Involvement of a Formally Copper(III) Nitrite Complex in Proton-Coupled Electron Transfer and Nitration of Phenols

Abstract

A unique high-valent copper nitrite species, LCuNO₂, was accessed via the reversible one-electron oxidation of [M][LCuNO₂] (M = NBu₄⁺ or PPN⁺). The complex LCuNO₂ reacts with 2,4,6-tri-tert-butylphenol via a typical proton-coupled electron transfer (PCET) to yield LCuTHF and the 2,4,6-tri-tert-butylphenoxyl radical. The reaction between LCuNO₂ and 2,4-di-tert-butylphenol was more complicated. It yielded two products: the coupled bisphenol product expected from a H-atom abstraction and 2,4-di-tert-butyl-6-nitrophenol, the product of an unusual anaerobic nitration. Various mechanisms for the latter transformation were considered.

Graphical Abstract

∎ INTRODUCTION

Copper nitrite complexes have been studied extensively as models for the active sites of copper-containing nitrite reductases (CuNIRs),^1–25 key enzymes in global denitrification and mammalian and plant signaling pathways.²⁶ These model complexes contain Cu^I or Cu^II, often in coordination geometries relevant to the CuNIR active site, and they typically reduce NO₂⁻ to NO, a reaction also performed by the enzymes. It has been proposed for CuNIRs that one-electron reduction of NO₂⁻ to NO involves an initial proton transfer facilitated by neighboring amino acid residues.^27–31 Inspired by this notion, recent work has probed the ability of copper(II) nitrite complexes to perform proton-coupled electron transfer (PCET).^10,19 Notably, PCET was invoked in the reaction of a copper(II) nitrite complex with 2,4-di-tert-butylphenol (DTBP) that underwent subsequent nitration, an unusual anaerobic transformation.¹⁹ Such a process is relevant to tyrosine nitration, which normally occurs via an attack by peroxynitrite.^32–34

Considering that higher-oxidation-state species generally exhibit enhanced PCET reactivity,^35–38 we sought to prepare a complex with a [CuNO₂]²⁺ core (formally containing Cu^III) that would represent a rare example of a high-valent metal nitrite species.^39–41 Specifically, previous success in using the hindered dianionic ligand bis(2,6-diisopropylphenylcarboxamido)-pyridine (L²⁻) to prepare reactive complexes with cores [CuX]²⁺ (X = OR,^42–47 SR,⁴⁸ OOR,⁴⁹ O₂CR,^50,51 F, Cl, and Br⁵²) led us to target LCuNO₂. Herein, we describe the successful generation and spectroscopic characterization of this novel high-valent metal nitrite species, as well as preliminary studies of PCET reactivity with phenolic substrates that result in an unusual anaerobic nitration.

■ RESULTS AND DISCUSSION

Synthesis and Characterization of [M][LCuNO₂] (M = NBu₄⁺ or PPN⁺).

The addition of the corresponding nitrite salt [M][NO₂] to LCu(MeCN) (MeCN = acetonitrile) in tetrahydrofuran (THF) resulted in the formation of copper(II) complexes [M][LCuNO₂] (M = NBu₄⁺ or PPN⁺), which were isolated as powder blue solids and characterized by elemental analysis, UV–vis and electron paramagnetic resonance (EPR) spectroscopy, and X-ray crystallography for M = PPN⁺ (Scheme 1). The X-ray structure of [PPN][LCuNO₂] shows nitrite bound strongly to the Cu ion through one O atom [Cu1–O3 = 1.9667(13) Å] with an additional weak interaction through the second O atom [Cu1–O4 = 2.4622(14) Å; Figure 1]. This binding geometry is similar to that observed for carboxylate ligands in the series [NBu₄][LCu(O₂CR)],^50,51 and the N–O bond distances are comparable to those in other complexes containing Cu^II(η¹-ONO) cores.^{3,13–15,17,19,24,51,53–59} The Cu ion adopts a square-planar geometry (τ₄ = 0.16),⁶⁰ and the Cu1–N1, Cu1–N2, and Cu1–N3 bond lengths are similar to those in the previously reported [LCuX]⁻ complexes.^{42,44,46–52}

Scheme 1. Copper Complexes Discussed in This Work, with Formal Oxidation States Indicated.

Figure 1.

X-ray crystal structure of [PPN][LCuNO₂] with the PPN cation and H atoms omitted for clarity. All non-H atoms are shown as 30% thermal ellipsoids. Selected bond distances (Å) and angles (deg): Cu1–O3, 1.9667(13); Cu1–O4, 2.4622(14); Cu1–N1, 1.9987(13); Cu1–N2, 1.9242(14); Cu1–N3, 1.9985(13); N1–Cu1–N3, 160.25(6); N1–Cu1–N2, 80.26(5); N2–Cu1–N3, 80.48(6); N2–Cu1–O3, 178.01(6).

The X-band EPR spectra for [M][LCuNO₂] (M = NBu₄⁺ or PPN⁺) in THF at 30 K are nearly identical and exhibit typical signals for S = ¹/₂ square-planar copper(II) complexes (Figures 2 and S1). Spectral parameters were estimated by simulation, with the best match to the 17-line experimental superhyperfine pattern resulting when only three N atoms were included (Table S1). From these results and the similarity of the spectrum to those of other [LCuX]⁻ complexes,^{42,44,46–52} we conclude that little spin density is present on the nitrite N atom and that the superhyperfine coupling arises from interactions with the N atoms of L²⁻. The UV–vis spectra for [M][LCuNO₂] (M = NBu₄⁺ or PPN⁺) in THF show typical d–d transitions (λ ~ 586 nm; ε ~ 480 M⁻¹ cm⁻¹; Figure S2). Cyclic voltammetry for [NBu₄][LCuNO₂] in THF (0.3 M [NBu₄][PF₆]) revealed a pseudoreversible wave with E_1/2 = +180 mV versus ferrocene/ferrocenium (Fc/Fc⁺; linear plot of i_pa vs ν^1/2; Figure S3). This oxidation potential is similar to those measured for the [LCu(O₂CR)]^−/0 series (range = 150–298 mV)⁵¹ and 347 mV higher than that for [LCuOH]^−/0,⁴⁷ all which track inversely with the basicity of the anionic “X” ligands (greater basicity, lower potential; Table S2).

Figure 2.

X-band EPR spectrum of [NBu₄][LCuNO₂] in THF (black, experimental; red, simulation). Parameters: temperature 30 K; microwave frequency 9.38 GHz; microwave power 0.0002 mW; modulation amplitude 9.8 G; modulation frequency 100 kHz. Parameters from the simulation are listed in Table S1.

Synthesis and Characterization of LCuNO₂.

The addition of 1 equiv of acetylferrocenium tetrakis(3,5-bis-(trifluoromethyl)phenyl)borate ([AcFc][BArF₂₄]) to a solution of [NBu₄][LCuNO₂] in THF at −80 °C resulted in the immediate development of a deep Prussian blue color and the appearance of intense features at 478 nm (ε = 5350 M⁻¹cm⁻¹), 655 nm (ε = 9170 M⁻¹ cm⁻¹), and 816 nm (ε = 7180 M⁻¹ cm⁻¹) in the UV–vis spectrum (Figure 3a). Variation of the amount of [AcFc]⁺ between 0.2 and 1.8 equiv showed the attainment of maximum absorbance for the new features when 1 equiv of [AcFc]⁺ was added (Figures 3b and S4). Also, the addition of 1 equiv of decamethylferrocene (Fc*) bleached the solution to yield the spectrum of [LCuNO₂]⁻, a process that could be repeated (2 times; Figure S5). Taken together, the evidence supports reversible one-electron oxidation of [NBu₄][LCuNO₂].

Figure 3.

(a) Overlay of UV–vis spectra of [NBu₄][LCuNO₂] (blue) and the species generated upon the addition of 1 equiv of [AcFc][BArF₂₄] to [NBu₄][LCuNO₂] (red). Conditions: −80 °C in THF. (b) Plot of corresponding molar absorptivity values at λ = 655 nm versus equivalents of [AcFc][BArF₂₄] added to [NBu₄][LCuNO₂] at −80 °C in THF.

The new UV–vis features that appear upon oxidation of [NBu₄][LCuNO₂] are similar to those seen for other LCuX complexes and bear particular resemblance to those found for the LCuO₂CR series.⁵¹ These features were assigned using time-dependent density functional theory to ligand-to-metal charge transfer (LMCT) involving N-aryl π → Cu d and N-amide π → Cu d transitions.⁵⁰ Consistent with these assignments, the energy of these transitions is inversely proportional to the electron-withdrawing characteristics of the carboxylate ligand. Thus, for example, LCuO₂CC₆F₅ with the most electron-withdrawing carboxylate and the lowest carboxylic acid aqueous pK_a of 1.48⁶¹ yields the most electrophilic Cu ion and the lowest-energy feature at 866 nm, whereas LCuO₂CCH₃ with its absorption feature at 809 nm has the least electrophilic Cu ion in the series with a carboxylic acid aqueous pK_a of 4.8.⁵¹ The lowest-energy peak for the product of one-electron oxidation of [NBu₄][LCuNO₂] has λ_max = 816 nm, intermediate in the LCuO₂CR series, consistent with a nitrous acid aqueous pK_a of 3.16, and in line with an analogous LMCT assignment.⁶² Finally, a peak at 634 cm⁻¹ was observed in the resonance Raman spectrum of LCuNO₂ (λ_ex = 660 nm), which we assign as ν(Cu–O) based on nearly identical peaks present in previously measured spectra for LCuX with X = O-based ligands (Figure S6).^47,49 Taken together, the UV–vis and resonance Raman spectra and titration/stoichiometry/reversibility data support formation of LCuNO₂ upon one-electron oxidation of [NBu₄][LCuNO₂].

PCET Reactivity.

Treatment of LCuNO₂ with 2,4,6-tri-tert-butylphenol (TTBP; 50 equiv) in THF at −80 °C led to decay of the absorptions associated with LCuNO₂. This decay was monitored over ~40 min, and a global fit of the decay spectra to a second-order reaction model using ReactLab Kinetics⁶³ yielded a k₂ of 1(1) × 10⁻¹ M⁻¹ s⁻¹. The UV–vis spectrum of the product solution (Figure S7) indicated formation of the 2,4,6-tri-tert-butylphenoxyl radical (characteristic peaks around 400 and 660 nm)^64,65 and LCu(THF) (d–d transition at 570 nm).⁵⁰ The radical was also identified by EPR spectroscopy, and from integration, a yield of 61% was determined (Figure S8). In a separate experiment designed to detect possible coproduct NO, a solution of TTBP (50 equiv) was added to a solution of LCuNO₂ at −40 °C and allowed to warm to room temperature in the presence of a solution of CoTPP (TPP = 5,10,15,20-tetraphenyl-21H,23H-porphine).^20,40,66,67 Subsequent analysis of the latter solution by UV–vis spectroscopy revealed the formation of (NO)CoTPP in an amount corresponding to ~15% yield of NO from the PCET reaction (Figure S9). We presume that HNO₂ also forms in the PCET reaction but decays via unidentified processes, which might also lead to NO. An alternative PCET pathway (analogous to the one previously identified in a reaction of the Cu^II-NO₂ complex)¹⁹ would involve the initial formation of NO and LCu^IIIOH, but the latter would be expected to react further with TTBP and generate an additional 1 equiv of the phenoxyl radical (Scheme S1). The observed low yield of the radical and the low yield of NO argue against this pathway being the dominant H-atom-transfer step.^68,69

The rate constant found for the reaction of TTBP with LCuNO₂ is similar to that for the reaction with LCuO₂CC₆H₄(Cl) under the same conditions [k₂ = 3(1) × 10⁻¹ M⁻¹ s⁻¹],⁵⁰ but it is ~100 times smaller than that for the reaction with LCuOH [k₂ = 2(1) × 10¹ M⁻¹ s⁻¹].⁵⁰ These rate constants are in line with thermodynamic considerations, particularly the E_1/2 and pk_a values. Thus, the [LCuNO₂]^−/0 E_1/2 of 180 mV versus Fc/Fc⁺ and the pK_a of HNO₂ of 3.16⁶² fall close to the corresponding E_1/2 and carboxylic acid pK_a values for the series [LCuO₂CR]^−/0, and the former E_1/2 values correlate with the log k₂ values for the reaction with TTBP (Figure S10).⁵¹ These relationships support similar driving forces for the PCET reactions. Likewise, the greater basicity of LCuOH that results in the formation of a stronger O–H bond underlies its faster PCET reactions.⁴³

Nitration Reactivity.

UV–vis monitoring of the reactions between LCuNO₂ and varying amounts of DTBP (1–60 equiv) in THF at −40 °C resulted in t_1/2 values of 613, 318, 252, and 176 s in the presence of 1, 20, 40, and 60 equiv of substrate, respectively (Figure S11). While the trend in the t_1/2 values indicates a dependence of the rate on the DTBP concentration, the decay data could not be fit to simple kinetic models (i.e., pseudo first order; Figure S11c,f,i,l). These findings suggest that a more complicated reaction occurs with DTBP compared to TTBP. Product analysis was performed in an attempt to understand the results from the kinetic experiments. A peak at 570 nm in the final UV–vis spectrum suggests formation of LCu(THF). To identify the organic products, reactions were performed on a larger scale (2.67 mM) at −40 °C in THF using 0.5 or 1 equiv of DTBP for 2 h or 10 equiv of DTBP for 1 h, and the residues were analyzed by ¹H NMR spectroscopy (Figures S12–S14). Two products were identified: the coupled bisphenol product, 3,3′,5,5′-tetra-tert-butyl-[1,1′-biphenyl]-2,2′-diol, and 2,4-di-tert-butyl-6-nitrophenol (Scheme 2). The yields of bisphenol and 2,4-di-tert-butyl-6-nitrophenol and the amount of unreacted DTBP varied with differing equivalents of DTBP (Table 1). The data show that the yield of nitrated product and the conversion of substrate increase in the reactions with fewer equivalents of DTBP used. We interpret these results (greater nitration when LCuNO₂ is in excess) to indicate that nitration involves multiple equivalents of LCuNO₂.

Scheme 2. Observed Reactants and Products When LCuNO₂ Is Reacted with DTBP.

Table 1.

Results from Product Analysis of the Reactions between LCuNO₂ and Various Equivalents of DTBP^a

equiv	nitrated phenol % yield	bisphenol % yieldb	% of unreacted phenol	mass balancec (%)	% of phenol convertedd
10	0.5	8.2	94	102.7	8.5
	0.7	8.8	95	104.5	9.1
1	15	42	45	102	56
	17	43	42	102	59
0.5	48	26	30	104	71
	46	24	28	98	71

^aAll data were from ¹H NMR spectral integrations using integration of the trimethoxybenzene peak at 6.09 ppm (3 protons) as a standard. All values are based on the DTBP loading.

^bValues take into consideration that 2 mol of DTBP is required to form 1 mol of bisphenol.

^cMass balance values over 100% are due to the standard error in the ¹H NMR integration values.

^dCalculated by the equation (nitrated phenol % yield + bisphenol % yield)/mass balance.

The product 2,4-di-tert-butyl-6-nitrophenol could be formed from the reaction between LCuNO₂ and DTBP via multiple possible mechanisms. LCuNO₂ could abstract a H atom from DTBP, resulting in a [LCuHNO₂] adduct (which presumably decays to LCuTHF and HNO₂) and the phenoxyl radical of DTBP, which could react with another 1 equiv of LCuNO₂ in a “rebound” type of reaction to form LCuTHF and a nitrated phenol product (Scheme 3, path A). There is precedence to support the “rebound” pathway given that LCuF was reported to functionalize substrates via H-atom abstraction and radical capture, or “rebound”.⁵² Conversely, multiple pathways where free NO₂ is liberated are possible (Scheme 3), and free NO₂ has been studied for its oxidation and nitration chemistry with DTBP.⁷⁰ If LCuNO₂ does disproportionate into LCuTHF and free NO₂, then this is an unusual transformation of mildly oxidizing nitrite to NO₂. The results of the stoichiometry experiments are taken as evidence for Scheme 3, with path A being the predominant nitration mechanism, because the dependence of the nitrated product yield on the stoichiometry of LCuNO₂ suggests that limiting consumption of LCuNO₂ by PCET favors nitration.

Scheme 3. Possible Mechanisms of Phenol Nitration by LCuNO₂.

LCuNO₂ could be involved in generation of the phenoxyl radical of DTBP via a nucleophilic attack of the phenol on the nitrite ligand of LCuNO₂, as previously seen for a Cu^II-NO₂ compound (Scheme S2).^66,71 The products of the nucleophilic attack would be O-nitrosated DTBP and LCu^IIIOH, which would react with another 1 equiv of DTBP to form bisphenol. The O-nitrosated DTBP would then release NO and the phenoxyl radical of DTBP, which could then combine to form 2,4-di-tert-butyl-6-nitrophenol.⁷¹ The stoichiometry experiments do not support this nitration mechanism because it does not use more than 1 equiv of LCuNO₂ to nitrate the phenol. However, we cannot rule out nucleophilic attack of the phenol on LCuNO₂ to form the phenoxyl radical of DTBP, which then could react with another 1 equiv of LCuNO₂ via the “rebound” mechanism.

Alternatively, 2 equiv of the [LCuHNO₂] adduct could functionalize DTBP directly, resulting in the nitrated phenol product of DTBP (Scheme 3, path B). This mechanism is consistent with the stoichiometry experiments, and there is literature precedence for the involvement of a nitrous acid adduct, [CoHNO₂]²⁺,⁷² nitrating DTBP. However, we view this mechanism unlikely for [LCuHNO₂] because one of the products would be an unprecedented {CuNO}¹⁰ complex.

Yet another pathway to generate 2,4-di-tert-butyl-6-nitrophenol from LCuNO₂ and DTBP would involve one-electron oxidation of DTBP by LCuNO₂, with the resulting cation-radical phenol species reacting with free nitrite, as proposed for the protonated cryptand-capped tripodal Cu^II-NO₂ compound.^19,73 We rule this pathway out because the reported E_ox value of DTBP is 1.03 V versus Fc/Fc⁺ (converted from 1.43 V vs SCE),⁷⁴ which is about 800 mV higher than the E_1/2 value found for the [LCuNO₂]^−/0 couple. As discussed for the protonated cryptand-capped tripodal Cu^II-NO₂ compound, phenol nitration typically occurs via O₂-dependent pathways through metal peroxynitrite species. The reactions of the compounds LCuNO₂ and the protonated cryptand-capped tripodal Cu^II-NO₂ species are two recent examples of phenol nitration that occur through uncommon anaerobic pathways.¹⁹

■ CONCLUSIONS

The copper(II) nitrite starting materials [M][LCuNO₂] (M = NBu₄⁺ or PPN⁺) were prepared and characterized by UV–vis and EPR spectroscopy, CHN analysis, and X-ray crystallography. Electrochemical and chemical oxidations of [LCuNO₂]⁻ revealed a reversible one-electron oxidation to an intriguing [CuNO₂]²⁺ core, and LCuNO₂ was characterized at low temperatures by UV–vis and resonance Raman spectroscopies. To our knowledge, this complex is a unique example of a high-valent copper nitrite complex. The reaction between LCuNO₂ and TTBP revealed that LCuNO₂ can abstract H atoms from O–H substrates at rates comparable to those in LCuO₂CR complexes.⁵¹ Interestingly, the reaction between LCuNO₂ and DTBP yielded not only the expected PCET product but the ortho-nitrated phenol as well. Although there are multiple possible mechanisms that could result in the nitrated phenol, detection of the nitrated product indicates that LCuNO₂ (or a derivative) can functionalize substrates, a type of transformation that has only been demonstrated by X = halides for LCuX complexes.⁵²