Sterically Directed Nitronate Complexes of 2,6-Di-tert-Butyl-4-Nitrophenoxide with Cu(II) and Zn(II) and Their H-Atom Transfer Reactivity

Abstract

The bulky 2,6-di-tert-butyl-4-nitrophenolate ligand forms complexes with [Tp^tBuCu^II]⁺ and [Tp^tBuZn^II]⁺ binding via the nitro group in an unusual nitronato-quinone resonance form (Tp^tBu = hydro-tris(3-tert-butyl-pyrazol-1-yl)borate). The Cu complex in the solid state has a five-coordinate κ²-nitronate structure, while the Zn analogue has a four-coordinate κ¹-nitronate ligand. 4-Nitrophenol, without the 2,6-di-tert-butyl substituents, instead binds to [Tp^tBuCu^II]⁺ throught the phenolate oxygen. This difference in binding is very likely due to the steric difficulty of binding a 2,6-di-tert-butyl-phenolate ligand to the [Tp^tBuM^II]⁺ unit. Tp^tBuCu^II(κ²-O₂N^tBu₂C₆H₂O) reacts with the hydroxylamine TEMPO-H (2,2,6,6-tetramethylpiperidin-1-ol) by abstracting a hydrogen atom. This system thus shows an unusual sterically enforced transition metal-ligand binding motif and a copper-phenolate interaction that differs from what is typically observed in biological and chemical catalysis.

Graphical Abstract

Introduction

Chemical reactions between copper(II) and phenols or phenol derivatives are widespread and significant in a variety of processes in chemistry and biology. The enzyme tyrosinase, for example, uses a di-copper containing active site to oxidize tyrosine to dopaquinone in the first biosynthetic step of the production of the pigment melanin.¹ In chemistry, copper catalysts are used for aerobic oxidations of phenols and napthols, for instance in the production of BINOLs (1,1′-bis-2,2′-napthols).²

Many of these reactions involve a copper(II) phenoxide intermediate,^2b in which the copper ion activates the phenoxide ligand through radical delocalization onto the 2, 4, and 6 positions of the ring. This facilitates radical-radical coupling or O₂ addition reactions. Because of the importance of this motif, many examples of copper(II) phenoxide complexes have been reported and their reactivities have been well studied.³

We report here an example of a copper(II) complex that binds 2,6-di-tert-butyl-4-nitro-phenoxide (O₂N^tBu₂C₆H₂O⁻) through the nitro group (nitronate) rather than through the phenoxide oxygen, Tp^tBuCu^II(κ²-O₂N^tBu₂C₆H₂O) (1) [Tp^tBu = hydro-tris(3-tert-butyl-pyrazol-1-yl)borate]. The nitronato ⁻O₂N=R unit binds κ² to the bulky [Tp^tBuCu^II]⁺ fragment, with the ligand better described as a nitronato-quinone rather than as a phenoxide (Scheme 1A). The [Tp^tBuZn^II]⁺ analog is similar except that the nitronate ligand binds κ¹ in this case. The same [Tp^tBuCu^II]⁺ complex binds the unsubstituted 4-nitrophenoxide through the phenolate oxygen. Similarly, Fujisawa, Moro-oka et al. reported some time ago that the slightly less bulky Tp^iPriPrCu unit makes phenoxide complexes with 2,6-di-methylphenoxide and even 2,6-di-tert-butyl-phenoxide ligands, although these species decomposed at ambient temperatures.^3j These results indicate that the unusual nitronate binding mode is the result of steric clash between the Tp^tBu ligand and the tert-butyl groups of the phenol.

Scheme 1.

A. Phenolate and nitronato-quinone resonance forms of ^tBu₂-NO₂-ArO⁻. B. α-Deprotonation of aliphatic nitro compounds.

Transition metal-nitronate complexes have been reported for several different metals.⁴ The nitronate ligands are typically derived from α-deprotonated primary and secondary aliphatic nitro compounds (Scheme 1B). With only a one exception, these bind to metals in a κ² fashion.^4a To our knowledge, the examples reported here are the first transition metal complexes with nitronate ligands derived from a nitrophenol compound, a binding mode that appears to be sterically enforced. In addition to the structure and characterization of these compounds, we describe the ability of the Cu^II-κ²-nitronate-quionone complex to abstract H^• from the hydroxylamine TEMPO–H (2,2,6,6-tetramethylpiperidin-1-ol).

Results and Discussions

Synthesis and Characterization of Tp^tBuCu^II(κ²-O₂N^tBu₂C₆H₂O) (1)

Treatment of the previously described Tp^tBuCu^II-OCH₂CF₃⁵ with one equivalent of 2,6-di-tert-butyl-4-nitro-phenol (O₂N^tBu₂C₆H₂OH) in toluene-d₈ resulted in a rapid colour change from red-orange to dark blue-green. The ¹H NMR spectrum of this reaction displayed new paramagnetically shifted resonances, the disappearance of free O₂N^tBu₂C₆H₂OH and the appearance of free 2,2,2-trifluoroethanol, providing evidence of a protolytic ligand exchange reaction. When performed on a larger scale in benzene, removal of the solvent and 2,2,2-trifluoroethanol in vacuo and crystallization from pentane at −30 °C yielded large green crystals of Tp^tBuCu^II(κ²-O₂N^tBu₂C₆H₂O) (1) (eq 1).

$${\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\text{OCH}}_2{\text{CF}}_3+{\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2\text{OH}\to \underset{\mathbf{(}\mathbf{1}\mathbf{)}}{{\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}({\mathrm{\kappa }}^2-{\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O})}+{\text{HOCH}}_2{\text{CF}}_3$$ (1)

The X-ray crystal structure of (1) (Figure 1) indicates that indeed eq 1 does involve this protolytic displacement, but surprisingly the O₂N^tBu₂C₆H₂O⁻ anion does not bind to the phenolic oxygen. Instead, the phenolate moiety binds κ² to the nitronate oxygens. The structure has a square pyramidal geometry about the Cu ion, with τ = 0.04 where τ = 0 for perfect square monopyramidal vs. τ = 1 for perfect trigonal bipyramidal.⁶ Two of the Tp^tBu ligand pyrazole nitrogens bind to Cu(II) in the same square plane with the nitronato oxygens. The third Tp^tBu pyrazole nitrogen binds axially with a notably longer Cu–N bond distance.

Figure 1.

ORTEP representations of Tp^tBuCu^II(κ²-O₂N^tBu₂C₆H₂O) (1), Tp^tBuZn^II(κ¹-O₂N^tBu₂C₆H₂O) (2) and Tp^tBuCu^II-OC₆H₄NO₂ (3) showing 50% probability thermal ellipsoids and select atom labels. Hydrogen atoms are omitted for clarity.

The bond lengths show that the O₂N^tBu₂C₆H₂O⁻ ligand is well described by the nitronato-quinone resonance form (Table 1, Scheme 1A). The dearomatization of the benzene ring is indicated by the contraction of the C23–C24 and C27–C26 bond distances, avg. 1.3575(6) A, while the other four are lengthened: C24–C25 and C25–C26, avg. 1.481(7) A, and C22–C23 and C22–C27, avg. 1.427(6) A. The C25–O1 bond length of 1.236(7) is consistent with a quinone-like carbonyl bond.⁷

Table 1.

Select Bond Distances (in Å) of Nitrophenol-Derived Ligands in Complexes 1, 2 and 3.

	TptBuCuII(κ2-O2NtBu2C6H2O) (1)	TptBuZnII(κ1-O2NtBu2C6H2O) (2)	TptBuCuII(OC6H4NO2) (3)
O2-N7	1.307(5)	1.3314(13)	1.232(3)
O3-N7	1.293(5)	1.2660(14)	1.221(3)
N7-C22	1.333(6)	1.3441(16)	1.449(3)
C22-C23	1.420(7)	1.4270(17)	1.390(3)
C23-C24	1.360(6)	1.3505(18)	1.379(3)
C24-C25	1.486(7)	1.4854(17)	1.410(3)
C25-C26	1.476(8)	1.4900(17)	1.406(3)
C26-C27	1.355(7)	1.3549(17)	1.371(3)
C27-C22	1.434(7)	1.4287(17)	1.392(3)
C25-O1	1.236(7)	1.2338(16)	1.324(3)

The nitronate C=N bond length of 1.333(6) A in 1 is slightly longer than those of related metal-κ²-O₂N=CR₂ complexes (R = H, CH₃), typically ca. 1.27–1.30 A.⁴ Conversely, the N-O bonds are slightly shorter: avg. 1.300(7) A vs. the typical ca. 1.32–1.34 A.⁴ These deviations are consistent with some π-delocalization from the nitronate group back into the quinoidal substituent.

The X-band EPR spectrum of (1) at 120 K in a toluene glass can be modelled with rhombic g values of g_x = 2.065, g_y = 2.156 and g_z = 2.318 (Fig. 2). Compared to axial copper EPR powder spectra (g_x = g_y), the derivative powder EPR spectrum of (1) has diminished intensity in the high field region indicative of a rhombic spectrum. This has been observed in other five-coordinate copper nitrate compounds.⁸ Additionally, we modelled the diminished intensity as a 0.1 broadening around the g_y value due to g strain. This substantial distribution in the value of g_y is indicative of large structural changes such as interchange between the 5 and 4 coordinate structures (compare (1) and (2) in Figure 1) in solution. The large shifts of all three values away from the free electron g value indicate a d_x2-y2 ground state where the z direction most likely points along the long Cu-N4 axial bond.⁹ The largest copper hyperfine value |A_z^Cu| = 128 G lies in the g_z direction. Nitrogen hyperfine coupling (not included in the simulation) along the g_x direction of |A_z^14N| = 12 G is consistent with coupling to the coordinated pyrazole nitrogens which has been reported for the related Tp^tBuCu^II-OAc complex.¹⁰

Figure 2.

X-Band EPR Spectra of (1) (top) and (3) (bottom) in toluene glasses at 120 K. Data are shown in black and simulations are shown in red.

The EPR spectrum of the structurally related Tp^tBuCu^II-OAc (τ = 0; Fig. S1, Table S4)¹⁰ displays a similar spectrum with slightly larger |A_z^Cu| and |A^N| splitting than observed for (1) (142 vs 128 G and 14 vs 12 G). These results seem consistent with a greater spin delocalization onto the nitronate ligand in (1) that the acetate ligand in Tp^tBuCu^II-OAc.

Synthesis and Characterization of Tp^tBuZn^II(κ¹-O₂N^tBu₂C₆H₂O) (2)

The addition of one equivalent of DBU (1,8-diazabicycloundec-7-ene) to a dichloromethane solution of equimolar Tp^tBuZn^II-OTf and 2,6-di-tert-butyl-4-nitrophenol (O₂N^tBu₂C₆H₂OH) results in the colourless solution turning light yellow (OTf = triflate, SO₃CF₃). Removal of the solvent followed by recrystallization from pentane at −30 °C yielded large yellow crystals of Tp^tBuZn^II-(κ¹-O₂N^tBu₂C₆H₂O) (2), the Zn(II) analogue of (1) (eq 2).

$${\text{Tp}}^{t\text{Bu}}{\text{Zn}}^{\text{II}}\text{-OTf}+{\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2\text{OH}+\text{DBU}\to \underset{\mathbf{(}\mathbf{2}\mathbf{)}}{{\text{Tp}}^{t\text{Bu}}{\text{Zn}}^{\text{II}}({\mathrm{\kappa }}^1-{\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O})}+[\text{DBU-}{\mathrm{H}}^{+}]{\text{OTf}}^{-}$$ (2)

The X-ray structure of (2) displays a distorted tetrahedral Zn(II) center with 3 Zn–N bonds to the Tp^tBu ligand (Figure 1). In the fourth position, the O₂N^tBu₂C₆H₂O⁻ ligand is bound in a similar nitronato-quinone form but through only one of the nitronate oxygens, unlike in (1). The bond lengths of the ligand are similar to those observed in (1) except for the N–O bond distances which are much more asymmetric. The Zn(II) bound nitronate oxygen has a substantially longer N–O bond length than the unbound oxygen (N7–O2 = 1.3314(13) vs 1.2660(14) A, respectively) (Table 1).

The geometry about Zn(II) deviates quite substantially from tetrahedral. A quantitative assessment of this distortion can be made using the pyramidalizion normalization parameter, τ′. (τ′ = [Σ(L_basal-M-L_basal) − Σ(L_basal-M-L_axial)]/90; τ′ = 0 for a perfect tetrahedral and τ′ = 1 for perfect trigonal monopyramidal).¹¹ The τ′ value of (2) is 0.69 and is therefore better described as having trigonal monopyramidal geometry. This contrasts the much more tetrahedral structures reported by Parkin and others for related Tp^tBuZn^II–X complexes, for which τ′ values typically range from 0.3 to 0.4.¹²

Synthesis and Characterization of Tp ^tBuCu^II-OArNO₂ (3)

Analogous to the synthesis of (1), the addition of one equivalent of 4-nitrophenol (HOArNO₂) to Tp^tBuCu^II-OCH₂CF₃ in benzene resulted in a rapid colour change from orange-red to green-brown. After removing the solvent in vacuo, the dark coloured solids were redissolved in diethyl ether and large brown and green dichroic crystals of Tp^tBuCu^II-OArNO₂ (3) were grown from slow evaporation (eq. 3).

$${\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\text{OCH}}_2{\text{CF}}_3+{\text{HOArNO}}_2\to \underset{\mathbf{(}\mathbf{3}\mathbf{)}}{{\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\text{OArNO}}_2}+{\text{HOCH}}_2{\text{CF}}_3$$ (3)

The X-ray structure of (3) shows a distorted tetrahedral geometry about copper (τ′ = 0.62) with 4-nitro-phenoxide bound to the metal center through the phenoxide oxygen (Figure 1, Table 1). The structure is quite similar to the previously reported and related 4-fluoro-phenoxide complex Tp^tBuiPrCu(OC₆H₄F).^3a

The X-band EPR spectrum of (3) in a toluene glass at 120 K was modelled with rhombic g values of g_z = 2.0021, g_y = 2.145, and g_x = 2.320 (Fig. 2). The g_z value nearly equals the free electron g value (g_e = 2.0023) indicating this molecule is in the d_z2 ground state.⁹ The crystal structure reveals a significant distortion of the geometry such that N2 and O1 are nearly trans to each other (≈160°). The direction of the g_z axis (and molecular z axis) is less clear without further experiments for (3); however, the d_z2 ground state would be favoured if z lies along this N2-Cu1-O1 axis making it rotated by ≈ 90° compared to 1. The largest copper hyperfine lies along the g_z axis with a value |A_z^Cu| = 145 G. Additional simulation parameters include copper hyperfine coupling of |A_x^Cu| and |A_y^Cu| < 13 G; however, the exact experimental values are indistinguishable from the inhomogeneous broadening in the spectrum. Again, a large g strain of 0.03 in the g_y value was used for the simulation. Similar EPR spectra have been observed for related 4-coordinate TpCu^II-X complexes with τ′ values < 0.64.^5,13

Attempts to Prepare 2,6-^tBu₂-phenoxyl complexes

In attempts to prepare copper(II) complexes with 2,6-^tBu₂-phenoxide ligands via radical coupling of a copper(I) precursor with stable organic radicals, the previously reported copper(I) dimer [Tp^tBuCu^I]₂¹⁴ was treated with two equivalents of the stable 2,4,6-tri-tert-butylphenoxyl radical (^tBu₃ArO^•)¹⁵ in benzene(-d₆). No reaction was observed by UV/Vis or ¹H NMR spectroscopies (eq 4). The related 4-nitro-phenyl-phenoxyl radical (^tBu₂NPArO^•)¹⁶ also was similarly unreactive. This result is consistent with our previous report showing that treatment of Tp^tBuCu^II-OCH₂CF₃ with ^tBu₂NPArOH yielded ½[Tp^tBuCu^I]₂, HOCH₂CF₃ and ^tBu₂NPArO^• in partial yield with no formation of a Cu^II-nitronate complex observed.⁵

(4)

To prepare the related zinc complex, Tp^tBuZn^II-OTf was reacted with one equivalent of ^tBu₃ArO–H and DBU, analogous to equation 2 above. Upon workup, only Tp^tBuZn^II-OTf was recovered. Similar attempts with Tp^tBuZn^II-Cl and ^tBu₃ArOLi were also unsuccessful.

Spectroscopic and Electrochemical Measurements

The ¹H NMR spectra for all three complexes have equivalent ^tBu groups of their Tp^tBu ligands, indicating that they are fluxional on the NMR timescale. For the paramagnetic Cu^II complex (1), the resonances due to these ^tBu groups and the ^tBu groups from the nitronate ligand were the only signals observed and displayed the expected 3:2 peak integration. For complex (3), only the ^tBu groups from the Tp^tBu ligand were observed (Supporting Information Fig S4, S5). The optical spectrum of compound (1) displays a weak optical absorption at 796 nm (ε = 180 M⁻¹ cm⁻¹) while compound (3) has absorptions at 583 nm (ε = 430 M⁻¹ cm⁻¹), 667 nm (ε = 400 M⁻¹ cm⁻¹) and 830 nm (ε = 130 M⁻¹ cm⁻¹) (Supporting information Fig S2, S3).

The cyclic voltammogram (CV) of (1) in dichloromethane with 0.1 M [ⁿBu₄N][PF₆] (Supporting information fig. S6) shows irreversible electrochemical responses at both oxidizing and reducing potentials. The cathodic peak current, E_c,a, occurs at −0.84 V while the anodic peak current, E_p,a occurs at 1.10 V (both vs Fc^+/0 and at 100 mV/s scan rates). Similar electrochemical irreversibility has been reported for related Tp^tBuCu^II-X complexes.^5,14

Reactivity with H-Atom Transfer Reagents

Combining 2.5 mM [Tp^tBuCu^I]₂ with 2 equivalents of O₂N^tBu₂C₆H₂OH in benzene-d₆ or toluene-d₈ showed no reaction as determined by ¹H NMR. When one equivalent of the hydrogen atom acceptor, ^tBu₃ArO^• was added to the same solution, a colour change from blue (^tBu₃ArO^•) to green-brown occurred over roughly one hour. The optical trace of this reaction (Figure 3) displays clear isosbestic points and shows that (1) is generated in approximately 60% yield. Similarly, monitoring a related reaction by ¹H NMR showed 50% yield of (1) as well as signals for ^tBu₃ArO–H (~80%) and unreacted O₂N^tBu₂C₆H₂O–H (~40%). All of the ^tBu₃ArO^• is consumed during this reaction as determined from the optical spectrum of the completed reaction.

Figure 3.

Schematic of the reaction between 2.5 mM [Tp^tBuCu^I]₂, 2 equivalents of O₂N^tBu₂C₆H₂O–H and two equivalents of ^tBu₃ArO^• in benzene and the optical traces of the reaction carried out over the course of 1 hour. The blue arrow shows the disappearance of ^tBu₃ArO^• with time while the green arrow shows the appearance of 1.

The reaction of (1) with the hydrogen atom donor TEMPO-H, each 12 mM in benzene, resulted in a colour change from green-brown to light pink over the course of an hour. ¹H NMR spectra showed quantitative formation of O₂N^tBu₂C₆H₂OH and [Tp^tBuCu^I]₂. The optical spectrum of the same reaction showed a small absorption indicative of TEMPO radical at 460 nm, in essentially quantitative yield (Σ₄₆₀ = 11.1 M⁻¹ cm^{−1 17}). The balanced reaction is therefore as described by equation 5.

(5)

Discussion

Sterically Directed Phenolate/Nitronate Linkage

The formation of the κ²-nitronate complex (1) from the 2,6-di-t-butyl-4-nitrophenol contrasts with the formation of the O-bond phenoxide complex (3) from the unhindered 4-nitrophenol. These results indicate that the preference for the nitronate coordination in (1) and (2) is the result of the steric clash between Tp^tBu tert-butyl groups and the 2,6-di-tert-butyl groups of the phenol. This is further supported by the inability to prepare related 2,6-di-tert-butyl-phenoxyl complexes by ligand metathesis from [Tp^tBuZn^II] precursors or from [Tp^tBuCu^I]₂ and bulky phenoxyl radicals (eq 4).⁵

The reaction of [Tp^tBuCu^I]₂ with O₂N^tBu₂C₆H₂O–H and ^tBu₃ArO^• to give (1) in 50–60% yield likely proceeds by initial equilibration of the radicals, reaction 5. Based on the known O–H BDFEs for the two phenols in benzene,¹⁸ the equilibrium constant K₅ = 3 × 10⁻³ (eq 6). The self-exchange rate constant for ^tBu₃ArO^• + ^tBu₃ArOH is 95 M⁻¹ s⁻¹ in benzene,¹⁹ indicating that reaction 6 will be at equilibrium, with ca. 5% of the O₂N^tBu₂C₆H₂O^• radical. This transiently stable radical is known to have a lifetime suitable for EPR studies but irreversibly decomposes in minutes.²⁰ This suggests that [Tp^tBuCu^I]₂ can react with the less sterically crowded nitro end of the O₂N^tBu₂C₆H₂O^• radical. This is consistent with the lack of reactivity with the oxygen end and with ^tBu₃C₆H₂O^• being due to sterics.

$${\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}-\mathrm{H}+{}^t\mathrm{B}{\mathrm{u}}_3{\text{ArO}}^{•}\rightleftharpoons {\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2{\mathrm{O}}^{•}+{}^t\mathrm{B}{\mathrm{u}}_3\text{ArO}-\mathrm{H}$$ (6)

H-Atom Transfer Reactivity and Thermochemistry

Treatment of (1) with the hydrogen atom donor TEMPO–H results in net H⁺/e⁻ transfer produces O₂N^tBu₂C₆H₂O–H, ½[Tp^tBuCu^I]₂, and TEMPO in quantitative yields. This reaction could proceed via several possible pathways, including concerted proton-electron transfer (CPET),²¹ protolytic ligand exchange followed by TEMPO dissociation, or phenoxyl radical dissociation. Initial electron transfer (ET) or proton transfer (PT) are unlikely given the properties of (1) and TEMPO–H, following standard thermochemical arguments.^18b

Regardless of mechanism, the well-defined stoichiometry of this reaction provides thermochemical information about this system. Because this reaction proceeds quantitatively, it can be concluded that ΔG < 0 under these conditions. Since this reaction was carried out at concentrations much lower than standard state, entropy is expected to provide 1.3 kcal mol⁻¹ to ΔG here relative to standard state free energy, ΔG°. A limit for ΔG° of this reaction is thus expected to be ΔG° ≤ 1.3 kcal mol⁻¹. With this limit, an ‘effective bond dissociation free energy (BDFE)²²’of [Tp^tBuCu^I]₂ + O₂N^tBu₂C₆H₂O–H can be calculated using the known O–H bond strength of TEMPO–H (eq 7–9).

$${\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}+\text{TEMPO}-\mathrm{H}\to ½{[{\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\mathrm{I}}]}_2+{\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}-\mathrm{H}+\text{TEMPO}\mathrm{\Delta }G°\le 1.3\text{kcal}{\text{mol}}^{-1}$$ (7)

$$\text{TEMPO}+{\mathrm{H}}^{•}\to \text{TEMPO}-\mathrm{H}\mathrm{\Delta }G°=-65.2\text{kcal}{\text{mol}}^{-118\mathrm{b}}$$ (8)

Summing (eq 7) and (eq 8) gives:

$${\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}+{\mathrm{H}}^{•}\to ½{[{\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\mathrm{I}}]}_2+{\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}-\mathrm{H}\mathrm{\Delta }G°\le -63.2\text{kcal}{\text{mol}}^{-1}$$ (9)

Since the O–H bond strength of O₂N^tBu₂C₆H₂O–H is known,¹⁶ this ‘effective BDFE’ limit can be used to extrapolate a thermodynamic limit for homolytically cleaving the Cu^II–κ²-O₂N=R chelating bond and forming ½ [Tp^tBuCu^I]₂.

$${\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}+{\mathrm{H}}^{•}\to ½{[{\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\mathrm{I}}]}_2+{\mathrm{O}}_2\mathrm{N}{}^t\mathrm{B}{\mathrm{u}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}-\mathrm{H}\mathrm{\Delta }G°\le -63.2\text{kcal}{\text{mol}}^{-1}$$ (9)

$${\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}-\mathrm{H}\to {\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2{\mathrm{O}}^{•}+{}^{•}\mathrm{H}\mathrm{\Delta }G°=80.4\text{kcal}{\text{mol}}^{-116}$$ (10)

Summing (eq 9) and (eq 10) gives:

$${\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\text{II}}-{\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2\mathrm{O}\to ½[{\text{Tp}}^{t\text{Bu}}{\text{Cu}}^{\mathrm{I}}]+{\mathrm{O}}_2{\mathrm{N}}^t{\text{Bu}}_2{\mathrm{C}}_6{\mathrm{H}}_2{\mathrm{O}}^{•}\mathrm{\Delta }G°\le 16.5\pm 2\text{kcal}{\text{mol}}^{-1}$$ (11)

It is important to note that this value encompasses both the free energy cost of homolytically cleaving the copper(II)-nitronate chelate as well as the free energy gained from the formation of the [Tp^tBuCu^I]₂ dimer. Dimer formation is likely a substantial free energy contribution to equation 11, since it was reported to be stable with respect to dissociation even at 362 K.¹⁴

The thermochemical analysis, even with these caveats, suggests that the loss of phenoxyl radical (O₂N^tBu₂C₆H₂O^•) from 1 to form the copper dimer is not so endoergic. With the 2,4,6-tri-tert-butyl-phenoxyl radical that can only ligate through the phenoxide oxygen, the reaction analogous to reaction (11) actually favors the free phenoxyl radical (ΔG° ≤ 0, eq 4).

The free radical is also more stable than the phenoxide or nitronate complex for the 4-nitrophenyl-2,6-di-tert-butyl phenoxyl radical¹⁶ (eq 4). In this case, the energetic cost to form the nitronate is evidently too large to de-aromatize two aromatic rings required for the quinonate structure.

This energetic analysis also provides a rationale for the facile decomposition of the Tp^iPriPrCu-2,6-di-methyl-phenoxide and 2,6-di-tert-butyl-phenoxide complexes, as reported by Fujisawa, Moro-oka, et al.^3j The decomposition products that they identified are characteristic of phenoxyl radicals.

Conclusions

In the system described here, the steric bulk surrounding [Tp^tBuCu^II]⁺ and [Tp^tBuZn^II]⁺ units dictates the binding mode of nitrophenoxide ligand O₂N^tBu₂C₆H₂O⁻. The 2,6-di-tert-butyl groups prevent coordination via the phenoxide oxygen, as indicated by unsuccessful attempts to prepare other 2,6-ditert-butyl-phenoxide complexes. The Fujisawa, Moro-oka, et al. 2,6-di-tert-butyl-phenoxide complex with the slightly smaller di-iso-propyl-Tp ligand appears to be just on the cusp of thermochemical stability.^3j The 4-nitro-2,6-di-tert-butylphenoxide ligand has another alternative, to bind as a nitronate with a quinonoidal electronic structure. This binding mode is sterically much less crowded. In the absence of crowding, as in the simple 4-nitrophenolate, binding through the phenolate oxygen to [Tp^tBuCu^II]⁺ is observed.

Tp^tBuCu^II(κ²-O₂N^tBu₂C₆H₂O (1) cleanly abstracts a hydrogen atom from the hydroxylamine TEMPO–H to form the Cu(I) dimer ½ [Tp^tBuCu^I]₂, O²N^tBu₂C₆H₂OH and the nitroxyl radical TEMPO. Thermochemical analysis of this reaction provides limits for an ‘effective BDFE’ in this system and for loss of the phenoxyl radical from (1). This analysis provides a rationale for the copper-phenoxyl chemistry observed, including the new and unusual sterically directed nitronate ligation mode of a bulky nitrophenoxide that differs from what is typically observed in biological and synthetic copper-phenol chemistries.

Experimental

Materials

Unless otherwise noted, all reactions were carried out in a nitrogen filled glovebox. All materials were purchased from Sigma-Aldrich and used without purification. Benzene-d₆ and toluene-d₈ were purchased from Cambridge Isotope Labs and dried over NaK prior to being vacuum distilled. Dichloromethane-d₂ was also purchased from Cambridge Isotope Labs and dried over CaH₂ prior to being vacuum distilled.

Other solvents were purchased from Fischer and dried using a “Grubbs type” Seca Solvent System installed by Glass-Contour. Tp^tBuCu^II-OCH₂CF₃,⁵ [Tp^tBuCu^I]₂,¹⁴ Tp^tBuZn^II-OTf,²³ Tp^tBuZn^II-Cl,^{24 t}Bu₂-NO₂-ArO–H,^{25 t}Bu₃ArO^•,^{15 t}Bu₂NPArO^•,¹⁶ and TEMPO–H²⁶ were prepared following published synthetic protocols. All glassware was dried in an oven at 150 °C overnight and pumped into a nitrogen filled glovebox while still hot.

Instrumentation

All ¹H NMR spectra were collected on Bruker 300 and 500 MHz instruments. ¹H chemical shifts were referenced to TMS using the residual solvent peak. UV-visible absorption spectra were collected with a Hewlett-Packard 8453 diode array spectrometer and are reported as λ_max in nm (ε in M⁻¹ cm⁻¹; uncertainties of ca. 10%). Cyclic voltammetry (CV) was performed using a CH instrument 600D potentiostat and the irreversible potentials given are ±0.05 V. EPR spectra were collected with a Bruker EMX X-band spectrometer and simulations were performed using EasySpin.²⁷ g values are reported with ±0.005 uncertainty. CHN elemental analysis was performed by Atlantic Microlabs, Inc., Norcross, GA.

Syntheses

hydro-tris(3-tert-butylpyrazol-1-yl)borate) copper(II) 3,5-bis(tert-butyl)-4-benzenone-κ²-nitronate, Tp^tBuCu^II(κ₂-O₂N^t Bu₂C₆H₂O (1)

To a ~5 mL solution of Tp^tBuCu^II-OCH₂CF₃ (58.7 mg, 0.11 mmoles) in benzene was added ^tBu₂-NO₂-ArOH (27.1 mg, 0.11 mmoles) with stirring. The solution quickly changed in colour from orange to blue-green. Removal of the solvent in vacuo gave dark green solid. Crystallization from pentane at −30 °C yielded 50 mg of large X-ray quality crystals (66 %). ¹H NMR (500 MHz, benzene-d₆) δ: 2.8 (broad, 27H), 1.0 (s, 18H); UV/Vis (toluene): 830 nm (180); Anal. Calcd. For C₃₅H₅CuN₇O₃: C, 60.47; H, 7.83; N, 14.01. Found: C, 60.74; H, 7.89; N, 14.16.

hydro-tris(3-tert-butylpyrazol-1-yl)borate) zinc(II) 3,5-bis(tert-butyl)-4-benzenone-κ¹-nitronate, Tp ^tBuZn^II(κ¹-O₂N^t Bu₂-C₆H₂O (2)

To a ~5 mL dichloromethane solution of Tp^tBuZn^II-OTf (101.3 mg, 0.17 mmoles) was added dropwise a ~2 mL dichloromethane solution of DBU (25.9 mg, 0.17 mmoles) and 2,6-di-tert-butyl-4-nitro phenol (42.7 mg, 0.17 mmoles) with stirring. Upon addition, the solution turned from colourless to yellow. After 10 minutes the solvent was removed in vacuo and the remaining solid was redissolved in a minimum amount of pentane and filtered. Large yellow X-ray quality crystals grew upon standing at −30 °C (73.1 mg, 62%). ¹H NMR (500 MHz, CD₂Cl₂) δ: 7.74 (s, 2H), 7.61 (d, ³J_HH = 2.3 Hz, 3H), 6.12 (d, ³J_HH = 2.3 Hz, 3H), 1.33 (s, 27H), 1.31 (s, 18H). Anal. Calcd. For C₃₅H₅₄BN₇O₃Zn: C, 60.31; H, 7.81; N, 14.07. Found: C, 60.42; H, 7.59; N, 14.20.

hydro-tris(3-tert-butypyrazol-1-yl)borate) copper(II) 4-nitrophenolate (3)

To a ~5 mL dichloromethane solution of Tp^tBuCu^II-OCH₂CF₃ (62.2 mg, 0.114 mmoles) was added a ~2 mL dichloromethane solution of 4-nitro-phenol (15.9 mg, 0.114 mmoles). Upon addition, the solution turned from orange to brown-green. After the solvent was removed in vacuo, the resulting solids were redissolved in diethyl ether, filtered over Celite and crystallized by slow evaporation yielding green/brown dichroic crystals of (3) (65.5 mg, 98%). ¹H NMR (300 MHz, CD₂Cl₂) δ: 7.6 (broad, 27H). UV/Vis (toluene): 583 (430), 667 (400), 830 (130). Anal. Calcd. For C₂₇H₃₈BCuN₇O₃: C, 55.63; H, 6.57; N, 16.82. Found: C, 55.44; H, 6.56; N, 16.85.

Reactions

Representative NMR Reaction Procedure

A TEMPO–H solution (300 mL, 24 mM, benzene-d₆) was added dropwise to a solution of Tp^tBuCu^II-O₂N^tBu₂C₆H₂O (300 mL, 24 mM, benzene-d₆) with stirring. The resulting solution was transferred to a J. Young NMR tube and stored at room temperature in the dark until the reaction had reached completion.

Representative UV/Vis Procedure

To a septum-capped 1 cm path length quartz cuvette containing a 2.5 mL benzene solution of [Tp^tBuCu^I]₂ (3 mM) and ^tBu₂NO₂ArO–H (6 mM) was added a 0.5 mL benzene solution of ^tBu₃ArO^• (30 mM) from a gas-tight syringe (Final concentrations 2.5, 5, and 5 mM for [Tp^tBuCu^I]₂, ^tBu₂NO₂ArO–H, and ^tBu₃ArO^•, respectively). The cuvette was inverted several times prior to beginning optical measurements to ensure complete mixing had occurred.

Attempts to prepare “Tp^tBuZn^II-OAr^tBu₃”

Attempt 1

To a ~5 mL dichloromethane solution of Tp^tBuZn^II-OTf (73.1 mg, 0.11 mmoles) was added dropwise a ~2 mL dichloromethane solution of DBU (18.6 mg, 0.11 mmoles) and 2,4,6-tri-tert-butyl phenol (32.0 mg, 0.11 mmoles) with stirring. After stirring for one hour, the solvent was removed in vacuo, the resulting solids taken up in pentane, filtered over Celite. After removal of the solvent, the crude white solid remaining was identified as the starting Zn complex, Tp^tBuZn^II-Otf, by ¹H NMR.

Attempt 2

LiOAr^tBu₃•THF was prepared in situ by the addition of one equivalent of lithium bis(trimethylsilyl)amide (LiHMDS) (22.2 mg, 0.13 mmoles) to ^tBu₃ArOH (34.8 mg, 0.13 mmol) in ~3 mL THF. Addition of this solution to a ~5 mL THF solution containing Tp^tBuZn^II-Cl (64.4 mg, 0.13 mmol) was carried out with stirring. After 1 hour, the solvent was removed in vacuo, the solids were taken up in toluene, filtered over Celite, and crystallized at −30 °C. The resulting colourless crystals were found to be LiOAr^tBu₃^•THF and the mother liquor from crystallization was found to contain Tp^tBuZn^II-Cl and LiOAr^tBu₃^•THF as determined by ¹H NMR.²⁸