively; Cu–Cl 2.3 Å in the LPMO structure and 2.21 and 2.27 Å in L¹CuCl and L²CuCl, respectively). However, the geometries about the Cu ions in L¹CuCl and L²CuCl deviate from square planar geometries more than in the LPMO structure (τ₄ of 0.11).⁷

Figure 5.

X-ray crystal structures of L¹CuCl (top) and L²CuCl (middle) and a representation of the solid-state electrostatic dimer of L²CuCl (bottom). H atoms are omitted for clarity, and all non-H atoms are shown as 30% thermal ellipsoids. Selected bond distances (Å) and angles (deg) for L¹CuCl: Cu1–N1, 1.9909(16); Cu1–N3, 1.9341(15); Cu1–N4, 1.9999(17); Cu1–Cl1, 2.2059(5); N1–Cu1–N4, 149.41(7); N3–Cu1–Cl1, 158.01(5). Selected bond distances (Å) and angles (deg) for L²CuCl: Cu1–N1, 2.032(3); Cu1–N3, 1.947(3); Cu1–N4, 2.030(3); Cu1–Cl1, 2.2701(9); Cu1–Cl2, 2.7025(9); N1–Cu1–N4, 144.64(12); N3–Cu1–Cl1, 171.00(9).

The complexes L¹CuCl and L²CuCl were also characterized by EPR spectroscopy (THF, X-band, 30 K, Figure S36) which showed signals typical for tetragonal, monomeric Cu(II) complexes, indicating that the dimeric solid-state structure for the latter complex does not persist in solution. The simulated g values (Table S1) support some degree of rhombicity. Both EPR spectra show Cu hyperfine and N superhyperfine splitting, the parameters for which were determined by simulation (Table S1). The UV–vis spectra of the complexes (in MeCN) have weak features consistent with d–d transitions in Cu(II) species (Figure S16; L¹CuCl: 463 nm (ε = 358 M^–1 cm^–1), 604 nm (ε = 152 M^–1 cm^–1), 775 nm (ε = 198 M^–1 cm^–1); L²CuCl: 458 nm (ε = 325 M^–1 cm^–1), 606 nm (ε = 105 M^–1 cm^–1), and 815 nm (ε = 177 M^–1 cm^–1)). Taken together, the data show that mononuclear Cu(II) complexes containing (L¹)^– and (L²)^– can be synthesized, and they model the histidine brace of LPMO.

Cu(II)–Solvento Complexes Bearing (L¹)^– and (L²)^–

While the chloro complexes are stable and useful for structural assessment, we sought compounds that could more readily be converted to various copper–oxygen species by virtue of the presence of a labile ligand such as MeCN. To this end, we deprotonated HL¹ and HL² with KHMDS in MeCN and then added [Cu(MeCN)₅][SbF₆]₂ to the solutions. The resulting green products were characterized by EPR and UV–vis spectroscopy, elemental analysis, and high-resolution electrospray ionization–mass spectrometry (HR ESI-MS, Figure S34). On the basis of these data we tentatively propose the formulations [L¹CuMeCN][SbF₆] and [L²CuMeCN][SbF₆]. Thus, the EPR spectra (Figure S37) reveal the unpaired spins are localized on the Cu(II) ions, as evidenced by the hyperfine interactions in the g_parallel region of the pseudoaxial signal, but superhyperfine splitting from the N atoms is not apparent (parameters listed in Table S2). The UV–vis spectra of the complexes in MeCN were measured, with that of [L¹CuMeCN][SbF₆] having features at 466 nm (ε = 390 M^–1 cm^–1) and 731 nm (ε = 220 M^–1 cm^–1) and that of [L²CuMeCN][SbF₆] having similar absorptions at 462 nm (ε = 255 M^–1 cm^–1) and 752 nm (ε = 122 M^–1 cm^–1) (Figure S18). These low absorbing peaks are consistent with d–d transitions typical in Cu(II) complexes, but they are different from the features found for L¹CuCl and L²CuCl, consistent with a different fourth ligand that perturbs the electronic structure of the complex. Unfortunately, attempted elemental analyses were inconclusive, as the results were inconsistent with the proposed structures and could not be fit to a combination of metals, ligands, solvents, and contaminants, possibly due to a variety of ligands being present in the fourth coordination site of the copper (i.e., other solvents or a carbonyl from a neighboring molecule), or additional solvent coordination on the fifth position. In view of this issue, high-resolution ESI-MS data were collected for acetonitrile solutions of the complexes and yielded positive ion peak envelopes that may be assigned to 1:1 Lⁿ/Cu (n = 1 and 2) species with no bound acetonitrile (Figure S34). Attempts to obtain X-ray crystal structures of [L¹CuMeCN][SbF₆] and [L²CuMeCN][SbF₆] were unsuccessful, despite exhaustive efforts to grow crystals suitable for X-ray diffraction and microcrystal electron diffraction. Ultimately, the nature of the species formed in solution remained uncertain, so alternative syntheses to discrete Cu(II) species were explored.

Cu(II)-OAc Complexes Bearing (L¹)⁻ and (L²)^–

In a different synthetic route, we reacted HL¹ and HL² with Cu(OAc)₂·H₂O in the presence of 3 Å molecular sieves in MeCN, which yielded vibrant green-blue products. Both products were characterized by X-ray crystallography, EPR and UV–vis spectroscopy, and elemental analysis. X-ray diffraction quality crystals of both products were obtained, and the structures revealed the products to be L¹CuOAc and L²CuOAc (Figure 6). The acetate ligands are bound in monodentate fashion with weak interactions from the second carboxylate oxygens in the axial position of the Cu(II) ion (Cu–O3 lengths: 2.611(5) Å for L¹CuOAc and 2.3310(19) Å for L²CuOAc). The geometric parameter, τ₄, of the metal center is about equal for each complex: 0.33 for L¹CuOAc and 0.36 for L²CuOAc. All other bond lengths in the two compounds are similar. The solid-state structures resemble those of other Cu(II)-acetate complexes supported by tridentate ligands.^{43,46,54,57,59,63,66,69,87,88} The Cu–N distances found for L¹CuOAc and L²CuOAc are all similar to previously published structures, and the Cu–O2 bond lengths of 1.920 and 1.9492 Å are within the range of the known Cu–O(proximal)OCH₃ bond lengths (1.92–2.08 Å). The Cu–O(distal)OCH₃ lengths of 2.611(5) and 2.3310(19) Å are also within or very similar to the range found for the previously reported similar Cu(II)–acetate structures (2.37–2.85 Å). Previously, LPMO and fungal LPMO-like copper proteins were discovered to contain a Cu(II) ligated by the histidine brace and a carboxylate residue. The binding of the carboxylate ligands to copper in L¹CuOAc and L²CuOAc bear a resemblance to those in the protein structures, where a Cu–O bond (1.98–2.16 Å) is in the histidine brace plane and a Cu–O interaction (2.68–3.07 Å) is in an axial position of the copper.⁸⁹⁻⁹¹

Figure 6.

X-ray crystal structures of L¹CuOAc (left) and L²CuOAc (right). H atoms are omitted for clarity, and all non-H atoms are shown as 30% thermal ellipsoids. Selected bond distances (Å) and angles (deg) for L¹CuOAc: Cu1–N1, 2.018(6); Cu1–N3, 1.923(6); Cu1–N4, 2.015(6); Cu1–O2, 1.921(5); Cu1–O3, 2.611(5); N1–Cu1–N4, 146.1(3); N3–Cu1–O2, 166.8(3). Selected bond distances (Å) and angles (deg) for L²CuCl: Cu1–N1, 2.044(2); Cu1–N3, 1.914(2); Cu1–N4, 2.002(2); Cu1–O2, 1.9500(17); Cu1–O3, 2.3296(18); N1–Cu1–N4, 141.48(8); N3–Cu1–O2, 167.67(8).

The complexes also were characterized by EPR spectroscopy (Figure S38), and the data show pseudoaxial signals with some degree of rhombicity as evidenced by the simulated g-values (Table S3), indicative of tetragonal Cu(II) complexes. Cu hyperfine and N superhyperfine couplings are observed in both spectra (parameters listed in Table S3). The UV–vis spectra for L¹CuOAc and L²CuOAc in MeCN (Figure S20) have peaks at 694 nm (ε = 208 M^–1 cm^–1) and 694 nm (ε = 292 M^–1 cm^–1), respectively, typical for d-d transitions of Cu(II) complexes but different than the features found for L¹CuCl, L²CuCl, [L¹CuMeCN][SbF₆], and [L²CuMeCN][SbF₆].

Cu(II)–Cl Complex Bearing (L³)⁻

Attempts to make a Cu(I) complex with HL³ afforded a very poorly soluble orange material, which appeared similarly unreactive toward oxygen as (L²Cu)₂ but was problematic to characterize. We therefore focused efforts on preparing Cu(II) complexes instead. The chloro complex L³CuCl was obtained via direct reaction between HL³, CuCl₂, and NEt₃, affording a brown-green species in 85% yield. The crystal structure (Figure 7) indicates a mononuclear complex with distorted square planar geometry (τ₄ = 0.40) and typical Cu–L bond lengths for Cu(II). The distortion in geometry is likely in large part due to the ligand’s unequal bite angles and nonplanar conformation and is similar to the structure we observed for L¹CuCl. EPR spectroscopy in CH₂Cl₂ at 30 K (Figure S39; simulated parameters in Table S4) shows a broad and mostly featureless resonance, save for some faint Cu hyperfine interactions. The origin of this broadness is unclear.

Figure 7.

Crystal structures of the Cu(II) complexes bearing L³, drawn at the 50% probability level. Hydrogen atoms, counterions, and solvent molecules omitted for clarity. Selected bond distances (Å) and angles (deg) for [L³CuMeCN][SbF₆]: Cu1–N1, 2.021(2); Cu1–N2, 1.926(2); Cu1–N3, 2.0514(19); Cu1–N4, 1.988(2); Cu1–O1, 2.199(2); N1–Cu1–N3, 144.20(7); N2–Cu1–N4, 175.07(8). Selected bond distances (Å) and angles (deg) for L³CuCl: Cu1–N1, 1.995(2); Cu1–N2, 1.912(2); Cu1–N3, 1.989(2); Cu1–Cl1, 2.2113(7); N1–Cu1–N3, 155.85(7); N2–Cu1–Cl1, 147.65(8). Selected bond distances (Å) and angles (deg) for [(L³Cu)₂OH][SbF₆]: Cu1–N1, 1.9819(17); Cu1–N2, 1.9088(16); Cu1–N3, 1.9466(16); Cu1–O2, 1.9182(11); N1–Cu1–N3, 150.07(6); N2–Cu1–O2, 147.18(6); Cu1–O2–Cu1′, 115.16(11). Selected bond distances (Å) and angles (deg) for L³CuOBz (averaged): Cu1–N1, 2.043(3); Cu1–N2, 1.929(3); Cu1–N3, 2.120(3); Cu1–O3, 1.922(3); Cu1–O5, 2.131(3); N1–Cu1–N3, 136.98(12); N2–Cu1–O3, 171.14(12); N1–Cu1–O5, 130.68(11). Selected bond distances (Å) and angles (deg) for (L³Cu)₂CO₃ (averaged): Cu1–N1, 2.034(5); Cu1–N2, 1.904(5); Cu1–N3, 2.020(5); Cu1–O3, 1.921(4); Cu1–O5, 2.131(3); N1–Cu1–N3, 139.54(12); N2–Cu1–O3, 166.29(19); N1–Cu1–O5, 98.04(16).

Cu(II)–Solvento Complex Bearing (L³)⁻

The solvento adduct [L³CuMeCN][SbF₆] was produced by reacting HL³, [Cu(MeCN)₅][SbF₆]₂, and NEt₃, affording the dark green complex in 78% yield. The structure of [L³CuMeCN][SbF₆] (Figure 7) is analogous to that of L³CuCl, but with axial coordination of a neighboring complex’s carbonyl group (2.20 Å bond length), resulting in a distorted square pyramidal (τ₅ = 0.35) coordination polymer (Figure S35). We speculate that this behavior is the result of a more electrophilic metal center than in the case of L³CuCl, and we note that it has been seen in other Cu(II) complexes with carbonyl-containing ligands.^41,43 The coordination polymer does not appear to persist in solution, as we observe a typical pseudoaxial monomeric Cu(II) EPR signal for the compound in THF at 30 K (Figure S39; simulated parameters in Table S5). UV–visible spectroscopy of both complexes in CH₂Cl₂ (Figures S17, S19) revealed weakly absorbing peaks in the visible region: 700 nm (260 M^–1 cm^–1) for [L³CuMeCN][SbF₆] and 566 nm (150 M^–1 cm^–1) and 760 nm (140 M^–1 cm^–1) for L³CuCl, typical values for Cu(II) d–d transitions.

μ-Hydroxo–Dicopper Complex

Seeking to prepare Cu–O species, we targeted a hydroxo complex via reaction between [L³CuMeCN][SbF₆] and NBu₄OH under inert atmosphere, which resulted in the formation of a red substance which could not be characterized. We then attempted to obtain the aquo complex by treating [L³CuMeCN][SbF₆] with water, but this unexpectedly led to the formation of a μ-hydroxo dimer, [(L³Cu)₂OH][SbF₆], a brown compound isolated in 54% yield. The crystal structure (Figure 7) indicates bond lengths typical for Cu(II) (e.g., Cu–OH bond length of 1.915 Å). A Cu–O–Cu bond angle of 115.5°, a prominent residual electron density peak at the cusp of the O atom, and the presence of one counterion per dimer reveal that the bridging moiety is a hydroxo rather than an oxo group. Similar to the monomeric complexes, the local geometry around each metal center is significantly distorted square planar (τ₄ = 0.45). UV–visible spectroscopy in CH₂Cl₂ (Figure S21) shows a peak at 534 nm (300 M^–1 cm^–1) and a shoulder feature at 747 nm (130 M^–1 cm^–1), consistent with d–d Cu(II) transitions. The absence of an EPR signal in frozen CH₂Cl₂ solution at 30 K (Figure S39) suggests that the dimer persists in solution and has an S = 0 ground state due to antiferromagnetic coupling of the unpaired spins. Such μ-hydroxo–dicopper(II) cores have been reported, with fairly wide-ranging bond metrics (reported Cu–OH bond lengths range in excess of 1.9–2.2 Å).⁹²⁻¹¹² The spontaneous formation of a hydroxo moiety likely occurs due to electronic deficiency in the ligand set of [L³CuMeCN]⁺ rendering the Cu center Lewis acidic; the incomplete yield and observed formation of a pale blue aqueous fraction during preparation (see Experimental section) suggests partial demetalation, releasing the ligand’s basic groups to facilitate the deprotonation.

Cu(II)–Benzoate Complex Bearing (L³)⁻

We also targeted the formation of a carboxylate complex by reacting [L³CuMeCN][SbF₆] with sodium benzoate; this resulted in the bright green compound L³CuOBz that crystallizes as a five-coordinate dimer (Figure 7) in 79% yield. In the solid state, two benzoate moieties bridge between the two Cu centers via each O atom, with prominent Jahn–Teller asymmetry between the two Cu–O bonds (avg 1.92 Å vs 2.13 Å), analogous to the Cu–Cl interactions in the dimeric solid state structure of L²CuCl. The local geometry at each Cu center is intermediate between square pyramidal and trigonal bipyramidal, leaning toward the latter (τ₅ = 0.57). On the other hand, we observed a typical square planar, pseudoaxial Cu(II) EPR signal in frozen CH₂Cl₂ solution at 30 K (Figure S39; simulated parameters in Table S6), indicating that a monomeric state, most likely with κ¹-coordinated benzoate, predominates in solution. UV–visible spectroscopy in CH₂Cl₂ (Figure S22) reveals one weakly absorbing peak in the visible region at 666 nm (220 M^–1 cm^–1) consistent with the Cu(II) d–d transition. Numerous examples of both mononuclear and carboxylate-bridged dimers of Cu exist in the literature.^45,112−115

Carbonate-Bridged Dicopper Complex

During our various metalation trials of HL³ with Cu(II), we observed that use of strong bases during the process, such as NaOH or NaOMe, led to formation of a bright green compound if ambient air was introduced into the reaction mixtures. The UV–visible spectrum of the green compound in CH₂Cl₂ (Figure S23) indicated a typical Cu(II) complex with a feature in the visible region at 679 nm (280 M^–1 cm^–1). Isolation and crystallization revealed this compound to be a carbonate-bridged dimer, (L³Cu)₂CO₃ (Figure 7), with each Cu center coordinating strongly to one O atom (avg bond length 1.92 Å) equatorially, and both interacting more weakly with the third O atom (avg bond length 2.41 Å) mostly aligned to the Jahn–Teller axis, giving a local coordination that is distorted square pyramidal (τ₅ = 0.25). EPR spectroscopy in frozen CH₂Cl₂ (Figure S39) revealed no signal, consistent with retention of the dimeric structure in solution. The route of formation of this species is most likely capture of atmospheric CO₂, either by the utilized bases themselves or some intermediate basic Cu species. We readily synthesized the same dimer by direct metalation using HL³, anhydrous CuCl₂, and K₂CO₃, in 93% yield. Copper(II) dimers bridged by carbonate in this manner are known in the literature and exhibit similar bond lengths.^116−123

Electrochemistry

We investigated the accessibility of additional oxidation states using cyclic voltammetry in the 11 Cu(II) complexes: L¹CuCl, L²CuCl, [L¹CuMeCN][SbF₆], [L²CuMeCN][SbF₆], L¹CuOAc, and L²CuOAc in MeCN solutions and [L³CuMeCN][SbF₆], L³CuCl, [(L³Cu)₂OH][SbF₆], L³CuOBz, and (L³Cu)₂CO₃ in CH₂Cl₂ solutions (Figures S25–S30). No reversible oxidation or reduction events were observed for any of the compounds, with several irregular irreversible oxidative features starting in the vicinity of 0.7–1 V vs Fc⁺/Fc, most consistent with oxidative degradation of the compounds. We therefore conclude that highly oxidized Cu complexes supported by (Lⁿ)^– (n = 1, 2, or 3) are not feasible species to access due to their apparent instabilities. Irreversible reductive features are observed below −0.2 V vs Fc⁺/Fc, and based on our Cu(I) metalation experiments, we suspect that these most likely correspond to Cu(I) reduction accompanied by extensive coordinative rearrangement to form dimeric or polymeric aggregates like (L²Cu)₂.

Attempted Synthesis of Copper-Superoxide Species

Since neither the Cu(I) nor Cu(III) states were accessible for reactivity studies, we sought to produce mononuclear Cu(II)–oxygen species via reaction of Cu(II) complexes with superoxide. In a method analogous to that used in previous work with pyridine(dicarboxamide) systems,^29,34 we reacted [LⁿCuMeCN][SbF₆] (n = 1–3) with [K(Krypt)][O₂] in THF at −80 °C. Unfortunately, we did not observe the formation of spectroscopic features indicative of Cu–O₂ species upon addition of the superoxide salt to the Cu(II) complexes under these conditions (Figures S31–S33). Additionally, we did not witness any hydrogen atom transfer reactivity with substrates like TEMPOH or 9,10-dihydroanthracene with the solutions comprising [L³CuMeCN][SbF₆] + [K(Krypt)][O₂].

Conclusions

We prepared three new monoanionic N,N′,N′′ ligands with a central anionic amido coordinating group flanked by two neutral N donors in an effort to examine in model complexes the effects of a single negative charge as debated for the LPMO active site. A Cu(I) dimer ligated by ^–L² was isolated from an attempt to synthesize a mononuclear Cu(I) complex. Attempts to form a 1:1 Cu/O₂ complex by reacting (L²Cu)₂ with O₂ were not successful. The monoanionic supporting ligands bind Cu(II) in a tridentate pincer fashion, affording a number of Cu(II) complexes in which a variety of neutral or anionic auxiliary ligands could be installed at the fourth and sometimes fifth coordination site, resulting in geometries for some of the monocopper(II) complexes that resemble those of the Cu(II) resting state of LPMO. Depending on the nature of the auxiliary ligands, we obtained mononuclear, dinuclear, and fluxional species. The challenges we faced with both obtaining copper–oxygen species as well as oxidizing these compounds to a formally Cu(III) state prevented us from studying substrate oxidation reactivity with [Cu-OR]²⁺-type species. While we had hoped that the lower donor strength of these ligands would lend access to O₂-reactive Cu(I) species, the lack of which is a major drawback of the pyridine dicarboxamide ligand framework, we observed a tendency to form an unreactive dimer. We speculate this may be due to a steric insufficiency or high flexibility in the explored ligands. Bulkier modifications to these frameworks could alleviate this issue and would be reasonable targets for further investigation. Nonetheless, the ligands presented in this work give access to a relatively unexplored coordination environment for Cu and could be useful in other lines of coordination chemistry research.

Experimental Section

Materials and Methods

All air-sensitive manipulations were carried out in a dinitrogen-filled glovebox or under argon using Schlenk techniques. All reagents and solvents were purchased from commercial vendors and used as received unless otherwise noted. Tetrahydrofuran, dichloromethane, acetonitrile, pentane, diethyl ether, and heptane were passed through activated alumina columns and used directly or plumbed into a glovebox. Tetrahydrofuran, dichloromethane, and acetonitrile were stored over activated 3 Å molecular sieves in a dinitrogen-filled glovebox and, prior to use in spectroscopy and cyclic voltammetry experiments, filtered using a 25 mm diameter, 0.2 mm hydrophobic polytetrafluoroethylene (PTFE) syringe filter. 3,5-Diisopropylpyrazole was either made according to the previously published synthesis¹²⁴ or purchased from Tokyo Chemical Industry Co., Ltd. and used without further purification. KHMDS was purchased from Sigma-Aldrich and was recrystallized from toluene at −35 °C, filtered, and dried in vacuo before use. Cu(MeCN)₄OTf,¹²⁵ [Cu(MeCN)₅][SbF₆]₂,¹²⁶ and N-(2-acetylphenyl)picolinamide⁷¹ were made according to the previously published syntheses. 3,5-Dimethylpyrazole, 1-fluoro-2-nitrobenzene, picolinic acid, and anhydrous CuCl₂ were purchased from Sigma-Aldrich and used without further purification.

UV–vis spectra were collected using a HP8453 (190–1100 nm) diode array spectrophotometer equipped with a Unisoku low-temperature UV–vis cell holder. EPR spectra were collected on frozen 1 mM samples with a CW Elexsys E500 EPR spectrometer using X band (9.38 GHz) radiation at 35 dB and 30 K with the following conditions: microwave power, 0.0002 mW; modulation amplitude, 9.6 G; and modulation frequency, 100 kHz. EPR spectra were simulated using the EasySpin EPR simulation package, v. 5.1.¹²⁷ NMR spectra were collected on a Bruker Avance III HD nanobay AX-400, a Bruker Avance III HD-500, or an Agilent DD2 (500 MHz) spectrometer. Deuterated tetrahydrofuran was purchased from Cambridge Isotopes Laboratories, degassed, and dried over 3 Å molecular sieves prior to use. Cyclic voltammograms were recorded using an EC Epsilon potentiostat from BASi, a glassy carbon working electrode, and a Ag wire pseudoreference electrode. All cyclic voltammograms were performed in THF or MeCN with 0.3 M tetrabutylammonium hexafluorophosphate (TBAP) electrolyte, which was recrystallized several times from ethanol and dried under high vacuum before use and were internally referenced to the ferrocene/ferrocenium (Fc/Fc⁺) couple. The spectra were converted vs the standard Fc/Fc⁺ couple using standard conversion factors.¹²⁸ Elemental analysis was performed by the CENTC Elemental Analysis Facility (University of Rochester). ESI-HR-MS data is supplied by instruments funded by NIH grant 8P41GM103422.

For X-ray crystallography experiments, crystals were placed onto the tip of a MiTeGen cryoloop and mounted on a Bruker D8 VENTURE diffactometer equipped with a Photon III CMOS and a Mo Kα source using normal parabolic mirrors as monochromators or a Bruker X8 diffractometer equipped with a Kappa Apex II CCD using a graphite monochromator, with the crystals cooled to 100 K in a nitrogen stream. Data collection and processing were performed within the Bruker APEX3¹²⁹ software suite, using SAINT¹²⁹ for data reduction and SADABS¹³⁰ for scaling an absorption correction. Structure solutions were performed with SHELXT¹³¹ or SHELXS¹³² using OLEX 2¹³³ or ShelXle¹³⁴ as graphical interfaces. The structures were refined against F² on all data by full matrix least-squares with SHELXL.¹³²

3,5-Dimethyl-1-(2-nitrophenyl)-1H-pyrazole

The following preparation was adapted from a previously published synthesis.⁷⁰ Under an argon atmosphere, a flame-dried 500 mL Schlenk flask containing dry THF (100 mL) was charged with NaH (60% in mineral oil, 1.931 g, 48.3 mmol) and then cooled to 0 °C using an ice bath. The slurry was stirred while a solution of 3,5-dimethylpyrazole (5.00 g, 52.0 mmol) in dry THF (100 mL) was added dropwise via an addition funnel over 30 min. After the addition was complete, a solution of 1-fluoro-2-nitrobenzene (5.49 mL, 7.339 g, 52.0 mmol) in dry THF (100 mL) was added dropwise via an addition funnel over 30 min. The ice bath was removed, and the orange solution was allowed to stir at rt for 2 h under argon. The reaction was transferred to a separatory funnel and slowly quenched with a saturated aqueous NH₄Cl solution (100 mL). After separation of the phases, the aqueous layer was extracted with ethyl acetate (2 × 200 mL). The organic layers were combined, washed with brine (50 mL), dried with Na₂SO₄, and decanted, and the solvents were removed in vacuo. The resulting orange residue was purified via column chromatography on silica gel (50% CH₂Cl₂, 50% ethyl acetate) affording the pure product as an orange oil (8.209 g, 73% yield). ¹H NMR (500 MHz, CDCl₃) δ (ppm): 7.93 (dd, J = 8.1, 1.5 Hz, 1H), 7.66 (td, J = 7.8, 1.5 Hz, 1H), 7.53 (td, J = 7.8, 1.4 Hz, 1H), 7.45 (dd, J = 7.9, 1.4 Hz, 1H), 5.99 (s, 1H), 2.21 (s, 3H), 2.15 (s, 3H). ¹³C NMR (500 MHz, CDCl₃) δ (ppm): 150.67, 146.59, 141.09, 133.37, 133.29, 129.67, 129.34, 125.25, 107.00, 13.64, 11.49. Anal. Calcd for C₁₁H₁₁N₃O₂: C 60.82, H 5.1, N 19.34. Found: C 60.57, H 5.15, N 19.02.

3,5-Diisopropyl-1-(2-nitrophenyl)-1H-pyrazole

The following preparation was adapted from a previously published synthesis.⁷⁰ Under an argon atmosphere, a flame-dried 500 mL Schlenk flask containing dry THF (100 mL) was charged with NaH (60% in mineral oil, 3.941 g, 104.5 mmol) then cooled to 0 °C using an ice bath. The slurry was stirred while a solution of 3,5-diisopropylpyrazole (10.00 g, 65.7 mmol) in dry THF (100 mL) was added dropwise via an addition funnel over 30 min. After the addition was complete, a solution of 1-fluoro-2-nitrobenzene (6.93 mL, 9.268 g, 65.7 mmol) in dry THF (100 mL) was added dropwise via an addition funnel over 30 min. The ice bath was removed, and the orange solution was allowed to stir at rt for 2 h under argon. The reaction was transferred to a separatory funnel and slowly quenched with a saturated aqueous NH₄Cl solution (150 mL). After separation of the phases, the aqueous layer was extracted with ethyl acetate (2 × 200 mL). The organic layers were combined, washed with brine (50 mL), dried with Na₂SO₄, and decanted, and the solvents were removed in vacuo. The resulting orange residue was purified via column chromatography on silica gel (80% hexanes, 20% ethyl acetate) affording the pure product as an orange oil (10.735 g, 66% yield). ¹H NMR (400 MHz, CDCl₃) δ (ppm): 7.97 (dd, J = 8.1, 1.5 Hz, 1H), 7.68 (td, J = 7.7, 1.2 Hz, 1H), 7.56 (td, J = 7.9, 1.2 Hz, 1H), 7.51 (dd, J = 8.0, 1.2 Hz, 1H), 6.06 (s, 1H), 2.94 (sept, J = 6.9 Hz, 1H), 2.81 (sept, J = 6.9 Hz, 1H), 1.25 (d, J = 6.9 Hz, 6H), 1.18 (d, J = 6.9 Hz, 6H). ¹³C NMR (500 MHz, CDCl₃) δ (ppm): 161.04, 151.91, 146.90, 133.77, 133.09, 129.60, 129.32, 125.35, 100.39, 28.09, 25.77, 22.85, 22.80. Anal. Calcd for C₁₅H₁₉N₃O₂: C 65.91, H 7.01, N 15.37. Found: C 66.09, H 7.23, N 15.27.

2-(3,5-Dimethyl-1H-pyrazol-1-yl)aniline

The following preparation was adapted from a previously published synthesis.⁷⁰ Under argon, a Schlenk flask was charged with 3,5-dimethyl-1-(2-nitrophenyl)-1H-pyrazole (8.097 g, 37.3 mmol), 10 wt % Pd/C (0.282 g, 0.265 mmol), and 200 proof ethanol (50 mL). The flask was flushed with H₂ then stirred at rt for 24 h under a balloon of H₂, refilling the H₂ balloon when deflated. The reaction mixture was filtered through Celite, and the solids were washed with ethanol (50 mL). Note: Residual Pd/C should be wetted with water after filtration to prevent ignition of the solid. Solvent was removed in vacuo yielding the pure product as a brown solid (6.609 g, 95% yield). ¹H NMR (400 MHz, CDCl₃) δ (ppm): 7.13 (ddd, J = 8.1, 7.3, 1.6 Hz, 1H), 7.04 (dd, J = 7.6, 1.6 Hz, 1H), 6.78–6.68 (m, 2H), 5.95 (s, 1H), 3.77 (s, 2H), 2.26 (s, 3H), 2.11 (s, 3H). ¹³C NMR (500 MHz, CDCl₃) δ (ppm): 149.52, 143.52, 140.85, 129.52, 127.68, 125.62, 118.15, 116.73, 105.75, 13.75, 11.62. Anal. Calcd for C₁₁H₁₃N₃: C 70.56, H 7.00, N 22.44. Found: C 70.56, H 7.09, N 22.37.

2-(3,5-Diisopropyl-1H-pyrazol-1-yl)aniline

The following preparation was adapted from a previously published synthesis.⁷⁰ Under argon, a Fisher Porter tube was charged with 3,5-diisopropyl-1-(2-nitrophenyl)-1H-pyrazole (2.264 g, 8.28 mmol), 10 wt % Pd/C (0.063 g, 0.059 mmol), and 200 proof ethanol (15 mL). The system was sealed and flushed with H₂ through three pressurization/vent cycles then pressurized once more to 100 psi H₂ and stirred at rt overnight. The vessel was repressurized once during the reaction. Note: A safety shield should be placed in front of the Fisher Porter tube during this reaction. Residual pressure in the vessel was vented before filtering the reaction mixture through Celite and washing the solids with ethanol (15 mL). Note: Residual Pd/C should be wetted with water after filtration to prevent ignition of the solid. Solvent was removed in vacuo to yield the product as a white solid (1.876 g, 93% yield). ¹H NMR (400 MHz, CDCl₃): δ_H 7.18 (ddd, J = 8.1, 7.3, 1.5 Hz, 1H), 7.11 (dd, J = 7.8, 1.5 Hz, 1H), 6.78 (m, 2H), 6.02 (s, 1H), 3.83 (s, 2H), 3.00 (sept, J = 6.9 Hz, 1H), 2.83 (sept, J = 6.9 Hz, 1H), 1.29 (d, J = 6.9 Hz, 6H), 1.14 (d, J = 6.9 Hz, 6H). ¹³C NMR (500 MHz, CDCl₃) δ (ppm): 160.08, 151.75, 144.00, 129.66, 128.00, 126.01, 118.15, 116.74, 99.02, 28.15, 25.60, 23.01, 22.82. Anal. Calcd for C₁₅H₂₁N₃: C 74.03, H 8.70, N 17.27. Found: C 74.54, H 8.93, N 17.40.

N-(2-(3,5-Dimethyl-1H-pyrazol-1-yl)phenyl)picolinamide (HL¹)

The following preparation was adapted from a previously published synthesis.⁷⁰ In a 250 mL Schlenk flask under argon, a solution of picolinic acid (3.117 g, 25.3 mmol) in CH₂Cl₂ (35 mL) was cooled to 0 °C using an ice bath. With stirring, DMF (5 drops) was added dropwise to the solution, followed by another dropwise addition of oxalyl chloride (2.96 mL, 34.5 mmol). The purple solution was brought to rt and stirred for 1 h. The solvent and excess oxalyl chloride were removed in vacuo to yield the acid chloride as a dark black/brown oily solid. Note: the excess oxalyl chloride was collected in an external trap to be quenched with a saturated aqueous solution of sodium bicarbonate. To the same flask under argon, the acid chloride was dissolved in CH₂Cl₂ (35 mL) and cooled back to 0 °C using an ice bath. A solution of 2-(3,5-dimethyl-1H-pyrazol-1-yl)aniline (4.310 g, 23.0 mmol) in CH₂Cl₂ (35 mL) was added to the flask, followed by the addition of triethylamine (3.53 mL, 25.3 mmol). The flask was removed from the ice bath, and the reaction was stirred at rt for 12 h. The reaction mixture was filtered into a separatory funnel and washed with a saturated aqueous ammonium chloride solution (2 × 20 mL) and brine (30 mL). The organic layer was dried with Na₂SO₄, decanted, and then concentrated in vacuo. The crude product was purified by via column chromatography on silica gel (90% hexanes, 10% ethyl acetate;) affording the pure product as an yellow/orange oil (5.686 g, 85% yield). Over a few months, the oil solidified into a brown solid. ¹H NMR (400 MHz, CDCl₃): δ_H 11.11 (s, 1H), 8.69 (dd, J = 8.3, 1.4 Hz, 1H), 8.52 (ddd, J = 4.8, 1.7, 1.0 Hz, 1H), 8.21 (dt, J = 7.9, 1.1 Hz, 1H), 7.83 (td, J = 7.7, 1.7 Hz, 1H), 7.45 (ddd, J = 8.6, 7.6, 1.6 Hz, 1H), 7.40 (ddd, J = 7.6, 4.7, 1.2 Hz, 1H), 7.28 (dd, J = 7.9, 1.6 Hz, 1H), 7.19 (td, J = 7.7, 1.4 Hz, 1H), 6.06 (s, 1H), 2.43 (s, 3H), 2.19 (s, 3H). ¹³C NMR (500 MHz, CDCl₃) δ (ppm): 162.65, 150.31, 150.13, 148.28, 141.24, 137.48, 134.46, 129.22, 129.19, 126.52, 126.38, 123.75, 122.46, 121.89, 106.71, 13.72, 12.01. Anal. Calcd for C₁₇H₁₆N₄O: C 69.85, H 5.52, N 19.17. Found: C 69.64, H 5.52, N 19.17.

N-(2-(3,5-Diisopropyl-1H-pyrazol-1-yl)phenyl)picolinamide (HL²)

The following preparation was adapted from a previously published synthesis.⁷⁰ In a 100 mL Schlenk flask under argon, a solution of picolinic acid (2.230 g, 18.1 mmol) in CH₂Cl₂ (25 mL) was cooled to 0 °C using an ice bath. With stirring, DMF (5 drops) was added dropwise to the solution, followed by another dropwise addition of oxalyl chloride (2.11 mL, 24.7 mmol). The purple solution was brought to rt and stirred for 1 h. The solvent and excess oxalyl chloride were removed in vacuo to yield the acid chloride as a dark black/brown oily solid. Note: the excess oxalyl chloride was collected in an external trap to be quenched with a saturated aqueous solution of sodium bicarbonate. To the same flask under argon, the acid chloride was dissolved in CH₂Cl₂ (25 mL) and cooled back to 0 °C using an ice bath. A solution of 2-(3,5-diisopropyl-1H-pyrazol-1-yl)aniline (4.000 g, 16.4 mmol) in CH₂Cl₂ (25 mL) was added to the flask, followed by the addition of triethylamine (2.52 mL, 18.1 mmol). The flask was removed from the ice bath, and the reaction was stirred at rt for 12 h. The reaction mixture was filtered into a separatory funnel and washed with a saturated aqueous ammonium chloride solution (2 × 10 mL) and brine (20 mL). The organic layer was dried with Na₂SO₄, decanted, and then concentrated in vacuo. The crude product was purified by via column chromatography on silica gel (90% hexanes, 10% ethyl acetate) affording the pure product as a white solid (4.050 g, 71% yield). ¹H NMR (400 MHz, CDCl₃): δ_H 10.58 (s, 1H), 8.69 (dd, J = 8.3, 1.4 Hz, 1H), 8.46 (ddd, J = 4.8, 1.8, 1.0 Hz, 1H), 8.22 (dt, J = 7.9, 1.1 Hz, 1H), 7.83 (tt, J = 7.7, 1.7 Hz, 1H), 7.47 (td, J = 7.9, 1.5 Hz, 1H), 7.39 (ddt, J = 7.6, 4.8, 1.4 Hz, 1H), 7.33 (dd, J = 7.8, 1.6 Hz, 1H), 7.19 (td, J = 7.6, 1.3 Hz, 1H), 6.10 (s, 1H), 3.12 (sept, J = 7.0 Hz, 1H), 2.85 (sept, J = 6.8 Hz, 1H), 1.40 (d, J = 7.0 Hz, 6H), 1.09 (d, J = 6.9 Hz, 6H). ¹³C NMR (500 MHz, CDCl₃) δ (ppm): 162.61, 160.77, 152.30, 149.96, 148.10, 137.42, 135.17, 129.68, 129.58, 127.23, 126.40, 123.89, 122.47, 121.94, 99.92, 28.32, 25.62, 22.95, 22.83. Anal. Calcd for C₂₁H₂₄N₄O: C 72.39, H 6.94, N 16.08. Found: C 72.41, H 7.20, N 15.95.

N-(2-(1-((2,6-Diisopropylphenyl)imino)ethyl)phenyl)picolinamide (HL³)

In a flame-dried 250 mL Schlenk flask under argon, 2,6-bis(isopropyl)aniline (5.6 mL, 30 mmol) was dissolved in 150 mL of dry, degassed toluene and cooled in an ice–water bath. Titanium tetrachloride (0.55 mL, 5.0 mmol) was added dropwise with stirring, resulting in the formation of a red solution progressing to a brown suspension over the course of the addition. After stirring for 1 h, the reaction was warmed to room temperature, and N-(2-acetylphenyl)picolinamide (2.39 g, 10 mmol) was added, after which the reaction was heated to reflux for 48 h. The crude mixture was washed with saturated aqueous K₂CO₃, and the organic fraction was concentrated under reduced pressure. The majority of unreacted picolinamide starting material was removed by extracting the evaporated residue with pentane and passing through a filter. The resulting filtrate was again concentrated under reduced pressure and loaded onto a silica gel column presoaked in hexanes (4 × 20 cm) using a minimal amount of pentane. Separation was carried out using an autocolumn eluting with a gradient solvent mixture starting at 25% CH₂Cl₂ in hexane and ramping to 100% CH₂Cl₂ over 5 column volumes. The desired product was the second major fraction, eluting at approximately 66% CH₂Cl₂ (first was 2,6-bis(isopropyl)aniline, at approximately 33% CH₂Cl₂). The product was concentrated under reduced pressure to a viscous yellow oil, which on standing under dynamic vacuum overnight afforded a glassy yellow residue. Final purification was carried out via recrystallization from a slowly evaporating concentrated pentane solution in the presence of seed crystals (initially obtained via prolonged standing in ambient air of a small amount of product in its oily state), affording light yellow crystalline material. Yield: 1.135 g, 29%. ¹H NMR (500 MHz, CDCl₃): δ_H (ppm) 15.38 (s, 1H), 9.16 (d, J = 8.3 Hz, 1H), 8.27 (d, J = 7.8 Hz, 1H), 8.09 (d, J = 4.5 Hz, 1H), 7.91 (d, J = 8.0 Hz, 1H), 7.83 (td, J = 7.7, 1.5 Hz, 1H), 7.61 (t, J = 7.8 Hz, 1H), 7.32–7.19 (m, 4H), 2.85 (sept, J = 6.9 Hz, 2H), 2.29 (s, 3H), 1.18 (d, J = 6.9 Hz, 6H), 1.01 (d, J = 6.8 Hz, 6H). ¹³C NMR (500 MHz, CDCl₃): δ (ppm) 168.27, 164.60, 151.29, 148.30, 145.06, 140.60, 137.08, 136.71, 131.85, 130.14, 125.89, 124.06, 123.78, 123.00, 122.75, 122.51, 121.37, 28.55, 23.52, 22.96, 19.39. Anal. Calcd for C₂₆H₂₉N₃O: C 78.16, H 7.32, N 10.52. Found: C 78.41, H 7.46, N 10.57.

(L²Cu)₂

In a glovebox, a solution of KHMDS (32 mg, 0.16 mmol) in MeCN (3 mL) was added dropwise to a solution of HL² (50 mg, 0.14 mmol) in MeCN (2 mL) in a 20 mL scintillation vial while stirring. After complete addition of the base, a solution of Cu(MeCN)₄OTf in MeCN (2 mL) was added dropwise to the vial with stirring, and a bright orange powder started to precipitate from the solution. After 10 min of stirring, the MeCN was decanted from the vial. The solid was rinsed with MeCN (2 × 3 mL), decanted, and dried in vacuo to afford the pure compound (39 mg, 33%). X-ray quality crystals (orange blocks) were formed upon slow diffusion of pentane into a concentrated THF solution of (L²Cu^I)₂ at −35 °C. ¹H NMR (400 MHz, THF-d₈): δ_H 7.80 (app d, J = 7.8 Hz, 1H), 7.57 (app t, J = 7.6 Hz, 1H), 7.45 (app d, J = 7.6 Hz, 1H), 7.32–7.11 (m, 4H), 7.01 (app t, J = 6.0 Hz, 1H), 5.91 (s, 1H), 3.27 (sept, J = 6.9 Hz, 1H), 3.09 (sept, J = 6.7 Hz, 1H), 1.38 (overlapping doublets, J = 6.8 Hz, 6H), 1.17 (d, J = 6.9 Hz, 3H), 0.47 (d, J = 6.9 Hz, 3H). ¹³C NMR (500 MHz, THF-d8) δ (ppm): 166.47, 160.30, 155.27, 154.97, 152.93, 147.35, 137.85, 134.45, 130.23, 129.87, 128.06, 124.88, 123.24, 122.07, 97.65, 30.81, 30.39, 27.63, 24.50, 23.48, 22.67. UV–vis (THF, −80 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 426 (16490), 592 (1365). Anal. Calcd for C₄₂H₄₆Cu₂N₈O₂: C 61.37, H 5.64, N 13.63. Found: C 60.83, H 5.62, N 13.29.

L¹CuCl

In a glovebox, a 50 mL Schlenk flask was charged with a solution of HL¹ (111.2 mg, 0.38 mmol) in THF (4 mL). A solution of KHMDS (75.9 mg, 0.38 mmol) in THF (4 mL) was added dropwise to the flask with stirring. After complete addition of the base, a slurry of anhydrous CuCl₂ (49.1 mg, 0.37 mmol) in THF (4 mL) was added to the flask with stirring, producing a dark green reaction mixture. After stirring at rt for 30 min, the solvent and HMDS byproduct were removed in vacuo. CH₂Cl₂ (5 mL) was added to the flask, and the solution was filtered into a 20 mL scintillation vial using a 25 mm diameter, 0.2 mm hydrophobic PTFE syringe filter. The solvent was removed in vacuo. MeCN (5 mL) was used to dissolve the product and then filtered into a 20 mL scintillation vial using a 25 mm diameter, 0.2 mm hydrophobic PTFE syringe filter. The solvent was removed in vacuo, and the resulting green oil was triturated with diethyl ether (2 × 5 mL) resulting in a green powder (44.6 mg, 31%). Crystalline material of the title product was obtained by dissolving the green oil in a minimum amount of MeCN and layering with diethyl ether at rt, which afforded dark green plates that were suitable for X-ray diffraction. UV–vis (MeCN, 25 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 463 (358), 604 (152), 775 (198). Anal. Calcd for C₁₇H₁₅CuN₄OCl: C 52.31, H 3.87, N 14.35. Found: C 54.18, H 4.13, N 14.61.

L²CuCl

In a glovebox, a 50 mL Schlenk flask was charged with a solution of HL² (95.1 mg, 0.27 mmol) in THF (4 mL). A solution of KHMDS (54.4 mg, 0.27 mmol) in THF (4 mL) was added dropwise to the flask with stirring. After complete addition of the base, a slurry of anhydrous CuCl₂ (35.2 mg, 0.26 mmol) in THF (4 mL) was added to the flask with stirring, producing a dark green reaction mixture. After stirring at rt for 30 min, the solvent and HMDS byproduct were removed in vacuo. CH₂Cl₂ (5 mL) was added to the flask, and the solution was filtered into a 20 mL scintillation vial using a 25 mm diameter, 0.2 mm hydrophobic PTFE syringe filter. The solvent was removed in vacuo. MeCN (5 mL) was used to dissolve the product and then filtered into a 20 mL scintillation vial using a 25 mm diameter, 0.2 mm hydrophobic PTFE syringe filter. The solvent was removed in vacuo, and the resulting green oil was triturated with diethyl ether (2 × 5 mL) resulting in a green/brown powder (70.0 mg, 60%). Crystalline material of the title product was obtained by dissolving the green oil in a minimum amount of CH₂Cl₂ and layering with heptane at −20 °C, which resulted in an oil that was subsequently allowed to sit at rt overnight resulting in dark green plates that were suitable for X-ray diffraction. UV–vis (MeCN, 25 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 458 (325), 606 (105), 815 (177). Anal. Calcd for C₂₁H₂₃CuN₄OCl: C 56.50, H 5.19, N 12.55. Found: C 57.02, H 5.24, N 12.53.

L³CuCl

A 25 mL heart flask under ambient atmosphere was loaded with HL³ (80 mg, 0.20 mmol), anhydrous CuCl₂ (27 mg, 0.20 mmol), and 4 mL of CH₂Cl₂ and then mixed until all ligand dissolved, giving a dull green solution along with some undissolved CuCl₂. Triethylamine (30 μL, 0.22 mmol) was added, and the solution was mixed with sonication until all CuCl₂ dissolved and a darker green hue developed (ca. 5 min). Pentane (ca. 9 mL) was added gradually with stirring, causing a pale precipitate (presumably HNEt₃Cl) to form, which was removed by filtration. The filtrate was concentrated under reduced pressure to a brown-green residue, redissolved in 1.5 mL of CH₂Cl₂ and layered with 18 mL of pentane, then allowed to stand undisturbed for 24 h. Dark brown crystals (suitable for diffractometry) were decanted and washed with pentane, then dried under vacuum overnight. Despite drying under reduced pressure, product was isolated as a solvated solid with 0.5 molecules of CH₂Cl₂ per [Cu]. Yield: 91.3 mg, 85%. Anal. Calcd for C₂₆H₂₈ClCuN₃O + 0.5CH₂Cl₂: C 58.94, H 5.41, N 7.78. Found: C 59.62, H 5.67, N 7.61. UV–vis (CH₂Cl₂, 25 °C): 233 nm (23000 M^–1 cm^–1), 258 nm (25000 M^–1 cm^–1), 365 nm (7800 M^–1 cm^–1), 566 nm (150 M^–1 cm^–1), 760 nm (140 M^–1 cm^–1).

[L¹Cu(MeCN)][SbF₆]

In a glovebox, a 50 mL Schlenk flask was charged with a solution of HL¹ (50 mg, 0.17 mmol) in MeCN (4 mL). A solution of KHMDS (34 mg, 0.17 mmol) in MeCN (4 mL) was added dropwise to the flask with stirring. After complete addition of the base, a solution of [Cu(MeCN)₅][SbF₆]₂ (122 mg, 0.16 mmol) in MeCN (4 mL) was added dropwise to the flask with stirring, producing a dark green reaction mixture. After stirring at rt for 30 min, the solvent and HMDS byproduct were removed in vacuo. The title product was obtained as a green solid after precipitating the compound out of MeCN (0.5 mL) with diethyl ether (10 mL) several times, decanting the liquid, and drying the solid in vacuo (108 mg, 104%). UV–vis (MeCN, 25 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 466 (390), 731 (220). HR-MS (ESI, acetonitrile, positive ion) m/z: [L¹Cu + SbF₆ + H]⁺ Calcd For [C₁₇H₁₆CuN₄OSbF₆]⁺ (without acetonitrile) 591.96; Found 591.9591. Anal. Calcd for C₁₉H₁₈CuN₅OSbF₆ (with acetonitrile): C 36.13, H 2.87, N 11.09. Found: C 38.18, H 3.32, N 10.77.

[L²Cu(MeCN)][SbF₆]

In a glovebox, a 50 mL Schlenk flask was charged with a solution of HL² (50 mg, 0.14 mmol) in MeCN (4 mL). A solution of KHMDS (29 mg, 0.14 mmol) in MeCN (4 mL) was added dropwise to the flask with stirring. After complete addition of the base, a solution of [Cu(MeCN)₅][SbF₆]₂ (102 mg, 0.14 mmol) in MeCN (4 mL) was added dropwise to the flask with stirring, producing a dark green reaction mixture. After stirring at rt for 30 min, the solvent and HMDS byproduct were removed in vacuo. The title product was obtained as a green solid after precipitating the compound out of MeCN (0.5 mL) with diethyl ether (10 mL) several times, decanting the liquid, and drying the solid in vacuo (89 mg, 94%). UV–vis (MeCN, 25 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 462 (255), 752 (122). HR-MS (ESI, acetonitrile, positive ion) m/z: [L²Cu + SbF₆ + H]⁺ Calcd For [C₂₁H₂₄CuN₄OSbF₆]⁺ (without acetonitrile) 648.02; Found 648.0219. [L²Cu]⁺ Calcd For [C₂₁H₂₃CuN₄O]⁺ (without acetonitrile) 410.12; Found 410.1185. Anal. Calcd for C₂₃H₂₆CuN₅OSbF₆ (without acetonitrile): C 40.17, H 3.81, N 10.18. Found: C 41.73, H 4.17, N 9.98.

[L³Cu(MeCN)][SbF₆]

In a 25 mL heart flask, HL⁴ (200 mg, 0.50 mmol) and [Cu(NCCH₃)₅][SbF₆]₂ (370 mg, 0.50 mmol) were combined and purged with argon then dissolved in 5 mL of dry, degassed MeCN, resulting in formation of a green solution. Triethylamine (0.10 mL, 0.74 mmol) was then added, resulting in a darker green solution. At this point, the mixture was no longer air and moisture sensitive, and further workup was carried out under ambient atmosphere. The crude products were concentrated under reduced pressure to approximately 1 mL volume and selectively precipitated with 20 mL of diethyl ether. The supernatant was decanted, and the semisolid green precipitate was redissolved with 1 mL of MeCN and reprecipitated with 20 mL of diethyl ether five more times; gradual removal of colored impurities and [HNEt₃][SbF₆] occurred, leaving crystalline green precipitate by the final step. The precipitate was then dissolved in 5 mL of MeCN, and an approximately equal amount of diethyl ether was gradually added until the solution became turbid yet free of green crystals. The mixture was passed through a glass fiber pad filter, removing a dark brown residue. The filtrate was concentrated under reduced pressure, dissolved fully in a minimum of MeCN (approximately 1.5 mL), layered with approximately 1.5 mL of a 1:1 MeCN/ether mixture, then approximately 20 mL of diethyl ether. After standing undisturbed for 2 days, the product was obtained as dark green crystals upon decantation of supernatant, washing with diethyl ether, and drying under vacuum. Crystals suitable for X-ray diffraction were collected from the above recrystallization prior to drying. Yield: 289 mg, 78%. Anal. Calcd for C₂₈H₃₁CuF₆N₄OSb: C 45.52, H 4.23, N 7.58. Found: C 45.51, H 4.13, N 7.55. UV–vis (THF, −60 °C): 700 nm (260 M^–1 cm^–1).

L¹CuOAc

A solution of HL¹ (100 mg, 0.34 mmol) in anhydrous MeCN (3 mL) was added dropwise into a 20 mL scintillation vial containing Cu(OAc)₂·H₂O (74 mg, 0.37 mmol), anhydrous MeCN (3 mL), and 3 Å molecular sieves (500 mg). The reaction was stirred for 2 h at rt, resulting in a green-blue solution, which was filtered into a clean 20 mL scintillation vial using a 25 mm diameter, 0.2 mm hydrophobic PTFE syringe filter. The solvent was removed in vacuo. Crystalline material of the title product was obtained by vapor diffusion of anhydrous diethyl ether into a concentrated dry THF solution of L¹CuOAc at rt, which afforded green plate crystals that were suitable for X-ray diffraction (120 mg, 85%). UV–vis (MeCN, 25 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 694 (208). Anal. Calcd for C₁₉H₁₈CuN₄O₃: C 55.13, H 4.38, N 13.54. Found: C 55.01, H 4.36, N 13.36.

L²CuOAc

A solution of HL² (100 mg, 0.29 mmol) in anhydrous MeCN (4 mL) was added dropwise into a 20 mL scintillation vial containing Cu(OAc)₂·H₂O (62.5 mg, 0.31 mmol), anhydrous MeCN (2 mL), and 3 Å molecular sieves (1 cm). The reaction was stirred for 2 h at rt, resulting in a green-blue solution, which was filtered into a clean 20 mL scintillation vial using a 25 mm diameter, 0.2 mm hydrophobic PTFE syringe filter. The solvent was removed in vacuo. The green oil was allowed to stir in hexanes (5 mL) overnight, which was then decanted, and was then triturated with hexanes (2 × 5 mL). The residual solvent was removed in vacuo yielding a green-blue solid (103.4 mg, 77%). Crystalline material of the title product was obtained by vapor diffusion of anhydrous diethyl ether into a concentrated dry THF solution of L²CuOAc at rt, which afforded blue block crystals that were suitable for X-ray diffraction. UV–vis (MeCN, 25 °C) [λ_max, nm (ε, M^–1 cm^–1)]: 694 (292). Anal. Calcd for C₂₃H₂₆CuN₄O₃: C 58.77, H 5.58, N 11.92. Found: C 59.77, H 5.85, N 11.76.

[(L³Cu)₂OH][SbF₆]

A solution of [L³Cu(MeCN)][SbF₆] (74 mg, 0.1 mmol) in ca. 10 mL of CH₂Cl₂ was repeatedly washed with water (ca. 50 mL total volume in 5–10 mL portions) until both layers were clear and the aqueous layer had a neutral pH. Over the course of the treatment, the aqueous washes progressed from turbid pale green and acidic to colorless and neutral, while the organic layer progressed from green to turbid brown to clear cinnamon brown. The organic portion was dried with MgSO₄ and filtered, then selectively precipitated by addition of 10 mL of pentane with mixing. The brown precipitate was allowed to settle and the pale green supernatant was decanted, after which the former was redissolved in minimal CH₂Cl₂ (ca. 6 mL) and the above precipitation was repeated, this time resulting in a virtually colorless supernatant. The precipitate was washed with pentane and dried under vacuum overnight. Crystals suitable for diffractometry were obtained by layering a CH₂Cl₂ solution of the product with pentane and allowing to stand undisturbed for at least 24 h. Despite drying under reduced pressure, product was isolated as a solvated solid with 1 molecule of CH₂Cl₂ per [Cu]. Yield: 34 mg, 54%. UV–vis (CH₂Cl₂, 25 °C): 245 nm (>26000 M^–1 cm^–1), 373 nm (8400 M^–1 cm^–1), 534 nm (300 M^–1 cm^–1), 747 nm (sh, 130 M^–1 cm^–1). Anal. Calcd for C₅₂H₅₇Cu₂F₆N₆O₃Sb + CH₂Cl₂: C 50.45, H 4.71, N 6.66. Found: C 50.20, H 4.58, N 6.59.

L³Cu(OBz)

To a solution of [L³Cu(MeCN)][SbF₆] (74 mg, 0.10 mmol) in 2 mL of CH₃CN under ambient atmosphere was added sodium benzoate (15 mg, 0.11 mmol), and the mixture was sonicated for ca. 5 min, changing color from dull green to bright sea green. After addition of 5 mL of CH₂Cl₂, undesired salts were removed by filtration through a glass fiber pad. The filtrate was evaporated under reduced pressure and redissolved in 2 mL of CH₂Cl₂, followed by selective precipitation with 5 mL of pentane, added gradually. This mixture was filtered through a glass fiber pad, yielding a bright sea green filtrate and a small amount of yellow-green precipitate. The filtrate was evaporated under reduced pressure, redissolved in 5 mL of a 1:1 CH₂Cl₂/heptane mixture, layered with 5 mL of heptane, and allowed to stand while slowly evaporating for 1 week. Bright green crystals (suitable for diffractometry) were decanted, washed with pentane, and dried under vacuum. Yield: 46 mg, 79%. UV–vis (CH₂Cl₂, 25 °C): 245 nm (>29000 M^–1 cm^–1), 365 nm (8000 M^–1 cm^–1), 666 nm (220 M^–1 cm^–1). Anal. Calcd for C₃₃H₃₃CuN₃O₃: C 66.96, H 5.70, N 7.21. Found: C 66.18, H 5.56, N 6.93.

(L³Cu)₂CO₃

To a 25 mL heart-shaped flask under ambient atmosphere were added HL³ (80 mg, 0.20 mmol), anhydrous CuCl₂ (27 mg, 0.20 mmol), and 1.5 mL of MeOH, then mixed until fully dissolved, yielding a dull green solution. At this point, 14 mg (0.20 mmol, 2 equiv) of finely ground K₂CO₃ was added, and the mixture was sonicated for 5 min, resulting in formation of a green precipitate, which was solubilized with addition of 5 mL of CH₂Cl₂. The mixture was sonicated for an additional 15 min, during which time its color changed to a deep vibrant emerald green. Solvent was removed under low pressure, and the product residue was redissolved in CH₂Cl₂ and filtered through a pipet fitted with a glass fiber pad plug, using a total of ca. 6 mL of solvent. The solution was layered with ca. 13 mL of heptane and allowed to stand while slowly evaporating for 5 days. The resulting emerald green crystals (suitable for diffractometry) were decanted and washed with pentane, then dried under vacuum overnight. Despite this drying step, product was isolated as a solvated solid with 1 molecule of CH₂Cl₂ per [Cu]. Yield: 99.3 mg, 93%. UV–vis (CH₂Cl₂, 25 °C): 358 nm (4900 M^–1 cm^–1), 679 nm (280 M^–1 cm^–1). Anal. Calcd for C₅₃H₅₆Cu₂N₆O₅: C 60.67, H 5.47, N 7.86 Found: C 60.62, H 5.56, N 7.82.

This compound also formed spontaneously when metalations involving HL³, Cu(II) salts, and NaOH or NaOMe were attempted under ambient atmosphere.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.2c04432.

• Spectroscopic and cyclic voltammetry data (PDF)

• All X-ray crystallographic data has been submitted to the CCDC. Accession numbers: L³CuCl, 2190092; [(L³Cu)₂OH][SbF₆], 2190093; L³CuOBz, 2190094; [L³CuMeCN][SbF₆], 2190095; (L³Cu)₂CO₃, 2190096; (L²Cu)₂, 2189536; L¹CuCl, 2189539; L²CuCl, 2189535; L¹CuOAc, 2189537; L²CuOAc, 2189538(ZIP)

Author Contributions

C.J.B. and D.Y.S. contributed equally.

The authors declare no competing financial interest.