Carboxylate Structural Effects on the Properties and Proton Coupled Electron Transfer Reactivity of [CuO₂CR]²⁺ Cores

Abstract

A series of complexes {[NBu₄][LCu^II(O₂CR)] (R = −C₆F₅, −C₆H₄(NO₂), −C₆H₅, −C₆H (OMe), −CH₃, and −C₆H₂(ⁱPr)₃)} were characterized (with the complex R = −C₆H₄(m-Cl) having been published elsewhere (Mandal et al. J. Am. Chem. Soc. 2019, 141, 17236)). All feature N,N′,N″-coordination of the supporting L²⁻ ligand, except for the complex with R = −C₆H₂(ⁱPr)₃, which exhibits N,N′,O- coordination. For the N,N′,N″-bound complexes, redox properties, UV–vis ligand-to-metal charge transfer (LMCT) features, and rates of hydrogen atom abstraction from 2,4,6,-tri-t-butylphenol using the oxidized, formally Cu(III) compounds LCu^III(O₂CR) correlated well with the electron donating nature of R as measured both experimentally and computationally. Specifically, the greater the electron donation, the lower is the energy for LMCT and the slower is the reaction rate. The results are interpreted to support an oxidatively asynchronous proton-coupled electron transfer mechanism that is sensitive to the oxidative power of the [Cu^III(O₂CR)]²⁺ core.

Graphical Abstract:

INTRODUCTION

With the ultimate aim of using a mechanistic understanding to inform the development of more efficient and selective oxidation catalysts, copper complexes relevant to postulated catalytic intermediates have been targeted for detailed study.¹ Species with [CuO]⁺, [CuOH]²⁺, [CuO₂]⁺, and [CuOOR]^+/2+ cores have garnered particular attention because of their hypothesized role as oxidizing intermediates in synthetic systems² and in enzymes.³ Studies of complexes with such cores aim to understand spectroscopic and redox properties and to unravel structure/reactivity relationships. For example, in previous work, we prepared formally Cu(III) complexes LCuOH (L = bis(2,6-diisopropylphenylcarboxamido)pyridine and derivatives) (Figure 1), defined their structural and spectroscopic attributes, and examined their proton-coupled electron transfer (PCET) reactivity with C–H and O–H bonds in organic substrates.^4,5 Rapid rates for these reactions were rationalized by the formation of a strong O–H bond in the product, LCu(OH₂), and the effects of changing the electron-donating properties of the supporting ligand L²⁻ were evaluated through variation of substituents on the aryl rings or reducing the pyridine moiety. Also, the PCET reactions of LCuOH and [LCuO₂]⁻ with a series of para-substituted phenols were compared, revealing intriguing differences in mechanisms traversed as the properties of the phenol substrates were varied.⁵ These differences were traced to divergent thermodynamic properties of the two cores supported by the same ligand L²⁻.

Figure 1.

Copper complexes supported by L²⁻ (focus of this work in box).

More recently,⁶ we prepared a new formally Cu(III) complex LCuO₂CR (R = −C₆H₄(m-Cl); Figure 1) and compared its rates of reactions with 9,10-dihydroanthracene (DHA) and 2,4,6,-tri-t-butylphenol (TTBP) to those of LCuOH and LCuOOC(Me₂)Ph.^7,8 Differences in the rates of PCET reactions were observed (LCuOH > LCuO₂CR ≫ LCuOOC(Me₂)Ph), and these differences were rationalized by invoking changes in the mechanisms followed. Importantly, in the current context, the reactions of LCuO₂CR (R = −C₆H₄(m-Cl)) were found via theory to involve a concerted, yet asynchronous PCET process, with the proton migrating to the carbonyl oxygen atom that is likely not (or only weakly) bound to the copper ion, well after the electron is transferred to the copper center. Such a process represents an example of site-separated PCET, which has attracted significant attention, in part because it is likely to be important in biological contexts.⁹

Intrigued by the novel mechanistic findings for LCuO₂CR (R = −C₆H₄(m-Cl)), we sought to probe the effects of changing R. Specifically, we asked how varying R might influence the structural, spectroscopic, redox, and reactivity properties of the [CuO₂CR]²⁺ core and its precursor [CuO₂CR]⁺. Herein, we report the results of such studies, which provide fundamental chemical insights, including into how thermodynamic attributes influence reactivity of formally Cu(III) species.

RESULTS AND DISCUSSION

Synthesis and Characterization of Complexes.

We targeted complexes [NBu₄][LCu^II (O₂CR)] by treating a THF solution of [NBu₄][LCu^IIOH]⁵ with 1 equiv of RCO₂H in the presence of molecular sieves, a method reported previously for the case of R = −C₆H₄(m-Cl). The complexes [NBu₄][LCu^II(O₂CR)] were isolated as teal/green solids in good yields (53–66%) for the R groups indicated in Scheme 1. For the case of R = CH₃, the complex was prepared by a reaction of [NBu₄][O₂ CCH₃] with LCu^II(MeCN) in THF (79%).^4a The indicated formulations shown in Scheme 1 were supported by electrospray ionization mass spectroscopy (ESI-MS), elemental analysis, and UV–vis and EPR spectroscopy and by analogy to X-ray crystal structures reported here for R = −CH₃ and –C₆H₂(ⁱPr)₃ and previously for R = −C₆H₄(m-Cl)⁶.

Scheme 1.

Syntheses of Complexes

The X-ray structures of the anionic portions of the complexes [NBu₄][LCuO₂ CR] (R = −CH₃ and − C₆H₂(ⁱPr)₃ are shown in Figure 2. The topology of the complex with R = −CH₃ (Figure 2a) is analogous to that reported for R = −C₆H₄(m-Cl)⁶ and features N,N′,N″ coordination of L²⁻ with a monodentate carboxylate completing the approximately square planar geometry (geometry index, τ₄ = 0.2).¹⁰ The second carboxylate oxygen (O2) atom interacts weakly with the copper ion in the axial position (Cu–O2 = 2.545(2) Å). On the basis of spectroscopic similarities (vide infra), we propose analogous structures for all the other carboxylate complexes, with the exception of the complex with R = −C₆H₂(ⁱPr)₃. For this latter case (Figure 2b), while similar carboxylate coordination is observed, the supporting ligand binds differently, with one carboxamido unit flipped to bind via its O atom (N,N′,O-coordination). Such isomerization of the carboxamide arms of L²⁻ have been reported in complexes of Ru, Ni, and Fe.¹¹ In contrast to previously reported O-coordination of a protonated carboxamide arm in Cu(II) complexes,¹² the indicated RN═C(R)–O⁻ formulation (Scheme 1) is supported by the overall charge (−1) and the short C–N (N3–C17 = 1.290(2) Å vs N1–C23 = 1.347(2) Å) and long C–O bond distances (O4–C17 = 1.305(2) Å vs O3–C23 = 1.248(2) Å). We speculate that the steric bulk of the −C₆H₂(ⁱPr)₃ moiety is responsible for the observed N,N′,O-coordination. Finally, we note a trend in Cu–O(carboxylate) distance that correlates with the steric profiles of the R groups; as R size decreases (−C₆H₂(ⁱPr)₃ > −C₆H₄(m-Cl) > −CH₃), the Cu1–O1 bond shortens(1.945(2) Å > 1.929(3) Å > 1.906(1) Å). Put differently, the Cu⋯O2 distance in the most hindered complex (R = −C₆H₂(ⁱPr)₃, 2.919(1) Å) is significantly longer than in the other two structures (R = −C₆H₄(m-Cl), 2.516(3) Å; R = −CH₃, 2.545(2) Å).

Figure 2.

Representations of the anionic portions of the X-ray crystal structures of (a) [NBu₄][LCuO₂CCH₃] and (b) [NBu₄][LCuO₂CC₆H₂(ⁱPr)₃] with all nonhydrogen atoms shown as 50% thermal ellipsoids and hydrogen atoms omitted for clarity. Selected bond distance (Å) and angles (deg): [NBu₄][LCuO₂CCH₃]: Cu–O1, 1.9452(17); Cu–O2, 2.5450(19); Cu–N1, 2.0158(18); Cu–N2,1.9400(18); Cu–N3, 2.0163(19); N1–Cu–N3, 159.59(7); N1–Cu–N2, 80.44(8); N2–Cu–O1, 172.50(8). [NBu₄][LCuO₂CC₆H₂(ⁱPr)₃]: Cu–O1, 1.9055(12), Cu–O2, 2.9190(14); Cu–N1, 1.9906(13); Cu–N2, 1.9148(14); Cu–O4, 2.0020(12); N3–C17, 1.290(2); O4–C17, 1.3049(18); N1–C23, 1.347(2); O3–C23, 1.2470(19), N1–Cu–O4, 162.12(5); N1–Cu–N2, 81.06(6); N2–Cu–O1, 170.92(6).

The UV–vis and EPR spectra of the series [NBu₄][LCu^II(O₂CR)] corroborate their structural features and distinguish between N,N′,N″- and N,N′,O-coordination. While all complexes exhibit typical d–d transitions in the UV–vis region (Figure S1, Table 1), the feature for the N,N′,O-complex with R = −C₆H₂(ⁱPr)₃ is distinctly shifted toward longer wavelength/lower energy (λ_max = 673 nm; ν_max =14859 cm⁻¹) compared to the other N,N′,N″-complexes (λ_max = 598–619 nm; ν_max = 16155–16722 cm⁻¹), consistent with the divergent coordination mode of L²⁻. Further supporting this assignment, the X-band EPR spectrum for this N,N′,O- complex is distinct from the quite similar spectra for all the other complexes, which are also closely analogous to previously reported Cu(II) complexes^4a,b,d featuring N,N′,N″-coordination of L²⁻ (Table S2 and Figure S2; illustrative spectra for R = −C₆H₅ and −C₆H₂(ⁱPr)₃ shown in Figure 3). Thus, while all spectra feature pseudoaxial signals typical for square planar Cu(II) compounds, the g_z (2.228) and Cu hyperfine (A_z^Cu = 545 MHz) values for the complex with R = −C₆H₂(ⁱPr)₃ differ significantly from other complexes (avg. g_z = 2.201, avg. A_z^Cu = 576 MHz), as illustrated for the specific example R = −C₆H₅ in Figure 3. Most notably, the N superhyperfine pattern for the complex with R = −C₆H₂(ⁱPr)₃ exhibits fewer lines (~12; Figure 3b, top) than for the other complexes (~17 for R = −C₆H₅, Figure 3b, bottom), consistent with fewer N atoms bound to the Cu(II) ion in the former complex and the postulated N,N′,O- rather than N,N′,N″-coordination. We thus conclude on the basis of UV–vis and EPR spectroscopy that the solid-state structures illustrated in Figure 3 are retained in solution.

Table 1.

Selected UV–Vis Data^a

complex	R	λmax (nm)	νmax (cm−1)	ε (M−1 cm−1)
[LCuII(O2CR)]−b	C6F5	598	16722	630
	C6H4(NO2)	600	16667	300
	C6H4(a)c	609	16420	400
	C6H5	612	16340	420
	C6H4(OMe)	617	16207	400
	CH3	619	16155	350
	C6H2(iPr)3	673	14859	300
[LCuIII(O2CR)]d	C6FS	506	19763	5000
		670	14925	10800
		866	11547	9600
	C6H4(NO2)	495	20202	5500
		653	15314	13200
		835	11976	10600
	C6H4(C1)c	491	20367	5000
		650	15385	13000
		830	12048	10500
	C6HS	493	20284	4700
		646	15480	11300
		827	12092	9200
	C6H4(OMe)	485	20619	4700
		636	15723	11200
		824	12136	8800
	CH3	467	21413	5800
		624	16026	13700
		809	12361	9700

^aFull data provided in Figures S1 and S5.

^bOnly dd features listed. Conditions: 25 °C, THF.

^cRef 6.

^dConditions: −80 °C, THF.

Figure 3.

X-band EPR data for [NBu₄][LCu^II(O₂CR)] (R = −C₆H₂(ⁱPr)₃, red, top; R = −C₆H₅, black, bottom). (a) Stacked spectra with dashed lines drawn to highlight differences in alignment of g_z and g_x features. (b) Second-derivative data for the g_y region with dashed lines indicating positive peaks to highlight the difference in the number of observable A^N superhyperfine lines. See Table S2 for EPR parameters derived from spectral simulations for all complexes.

Cyclic voltammograms were measured for the carboxylate complexes dissolved in THF with 0.3 M Bu₄NPF₆ at room temperature using a glassy carbon electrode and the ferrocene/ferrocenium (Fc/Fc⁺) or the decamethyl ferrocene/decamethyl ferrocenium (Fc*/Fc*⁺) couples as internal standards (Figure S6). With the exception of the N,N′,O-bound complex with R = −C₆H₂(ⁱPr)₃ (which exhibits irreversible features; see the Supporting Information), all of the [NBu₄][LCu^II(O₂CR)] systems exhibited pseudoreversible Cu^III/II redox couples (i_pa/i_pc ratios ranging from 1.00 to 1.34), which deviated from ideality to varying extents as reflected by peak separations (ΔE_p) at 100 mV/s, ranging from 0.112 to 0.173 V (Table 2). Estimated E_1/2 values within the series spanned a range of 150 mV (from 0.150 to 0.298 V vs Fc/Fc⁺) and were correlated with the electron donating nature of the bound carboxylate (less electron donating, higher E_1/2). This correlation is illustrated in a plot of E_1/2 versus the pK_a (in H₂O) of the free acid of the carboxylate ligand (Figure 4), which shows a linear dependence and a slope indicating how E_1/2 increases with lower pK_a (lower basicity of the carboxylate ligand). The E_1/2 values for [NBu₄][LCu^II(O₂CR)] are also compared to those of [LCuX]⁻ (X = OH, OOC(Me₂)Ph) in Table 2; the lower redox potentials for the latter compounds are consistent with greater basicity of their X moieties compared to the carboxylates.

Table 2.

Cyclic Voltammetry Data for [LCu(O₂CR)]⁻ and [LCuX]⁻ (X = OH, O C(Me₂)Ph)^a

complex	R	E1/2b	ipa/ipcc	ΔEpc
[LCu(O2CR)]−	C6F5	0.298(2)	1.0	0.15
	C6H4(NO2)	0.239(5)	1.2	0.17
	C6H4(Cl)d	0.228(8)	1.0	0.16
	C6H5	0.169(5)	1.1	0.11
	C6H4(OMe)	0.151(6)	1.3	0.14
	CH3	0.150(3)	1.3	0.12
[LCuOH]−e		−0.074
[LCuO2CMe2Ph]−f		−0.154

^aUnless noted otherwise, all data were acquired in THF at room temperature using 0.3 M Bu₄NPF₆, glassy carbon electrode.

^bV vs Fc/Fc⁺.

^cListed for carboxylate complexes only; for ΔE_p, units are V at scan rate = 100 mV s⁻¹.

^dRef 6.

^eRef 4b.

^fRef 7.

Figure 4.

Plot of E_1/2 for the [LCu(O₂CR)]^0/−couple (THF) vs the pK_a of RCO₂H (H₂O), with the data points labeled with the respective R groups. The red line is a linear fit to the data: E_1/2 = −0.052pK_a + 0.40 (R² = 0.91).

Characterization of [Cu^IIIO₂CR]²⁺ Complexes.

Treatment of solutions of [NBu₄][LCu^II (O₂CR)] (all R except −C₆H₂(ⁱPr)₃ because of its irreversible electrochemistry) in THF with [AcFc][BAr^F₄] at −80 °C resulted in an immediate color change from pale teal to deep blue. UV–vis spectra of the product solutions were similar and exhibited intense (ε ~ 5 × 10³ − 1 × 10⁴ M⁻¹ cm⁻¹) features between ~500 and 1000 nm (illustrative spectra for R = −C₆F₅ and −C₆H₄(OMe) in Figure 5; all spectra overlaid in Figure S5 with λ_max/ν_max data listed in Table 1). Consistent with formulation of the products as the 1-electron oxidized [Cu^IIIO₂CR]²⁺ complexes, (a) for the case of R = C₆F₅, titration experiments showed that the maximum absorbance values were attained upon addition of 1 equiv of [AcFc][BAr^F₄] (Figure S7), and (b) for the cases of R = C₆H₄(Y) (Y = NO₂ and OMe), the addition of 1 equiv of Fc* bleached the solutions, and the UV–vis features reformed upon subsequent addition of 1 equiv of [AcFc][BAr^F₄], with the reversibility indicated by closely similar results after 3 repeated cycles (Figure S8). Lastly, we note that similar spectra were observed (t_1/2 ~20 min) when the oxidation reactions were performed at −25 °C in 1,2-difluorobenzene (DFB, Figure S5), conditions which were used in previous studies^4b,d and were needed to observe some subsequent reactions with substrates (vide infra).

Figure 5.

UV–vis spectra of [LCu^III(O₂CR)] (R = −C₆F₅ and −C₆H₄(OMe)) in THF at −80 °C.

On the basis of computational results reported previously for the system with R = −C₆H₄(m-Cl)⁶, we attribute the UV–vis spectral features of all the N,N′,N″-complexes reported herein to varying compositions of two different ligand-to-metal charge transfer (LMCT) transitions involving the supporting ligand (L²⁻) and the copper center, N-aryl π → Cu d and N-amide π → Cu d. These assignments were further confirmed by performing time-dependent density functional theory (TD-DFT) calculations on individual systems (see the Supporting Information). The absorption at >800 nm (<12300 cm⁻¹) for the [CuO₂CR]²⁺ cores occur at uniquely high wavelengths/low energies relative to those observed for other formally Cu(III) complexes supported by L²⁻.⁴ Differences along the series were observed that correlate with the electron-donating nature of the bound carboxylate ligand (longest wavelength, highest energy for most electron-withdrawing −C₆F₅: 506, 670, 866 nm). The trend may be viewed graphically by plotting the energy at the peak maximum near 800 nm (ν_max) vs the pK_a of the free acid of the corresponding carboxylate ligand (Figure 6a) or σ _p¹³ for the case of R = −C₆H₄(Y) (Y = OMe, H, m-Cl, NO₂) (Figure 6b). The indicated linear correlations illustrate how the observed ν_max decreases as the bound carboxylate becomes increasingly electron-withdrawing (as pK_a decreases and σ_p increases) and the metal center becomes more electrophilic. These trends are consistent with the LMCT assignment of the UV–vis features, as greater electrophilicity of Cu corresponds to lower d orbital (LUMO) energies, resulting in lower energies (ν_max) for the LMCT transitions.

Figure 6.

(a) Plot of ν_max of the lowest energy feature in the UV–vis spectrum of [LCu^III(O₂CR)] (THF, −80 °C) vs the pK_a of RCO₂H (H₂O). The red line is the linear fit to ν_max = 242pK_a + 11000, R² =0.96. (b) Plot of ν_max for the lowest energy feature in the UV–vis spectrum of LCu^III(O₂CC₆H₄(Y)) (THF, −80 °C) vs σ_p for the indicated Y. The red line is the linear fit to ν_max = −149σ_p + 12100, R² = 0.99. Note: both σ_p and σ_m values for the case of R = m-Cl are plotted; see ref 12.

Reactivity of [Cu^IIIO₂CR]²⁺ Complexes.

We examined the reactions of LCu^III(O₂CR) (all R except −C₆H₂(ⁱPr)₃ with TTBP and DHA in order to evaluate trends in reactivity with relatively weak O–H and C–H bonds. In the case of both substrates, reactions were indicated by a decay of the UV–vis spectroscopic features associated with LCu^III(O₂CR), the observation of the aryloxyl radical by EPR spectroscopy in the reaction of LCu^III(O₂CC₆H₄(Y))⁶ with TTBP, and the observation of the formation of anthracene by UV–vis spectroscopy in the reactions with DHA. Second-order rate constants (k₂) for the reactions of LCu^III(O₂CR) with TTBP (THF, −80 °C) and DHA (DFB, −25 °C) are listed in Table 3 (see Figures S9–S13 for more detailed kinetic information). A clear trend in k₂ with the electron donating properties of R is seen for the reactions with TTBP, which is illustrated in a plot of log k₂ vs E_1/2 for the [LCu^III,II(O₂CR)]^0/− couple in Figure 7. A similar linear relationship is seen in a plot of log k₂ vs pK_a of RCO₂H (Figure S15), which follows from the dependence of E_1/2 on the pK_a of the free acid RCO₂H (Figure 4). Together, these plots show that, the more electron withdrawing R, the more oxidizing is the complex and the faster is the rate of reaction with TTBP.

Table 3.

Second-Order Rate Constants (k₂) for Reactions of [LCu^III(O₂CR)] with the Indicated Substrates^a

R	TTBPb	DHAc
C6F5	1.8(4)	0.014(3)
C6H4(m-Cl)d	0.3(1)	0.11(4)
C6H4(NO2)	0.23(3)	0.0079(4)
C6H5	0.052(8)	0.05(2)
C6H4(OMe)	0.09(1)	0.015(7)
CH,	0.025(3)	0.015(9)

^aUnits: M⁻¹ s⁻¹.

^bTHF, −80 °C.

^cDFB, −25 °C.

^dRef 6.

Figure 7.

Plot of log k₂ for the reaction of LCu^IIIO₂CR with TTBP at −80 °C in THF vs E_1/2 for the [LCu(O₂CR)]^0/− couple. Red line is the linear fit to log k₂ = 10.7E_1/2− 2.98, R² = 0.90.

To gain insight into the experimentally characterized variation in reactivity as a function of the substituents on the benzoates in LCu^IIIO₂CR described above, density functional modeling of the PCET reaction coordinate was undertaken at the B3LYP-D3(BJ)/basis-II/SMD(THF)//B3LYP-D3(BJ)/basis-I level of theory^14–16 (basis-I: 6–31G(d)¹⁷ for light atoms and SDD¹⁸ for Cu; basis-II: def2-TZVP¹⁹ basis for nonmetals and SDD for Cu) along the broken-symmetry singlet potential energy surface (see the Supporting Information for details and coordinates file). Table 4 displays computed free energies of activation (ΔG^‡) for reactions of the indicated complexes with the substrate 2,6-di-t-butylphenol (DTBP, lacking the third t-butyl group of TTBP for computational expediency). In fair agreement with the experimentally observed trend, increased electron-withdrawing ability of R correlates with an increase in the PCET rate, as quantified by the general decrease (within the error of DFT) in the computed ΔG^‡ values (Table 4).

Table 4.

Computed ΔG^‡ and Other Parameters for the Reactions of [LCu^IIIO₂CR] with DTBP^a

R	ΔG‡b	Cu–O1 (Å)	TS Cu–O1 (Å)	〈S2〉c
C6FS	2.3	1.868	1.984	0.853
C6H4(NO2)	3.7	1.864	1.966	0.827
C6H4(m-Cl)d	2.9	1.859	1.9S6	0.817
C6HS	7.9	1.861	1.948	0.730
C6H4(OMe)	7.2	1.860	1.943	0.700

^aSMD(THF)/B3LYP-D3(BJ)/basis-II//B3LYP-D3(BJ)/basis-I level of theory at 193.15 K.

^bUnits: kcal mol⁻¹.

^cCalculated at the UB3LYP-D3(BJ)/basis-I level of theory.

^dRef 6.

In a related work,⁶ we have demonstrated that the reaction of LCu^III(O₂CC₆H₄(m-Cl)) with DTBP proceeds via a concerted proton-coupled electron transfer pathway having a high degree of oxidative asynchronicity, meaning that the transition-state (TS) structure for the concerted event has more electron transfer-like character. Similar asynchronicity in electron/proton transfer may also be invoked for all the N,N′,N″-carboxylates described herein, with the degree of asynchronicity increasing as the electron-withdrawing ability of R gradually increases (recall that the asynchronicity is oxidative in nature, and the Cu center systematically becomes more oxidizing). Along with other influencing factors, such increasing asynchronicity in electron/proton transfer has been documented to lower the free energy barrier of a concerted PCET process,²⁰ so increased PCET reaction rates with more electron-withdrawing R is anticipated.

Furthermore, we note that, with increasing electron withdrawing nature of R, the cleavage of the Cu–O1 bond in individual TS structures (Figure 8) becomes easier, facilitating the formation of the product Cu^II species. Thus, while the equatorial Cu–O1 bond length in the Cu^III reagent remains similar across the series (variation of ~0.009 Å, Table 4), it varies significantly in the respective TS structures (variation of ~0.041 Å, Table 4, Figure 8). This interpretation is supported by the comparison of the computed values of the 〈S²〉 operator, which gradually increase toward unity with increasing electron-withdrawing ability of R, consistent with increased separation of the developing radical loci in the products. The significant computational challenges associated with the broken-symmetry approach in the region of the transition state suggests that the quantitative variation as a function of substituent may not be particularly accurately predicted, but the consistency of the qualitative variation with the experiment suggests that the observed variations in geometric and electronic structures as a function of substituent are likely germane to the reactivity differences.

Figure 8.

Representative transition-state structure for the reaction of LCu^IIIO₂CC₆H₅ with DTBP. Hydrogen atoms and the 2,6-diisopropylphenyl groups on the L²⁻ ligand are omitted for clarity. Selected distances (Å): Cu–O1, 1.948; Cu–O2, 2.699; O2–H,1.178; H–O3, 1.211.

It is informative to compare the trends in LCu^III(O₂CR)/TTBP reactivity to similar studies of the reactions of copper–oxygen species (supported by L²⁻ and related ligands) with phenols or DHA.^4d,5 Reactions of LCu^IIIOH with a series of para-substituted phenols (^XArOH, X = NO₂, CF₃, Cl, H, Me, OMe, NMe₂) showed a general trend of increasing rate with more electron rich substrates (greater electron donating substituents, lower E_1/2, higher pK_a), albeit with outliers for the most electron poor phenols (e.g., X = NO₂).^5a These and other results were interpreted to indicate that these reactions involved concerted proton-coupled electron transfer but with asynchronicity favoring ET (X = NMe₂) or favoring PT (X = OMe, Me, H, Cl).⁵ The rates of these reactions are significantly faster than those of LCu^III(O₂CR) with TTBP, but the trends are analogous, with faster rates exhibited by the most electron-poor, oxidizing reagents (^XArOH or LCu^III(O₂CR), respectively). In another work,^4d the reactivity of [CuOH]²⁺ cores supported by derivatives of L²⁻ with perturbed electron donating capabilities was probed by incorporating −NO₂ groups at the para positions of the 2,6-diisopropylaryl rings (^NO2L²⁻, electron poor) or hydrogenating the central pyridine ring to generate a piperidine donor (^pipL²⁻, electron rich). The observed trend in rates of PCET with DHA (^NO2LCuOH > LCuOH > ^pipLCuOH) were rationalized by the same trend in product bond dissociation enthalpies (BDEs), which in turn resulted from greater influences of the ligand changes on redox potential than on hydroxide basicity. For the reactions of LCuO₂CR, the trend in rate of reaction with TTBP is also dominated by the effect of R on the redox potential, as the rate increases as E_1/2 increases and the basicity of the (free) carboxylate ion decreases. However, in these reactions, the trend in rates does not correlate with the free carboxylic acid O–H BDEs (Table S3), which notably are essentially identical for R = −NO₂ and − OMe (107.1 and 106.8 kcal mol⁻¹, respectively).²¹ We speculate that the absence of such a thermodynamic rationale reflects asynchronicity of the PCET reaction and/or perturbation of the BDEs through interactions of the carboxylate ion or product carboxylic acid with the copper ion (effects that we have not measured).

In contrast to the clear reactivity trend for the reactions of LCu^IIIO₂CR with TTBP (Figure 7), no such obvious trend is apparent for the (slower) reactions with DHA (Figure 9). Particularly striking are the essentially identical values of k₂ for the reactions in the cases of R = −C₆F₅ and −C₆H₄(OMe) for which the [LCu^III,II(O₂CR)]^0,− E_1/2 differs by ~0.15 V and the order of magnitude difference in rate constants for the cases of R = −C₆H₄(m-Cl) and − C₆H₄(NO₂), which have very similar E_1/2 values (~10 mV difference). DFT calculations also revealed no clear reactivity trend for reactions of [LCu^IIIO₂CR] with DHA (TTable S7). We speculate that the absence of a clear reactivity trend in the reactions with DHA may result from a change in mechanism across the series and/or steric effects that confound interpretation of the electronic influences of R.²²

Figure 9.

Plot of log k₂ for the reaction of [LCu^IIIO₂CR] with DHA at −25 °C in DFB vs E_1/2 for the [LCu(O₂CR)]^0/−couple.

CONCLUSIONS

A series of complexes {[NBu₄][LCu^II(O₂CR)] (R = −C₆F₅, −C₆H₄(NO₂), −C₆H₅, −C₆H₄(OMe), −CH₃, and − C₆H₂(ⁱPr)₃)} were prepared, characterized, and compared to the data for the complex with R = −C₆H₄(m-Cl) published elsewhere.⁶ All adopted structures featuring N,N′,N″-coordination of the supporting L²⁻ ligand, except for the case where R = −C₆H₂(ⁱPr)₃, in which the supporting ligand isomerized to N,N′,O-coordination as revealed by X-ray crystallography and UV–vis and EPR spectroscopy. For all complexes except the one with R = −C₆H₂(ⁱPr)₃, cyclic voltammograms exhibited pseudoreversible waves with E_1/2 values that correlated with the electron donating properties of the carboxylate, as indicated by a plot of E_1/2 vs the pK_a of the carboxylic acid (Figure 4). Chemical oxidations of these complexes at −80 °C in THF or −25 °C in DFB resulted in the formation of LCu^III(O₂CR), which exhibited appropriate LMCT features in their UV–vis spectra. These features as well as the rates of the reactions of LCu^III(O₂CR) with the O–H bond in TTBP correlated well with the electron donating/withdrawing nature of R, as revealed by linear fits to plots of ν_max with carboxylic acid pK_a and (for the benzoates) σ_p (Figure 6), and a logarithmic plot of the second order rate constant for the reaction with TTBP vs the [LCu^III,II(O₂CR)]^0,− E_1/2 (Figure 7). DFT calculations corroborated the experimentally observed reactivity trend during hydrogen atom abstraction from 2,6-di-t-butylphenol by N,N′,N″-complexes. A systematic increase in the reaction rate is observed with the increasing electron withdrawing nature of R, quantified by, within error, a gradual decrease in the computed ΔG^‡ values, attributable to a gradual increase in the degree of asynchronicity in electron/proton transfer. No clear trend was observed in the slower reactions with DHA, pointing to possible mechanism changes across the series and/or complicating steric effects. Taken together, the evidence supports oxidatively asynchronous proton-coupled electron transfer pathways for the reactions of LCu^III(O₂CR) with the hindered phenol, with the oxidizing power of the complexes being predominant in controlling the reaction rate.