Reducing properties of 1,2-dipyridyl-1,2-disodioethanes: chemical validation of theoretical and electrochemical predictions

Abstract

The reducing properties of highly delocalized radical anions and dianions of 1,2-di(hetero)arylethenes were investigated by theoretical calculations at the PBE0/6-311+G(d,p)/IEFPCM level. The results correlated nicely with the reduction potentials determined by analysis of the voltammetric curves for the reduction of the parent alkenes, and this allowed a reliable scale for their relative reducing strength to be established. In full agreement with calculations and electrochemical results, use of the appropriate 1,2-dipyridyl-1,2-disodioethane as a base led to the successful α-alkylation of bromophenylacetic acids under mild reaction conditions, thus avoiding the competitive reductive cleavage of aromatic C–Br bonds.

---

Introduction

The tunability of redox reactions is a topic of primary importance in chemistry, finding application in a variety of areas such as extension of the synthetic usefulness of reagents such as SmX₂,^1,2 Ni(II)-salen^3,4 or neutral organic electron donors,⁵ control of the particle size during the formation of metal nanoparticles,⁶ or electrochemical performance of fuel cells.⁷ 1,2-Diaryl-1,2-disodioethanes⁸ constitute an easily accessible class of organometallic compounds endowed with interesting reducing properties. Besides applications as dinucleophiles in the diastereoselective synthesis of trans-1,2-diarylcyclopentanes⁹ and borinanes,¹⁰ we previously described their use as electron-transfer reagents for highly effective reductive eliminations in 1,2-hetero-disubstituted compounds^11,12 and degradations of aromatic persistent organic pollutants.^13–17

A comparison between the electroreduction potentials of the 1,2-diarylethenes and the results obtained for the reduction of model compounds with the corresponding 1,2-diaryl-1,2-disodioethanes bearing different substituents allowed us to observe that the less delocalized dianions are the most powerful reducing agents, and vice versa.¹⁸ These results were substantiated by establishing that the outcome of the reactions between halogenated arylacetic acids and 1,2-diaryl-1,2-disodioethanes strongly depends on the nature of both partners. The reaction pattern was rationalized in terms of a competition between the ease of the reductive cleavage of the carbon halide bond and the reducing and basic properties of the vic-diorganometals. These observations were instrumental to set up mild reaction conditions for the chemoselective generation of the enediolates of 2-fluoro- and 2-chloro-phenylacetic acids, as well as trapping of these reactive species with different electrophiles.¹⁹ However, our earlier attempts to extend such a procedure to the generation of enediolates of bromophenylacetic acids invariably led to the formation of phenylacetic acid as a main product. Reductive dehalogenation occurred even in the presence of 1,2-disodio-1-(2-pyridyl)phenylethane, i.e., the mildest reducing reagent employed (Scheme 1).^18,19 On these grounds and looking for more selective reagents, we speculated that addition of further pyridyl rings could generate dianions endowed with an appropriate balance between reducing and basic properties thereby allowing, inter alia, their use in effective α-functionalizations of bromophenylacetic acids.

Scheme 1 Competing basic and reducing properties of 1,2-diaryl-1,2-disodioethanes (from ref. 19). E⁺ = electrophile.

To check this hypothesis, we carried out a study aimed to verify whether the reducing properties of a series of dianions of 1,2-diarylethanes (3a–d, Chart 1) could be predicted by density functional theory (DFT) calculations, verified electrochemically, and then tested experimentally. The DFT computed electron affinities nicely correlate with the electrochemical reduction potentials of the parent alkenes (1a–d, Chart 1). The outcome of the DFT and electrochemical analyses was checked by preparing the sodium dianions of 1-(2-pyridyl)-2-(4-pyridyl)ethane (1a) and 1,2-di(2-pyridyl)ethane (1b) and then using them in reactions with halogenated benzoic and arylacetic acids.

Chart 1 Formulas of the investigated 1,2-diarylethenes (1a–d) and corresponding radical anions (2a–d) and dianions (3a–d).

Experimental section

Materials

Diarylalkenes 1a–d are commercially available and were purified by distillation or recrystallization immediately prior to use. Benzoic acids, 5a–d, and arylacetic acids, 5e–g, are commercially available. 18-crown-5 was distilled in vacuo immediately prior to use. Tetrahydrofuran (THF) was distilled from Na/K alloy under N₂ immediately prior to use. N,N-Dimethylformamide (DMF) was treated with anhydrous Na₂CO₃ and distilled under a nitrogen atmosphere. Tetra-n-butylammonium perchlorate (TBAP) was recrystallized from a 2 : 1 ethanol–water solution and dried at 60 °C under vacuum.

Instruments and methods

¹H NMR spectra were recorded at 300 or 400 MHz and ¹³C NMR spectra were recorded at 75 or 100 MHz in CDCl₃ with SiMe₄ as internal standard. Deuterium incorporation was calculated by monitoring the ¹H NMR spectra of crude reaction mixtures, and comparing the integration of the signal corresponding to protons in the arylmethyl (or heteroarylmethyl) position with that of known signals. Resonances of the CHD protons are shifted 0.02–0.04 ppm (δ) upfield relative to the resonances of the corresponding CH₂ protons; the resonances of the arylmethyl CHD carbons appear as triplets (J = 18–20 Hz) shifted 0.3–0.5 ppm (δ) upfield relatively to the corresponding arylmethyl (or heteroarylmethyl) CH₂ carbons. IR spectra were recorded on a FT-IR Jasco 680 P. Flash chromatography was performed on Merck silica gel 60 (40–63 μm), and TLC analyses on Macherey-Nagel silica gel pre-coated plastic sheets (0.20 mm). Elemental analyses were performed by the microanalytical laboratory of the Dipartimento di Chimica e Farmacia, Università di Sassari.

Computational details

Structure optimizations of compounds 1, of the corresponding radical anions, 2 and dianions, 3, were performed at the DFT level employing the PBE0 functional,²⁰ a parameter-free hybrid variant of the Perdew, Burke and Ernzerhof (PBE) generalized gradient functional,²¹ as implemented in the commercially available suite of programs GAUSSIAN 09.²² The 6-311+G(d,p) basis set was employed for all atoms. Unconstrained geometry vibrational analysis was carried out at the same level of theory to check the character of the stationary points and to calculate the thermochemistry data. In absence of coordinating cations a planar geometry was observed for all studied compounds. Calculations were performed in vacuo and with a polarizable continuum model of the solvent (DMF, ε = 37). We used the current implementation in Gaussian 09 (ref. 22) of PCM²³ performing a reaction field calculation with the integral equation formalism IEFPCM model.²⁴

Electrochemical investigations

The cyclic-voltammetry experiments were carried out in DMF containing 0.1 M TBAP, under an argon atmosphere in a glass cell thermostatted at 25 °C. The working electrode was a glassy carbon disk (9.64 × 10⁻³ cm²), prepared and activated as already described.²⁵ A Pt plate was used as the counter electrode and Ag/AgCl as the reference electrode. Calibration of the latter was performed by addition of ferrocene at the end of the experiments; in this specific solvent/electrolyte system, the ferricenium/ferrocene redox couple has E° = 0.464 V against the KCl saturated calomel electrode (SCE). Potential values are reported against SCE. We used a CHI 660c electrochemical workstation, and the feedback correction was applied to minimize the ohmic drop between the working and the reference electrodes.

General procedure for the generation of dianions 3a–d

0.1 M solutions of diorganometal 3bNa₂–dNa₂ in dry THF were prepared as described previously.^9,11,13–15 0.1 M solutions of diorganometal 3aNa₂ were generated according to a similar procedure by reacting 1-(2-pyridyl)-2-(4-pyridyl)ethene, 1a, under an Ar atmosphere with an excess of Na metal (4 equiv.) in dry THF at 0 °C during 12 h. After removing the excess of the metal, formation of the desired dianion can be evidenced by quenching separate aliquots of the resulting mixture with H₂O (or D₂O), followed by ¹H- and ¹³C-NMR analyses of the reaction product, characterized as follows:

1-(2-Pyridyl)-2-(4-pirydyl)ethane, 4a. Purified by flash chromatography (CH₂Cl₂/Et₃N = 10 : 1), light yellow oil. R_f = 0.48 (CH₂Cl₂/Et₃N = 10 : 1); Anal. found: C, 78.02; H, 6.81; N = 15.07; C₁₂H₁₂N₂ requires: C, 78.23; H, 6.57; N, 15.21; IR (nujol) 1589 cm⁻¹; ¹H NMR (400 MHz, CDCl₃): δ 3.05–3.13 (4H, m), 7.06 (1H, d, J = 7.6 Hz), 7.09–7.17 (3H, m), 7.57 (td, 1H, J = 7.8, 1.8 Hz), 8.47 (2H, d, J = 4.4 Hz), 8.57 (1H, d, J = 4.0 Hz); ¹³C NMR (100 MHz, CDCl₃): δ 34.9, 38.6, 121.4, 122.9, 123.8, 136.3, 149.4, 149.6, 150.3, 160.1.

General procedure for the reductive dehalogenation of halogenated benzoic acids 5a–c

24 mL of a metal free 0.1 M solution of a diorganometal 3Na₂ (2.4 mmol), was added to a solution of the appropriate benzoic acid 5a–c (1.2 mmol) dissolved in 5 mL of dry THF and chilled at 0 °C. The mixture was vigorously stirred and allowed to reach rt over 12 h, after which time it was quenched by slow dropwise addition of H₂O (15 mL). The organic solvent was evaporated in vacuo and the resulting mixture was extracted with CH₂Cl₂ (3 × 10 mL). The aqueous phase was acidified with 1 N HCl, extracted with CH₂Cl₂ (3 × 10 mL), and the organic phases were collected, washed with H₂O (1 × 10 mL), brine (10 mL), and dried (Na₂SO₄). After filtration and evaporation of the solvent, the resulting mixture was analyzed by ¹H NMR. The outcome of the reaction was assessed by comparing the spectrum with those of commercially available samples. In some reactions, the metal free diorganometal solution was allowed to equilibrate during 1 h at 0 °C in the presence of 0.95 mL of 15-crown-5 (4.8 mmol, 1.05 g) before reacting it with the appropriate benzoic acid under otherwise identical conditions.

General procedure for the metalation of arylacetic acids 5e,f

24 mL of a metal free 0.1 M solution of diorganometal 3Na₂ (2.4 mmol) was added at 0 °C to a solution of the appropriate arylacetic acid 5e,f (1.2 mmol) dissolved in 5 mL of dry THF. The resulting mixture was vigorously stirred for 2 h at the same temperature, after which time it was quenched by slow dropwise addition of H₂O (15 mL), followed by work up of the reaction mixture as described above. Quenching with D₂O was realized by adding 0.75 mL of D₂O to the reduction mixture chilled at 0 °C, followed after 10 minutes stirring by slow dropwise addition of H₂O (15 mL), and work up as described above. The outcome of the reaction was assessed by comparing the ¹H- and ¹³C-NMR spectra of recovered product with those of commercially available samples.

General procedure for the alkylation of arylacetic acids 5e,f

24 mL of a metal free 0.1 M solution of diorganometal 3aNa₂ (2.4 mmol) was added to a solution of the appropriate bromoarylacetic acid 5e,f (1.2 mmol) dissolved in 5 mL of dry THF and chilled at −20 °C, and the resulting mixture was vigorously stirred for 2 h at the same temperature. To the resulting dark brown mixture, chilled at the same temperature, were added 1.5 mmol of the appropriate electrophile. The resulting mixture was vigorously stirred and allowed to reach rt overnight, after which time it was quenched by slow dropwise addition of H₂O (15 mL), followed by work up of the reaction mixture as described above. Crude reaction products were purified and characterized as reported below.

2-(2-Bromophenyl)-3-methylbutanoic acid, 6a. Purified by flash chromatography (petroleum ether/AcOEt = 7 : 3), white powder. R_f = 0.65 (petroleum ether/AcOEt = 7 : 3); Anal. found: C, 51.32; H, 5.06; C₁₁H₁₃BrO₂ requires: C, 51.38; H, 5.10; IR (nujol) 1708 cm⁻¹; ¹H NMR (300 MHz, CDCl₃): δ 0.74 (3H, d, J = 6.9 Hz), 1.12 (3H, d, J = 6.3 Hz), 2.25–2.40 (1H, m), 3.97 (1H, d, J = 10.5 Hz), 7.11 (1H, ddd, J = 7.9, 7.4, 1.5 Hz), 7.29 (1H, td, J = 8.1, 1.2 Hz), 7.54 (2H, dd, J = 8.1, 1.2 Hz); ¹³C NMR (75 MHz, CDCl₃): δ 19.5, 21.2, 32.3, 56.9, 125.7, 127.7, 128.7, 129.0, 132.9, 137.2, 179.4.

2-(2-Bromophenyl)hexanoic acid, 6b. Purified by flash chromatography (petroleum ether/AcOEt = 7 : 3), pale yellow oil. R_f = 0.55 (petroleum ether/AcOEt = 7 : 3); Anal. found: C, 53.19; H, 5.54; C₁₂H₁₅BrO₂ requires: C, 53.15; H, 5.58; IR (neat) 1708 cm⁻¹; ¹H NMR (300 MHz, CDCl₃): δ 0.90 (3H, d, J = 6.9 Hz), 1.24–1.42 (4H, m), 1.75–1.90 (1H, m), 2.05–2.18 (1H, m, CH), 4.24 (1H, d, J = 7.5 Hz), 7.14 (1H, td, J = 7.8, 1.8 Hz), 7.31 (1H, td, J = 7.2, 1.2 Hz); 7.42 (1H, dd, J = 7.8, 1.8 Hz); 7.59 (1H, dd, J = 7.8, 1.2 Hz); ¹³C NMR (75 MHz, CDCl₃): δ 13.8, 22.4, 29.4, 32.5, 49.8, 125.0, 127.7, 128.7, 128.7, 133.0, 138.1, 180.0.

2-(4-Bromophenyl)propanoic acid, 6c²⁶. Purified by flash chromatography (petroleum ether/AcOEt = 7 : 3), R_f = 0.49 (petroleum ether/AcOEt = 7 : 3); white powder, mp 79–81 °C (heptane); IR (nujol) 1704 cm⁻¹; ¹H NMR (400 MHz, CDCl₃): δ 0.90 (3H, d, J = 7.2 Hz), 1.78 (1H, dpent, J = 14.4, 7.2 Hz), 2.08 (1H, dpent, J = 14.4, 7.2 Hz), 3.42 (1H, d, J = 7.2 Hz), 7.15–7.21 (2H, m), 7.42–7.47 (2H, m); ¹³C NMR (100 MHz, CDCl₃): δ 12.0, 26.2, 52.7, 121.4, 139.8, 131.7, 137.3, 179.5.

Results and discussion

Computational results

The structures and energies of the neutral, radical-anion, and dianion forms of compounds a–d in vacuo were modelled at the PBE0/6-311+G(d,p) level.²⁰ The solvent effect on the total molecular energy was estimated for DMF (ε = 37) by using the polarized continuum model²⁴ at the PBE0/6-311+G(d,p)/IEFPCM(DMF) level. As expected for structures in the absence of coordinating ions, all compounds are planar independently of the oxidation state, a geometry that allows attaining maximum electron delocalization.

A single minimum is observed for stilbene 1d and its radical anion 2d and dianion 3d. On the other hand, for stilbazole 1c, 1,2-di(2-pyridyl)ethane 1b, and 1-(2-pyridyl)-2-(4-pyridyl)ethane 1a, as well as for their radical anions and dianions, the minimum energy structures are two, three, and two, respectively. For all compounds the structural difference between the minima consists in the value of the dihedral angle C_e=C_e–C_i=N, where C_e is the C atoms of the ethene moiety (or of its reduced forms) and C_i is the ipso carbon atom of the aromatic ring. This dihedral angle may attain a value of either 0° or 180°, but the latter is favoured in most of the compounds. No exception to such preference is observed when the solvent is included in the calculations, whereas for calculations in vacuo we observed some exceptions for the dianions (Table S1, ESI†); global minima of compounds 1a–d are represented in Fig. S1 (ESI†). Another important difference between the results in vacuo and in solution is the positive formation energy of the dianions in the former, which indicates that without inclusion of the solvent contribution the uptake of two negative charges makes the system less stable.

For compounds 1a–d, the single (EA₁) and double (EA₂) electron affinities were calculated as the negative of the enthalpy change at 0 K and 1 atm,²⁷ from the neutral species to the corresponding radical anions or dianions, respectively. For each compound and charge state, the geometry was individually optimized. The calculated EA values decrease in the order a > b > c > d for both the radical anion and the dianion series no matter whether the calculations refer to vacuo or include the solvent effect (Table 1). These results, therefore, support our hypothesis that more delocalized dianions are milder reducing agents.

Table 1 Electron affinity data for 1,2-diarylethenes 1

Entry	Substrate (eV)	EA1^a (vacuo) (eV)	EA2^b (vacuo) (eV)	EA1^a (DMF) (eV)	EA2^b (DMF) (eV)
a Single electron affinity.b Double electron affinity.
1	1a	1.081	−1.822	2.796	4.961
2	1b	0.849	−2.238	2.700	4.775
3	1c	0.726	−2.475	2.498	4.322
4	1d	0.597	−2.698	2.269	3.847

---

In the stilbene radical anion 2d, the highest spin density sites are found at the ethene moiety, the para position of the benzene rings and, to a smaller extent, the ortho positions. When CH groups in these ring positions are substituted by the more electronegative nitrogen atom (2a, 2b, 2c), the spin density changes by involving the nitrogen atom(s) quite significantly (Fig. S2, ESI†). The corresponding molecular electrostatic potential (MEP) profiles (Fig. 1) show that except for the stilbene radical anion 2d and dianion 3d, in which the region with the highest electrostatic potential corresponds to the ethene portion, for all other ions the highest potential is found in the ring, in the area close to nitrogen atoms, which withdraw the charge from the ethene portion.

Fig. 1 Electrostatic potential mapped on an electron density isosurface (0.02 au A⁻²). Red coloured regions are those with the largest negative potential (varying from −0.19 to −0.46 au, depending on the compound and on the oxidation state); as the potential become less negative or even positive, the colour changes according to the scale reported on the top of the figure. The blue colour on the right-hand side of the scale corresponds to a positive potential having an absolute value equal to that of the largest negative potential (i.e. varying from 0.19 to 0.46 au depending on the compound).

Electrochemical analysis

The redox behaviour of 1a and 1b was studied by cyclic voltammetry (CV) to determine their formal reduction potential (E°) values, and then compare them to those determined previously for 1c and 1d (ref. 18) and the outcome of the calculations. The measurements were carried out in DMF/0.1 M TBAP with a glassy carbon microdisk electrode. For both compounds, the CV pattern shows two reduction peaks (Fig. 2). The first peak corresponds to the reversible formation of the radical anion, whereas the dianion forms at the second peak. At low potential-scan rates (v), the latter is chemically irreversible. As previously observed for 1c and 1d,¹⁸ irreversibility is attributed to protonation of the basic dianion by the residual water present in DMF, a quite common behaviour observed for the electroreduction of many unsaturated and aromatic compounds in formally aprotic solvents.²⁸ By increasing v and for both compounds, an anodic peak associated with the second cathodic peak starts becoming visible, which means that the lifetime of the dianions is now comparable with the CV timescale. The electroreduction of 1a and 1b differs from that of 1c and 1d in two main aspects. First, the formation of the radical anions of 1a and 1b is thermodynamically easier than that of 1c and 1d (Table 2). Second, as opposed to 1c and 1d, the chemical reversibility of the second peak of 1a and 1b starts emerging upon increasing v, which points to a relatively slower protonation rate and thus lower basicity of the corresponding dianions. The comparison between the v dependence of the second peak of 1a and 1b shows that the former yields a less basic dianion than the latter.

Fig. 2 CV curves for the reduction of 2.4 mM 1a (upper graph) and 2.3 mM 1b (lower graph) in DMF/0.1 M TBAP. The current is normalized for the scan rate. The curves were obtained at 0.1, 0.2, 0.5, 1, 2, and 5 V s⁻¹ (larger v values correspond to a more chemically-reversible, negatively-shifted second peak). Glassy-carbon electrode, 25 °C.

Table 2 E° data for 1,2-diarylethenes 1^a

Entry	Substrate	E^°1 (V)	E^°2 (V)
a DMF/0.1 M TBAP, glassy carbon electrode, 25 °C.b Determined as explained in the text.
1	1a	−1.729	−2.188
2	1b	−1.846	−2.261
3	1c	−2.003^b	−2.485^b
4	1d	−2.182^b	−2.700^b

---

For both reduction processes, the E° could be calculated as the half-sum of the anodic and the cathodic peak potentials, using the CV data at high-scan rates for the second electron transfer. The E° values are listed in Table 2 in comparison with the corresponding values for 1c and 1d. The dianions of 1c and 1d are very basic species and thus even in freshly distilled DMF no chemical reversibility can be attained at v as large as 100 V s⁻¹. For both compounds at low scan rates, the pertinent peak potential (E_p) shifts negatively by 30 mV per decade change in v. This value is as expected for a first-order reaction²⁹ and is thus in agreement with protonation of the dianion by residual water present in sufficiently large amount (pseudo-first order conditions). By working in the presence of activated alumina, Parker and co-workers³⁰ could determine the E^°₂ of 1d to be −2.700 V. From such value, use of the observed E_p values and the equation describing the effect of a first-order reaction on E_p¹⁶ yields a pseudo first-order protonation rate constant of 5 × 10⁴ s⁻¹. The reduction of 1c was studied in the same solvent and thus in the presence of the same amount of water. A comparison between the voltammetric behaviour of the second peak of 1b and 1c indicates that introduction of nitrogen makes the dianion less basic, and thus one can safely infer that the pseudo first-order protonation rate constant of 1c is smaller than that of 1d. We can thus bracket 1c's k value as being smaller than 5 × 10⁴ s⁻¹ but still larger than 5 × 10² s⁻¹, the minimum value compatible with the observation of full chemical irreversibility of 1d's second peak at 100 V s⁻¹. This corresponds to the E^°₂ value shown in Table 2, with an associated uncertainty of ±30 mV.

The E° values become more and more negative as one goes progressively from 1a to 1d. Fig. 3 shows that the computed EA values correlate with the corresponding E^°₁ and E^°₂ values of Table 2. The two correlations pertain to the calculations carried out in vacuo (Fig. 3A and C) and in DMF (Fig. 3B and D). This shows that addition of the solvent effect improves the quality of the two correlations very substantially, yielding r² values of 0.994 and 0.999 for the first and second reduction, respectively. In energy terms, the slopes of the plots of Fig. 3 are 1.015 (A), 1.513 (B), 1.185 (C), and 2.148 (D). We note that for this kind of relationship the slope should be one, provided the difference between the solvation energy of the radical anion (or dianion) and the neutral molecule (or radical anion) is kept constant along the series.^31,32 Our results, however, show that the EA values change more rapidly than the experimental values, particularly for the dianions and when the solvent is taken into account. Recently, Davis and Fry³² have shown how an accurate selection of the computational and solvation methods may improve the slope value. On the other hand, the performance of quantum methods on double electron affinity calculations are much less tested and, to the best of our knowledge, there is no previous report on such a correlation between EA₂ and E^°₂. It should be also noted that the presence of a supporting electrolyte, generally neglected in calculations, is known to affect the second reduction potential of even highly delocalized aromatic compounds quite significantly, whereas comparatively the first step is very little affected.²⁸

Fig. 3 Correlation between the computed single (A and B) and double (C and D) electron affinities of compounds 1a–d and the formal potentials E^°₁ and E^°₂. The graphs A and C refer to the calculations performed in vacuo at the PBE0/6-311+G(d,p) level, whereas the graphs B and D refer to the calculations performed by including the presence of DMF (ε = 37). The r² values of the linear regressions are: 0.938 (A), 0.994 (B), 0.877 (C), 0.999 (D). For the error bar associated to the formation of 3c, see text.

On these grounds, although a strict quantitative match between the sensitivities of EA₂ and E^°₂ on the substitution pattern could not be expected, we can still conclude that the observed correlations are valid indicators of the relative reducibility of the investigated compounds and, therefore, the reactivity of the ensuing electrogenerated species as bases and reductants.³³

Generation of 1,2-dipyridyl-1,2-disodioethanes and reaction with halogenated carboxylic acids

To validate the expectations based on the computational and the electrochemical results, we planned a series of experiments meant to check the different reducing properties of dianions 3a and 3b, in the form of their corresponding Na organometals (3aNa₂ and 3bNa₂), easily generated in dry THF (see below). Although for the chemical reactions the solvent of choice was THF, it is worth recalling that we could previously show that the electrochemical results obtained for 1,2-diarylethenes in THF compare well with those obtained in DMF, in terms of both CV pattern and relative position of the alkene reduction peaks.¹⁸ The present electrochemical results (and calculations) were thus considered to provide a reasonable reference also for reactivity data obtained in THF.

Deep red (0.1 M) solutions of 1,2-dipyridyl-1,2-disodioethanes 3aNa₂ and 3bNa₂, were obtained by reacting the appropriate alkene with an excess (4 equiv.) of Na metal in THF during 12 h at 0 °C and filtered from unreacted metal according to a previously described general procedure.^9,12,13 In separate experiments, efficient formation of the diorganosodio derivatives was verified by quenching the reaction mixtures with D₂O, followed by ¹H- and ¹³C-NMR analyses of the crude reaction products; according to this procedure, an almost quantitative incorporation of deuterium at the heteroarylmethyl positions of recovered 1,2-dipyridylethanes was typically observed (Scheme 2).

Scheme 2 Generation of dianions 3a,b and quenching with H₂O (D₂O).

The reaction of a diorganometal with a halogenated carboxylic acid can be foreseen to give rise to a competition between deprotonation and dehalogenation pathways.^13,14,19 By taking into account that whereas a deprotonation reaction requires a dianion versus acid ratio of 0.5 : 1, a reductive cleavage reaction requires a dianion versus aromatic halide ratio of 1 : 1, we investigated the reactivity of the halogenated carboxylic acids of Schemes 3 and 4 in the presence of an excess (2 : 1 molar ratio) of the different dianions, also for the sake of comparison with previous results from our laboratory.

Scheme 3 Reaction of 1,2-diaryl-1,2-disodioethanes with halogenated benzoic acids.

Scheme 4 Reaction of 1,2-diaryl-1,2-disodioethanes 3Na₂ with bromophenylacetic acids.

0.1 M THF solutions of the appropriate carboxylic acids were added dropwise to a solution of the appropriate diorganometal kept at 0 °C under an argon atmosphere. After 12 h the reaction mixtures were elaborated and the carboxylic acids recovered as described in the Experimental section (Scheme 3). Table 3 gathers our findings together with those for the reactions of benzoic acids 5a–c with dianions 3cNa₂ and 3dNa₂ (entries 7–12).¹³ Bearing in mind that the relative ease of reductive dehalogenation of organic halides follows the order I > Br > Cl ≫ F,³⁵ these results confirm the computational/electrochemical prediction of a relatively lower reducing power of dianions 3aNa₂ and 3bNa₂. Whereas reaction with 4-fluorobenzoic acid 5a led to the recovery of the unreacted acid in quantitative yield (Table 3, entries 1 and 4), the reductions of 4-chlorobenzoic acid 5b (Table 3, entry 2 vs. entry 5) and 4-bromobenzoic acid 5c (Table 3, entry 3 vs. entry 6) allow discriminating that dianion 3aNa₂ is a less powerful reductant than 3bNa₂.

Table 3 Reactions of 1,2-diaryl-1,2-disodioethanes 3Na₂ with halogenated benzoic acids 5a–c^a

Entry	Diorganometal	Substrate (X=)	Recovered substrate^b (%)	5d^b (%)
a All reactions were run in dry THF, during 12 h, with a dianion/carboxylic acid molar ratio = 2 : 1.b As determined by ^1H NMR spectroscopic analyses of crude reaction mixtures.c Addition of 2 equiv. of 15-crown-5 to the dianion, followed by 1 h equilibration at 0 °C before addition of the benzoic acid, afforded a comparable result.d Ref. 13.
1	3aNa2	5a, X = F	5a, >99	<1^c
2	3aNa2	5b, X = Cl	5b, 61	39
3	3aNa2	5c, X = Br	5c, 17	83^c
4	3bNa2	5a, X = F	5a, >99	<1
5	3bNa2	5b, X = Cl	5b, 6	94
6	3bNa2	5c, X = Br	5c, <1	>99
7	3cNa2	5a, X = F	5a, 69^d	31^d
8	3cNa2	5b, X = Cl	5b, <1^d	>99^d
9	3cNa2	5c, X = Br	5c, <1	>99
10	3dNa2	5a, X = F	5a, 20^d	80^d
11	3dNa2	5b, X = Cl	5b, <1^c^,^d	>99^c^,^d
12	3dNa2	5c, X = Br	5c, <1^c	>99^c

---

Finally, for the sake of better comparison with the reducing power assessed electrochemically, in which only loose ion pairs may form, some reductions (Table 3, entries 1, 3, 11 and 12) were performed after equilibrating solutions of dianions 3aNa₂ and 3dNa₂ with 15-crown-5 (see the Experimental section), which has a high binding affinity for sodium cation. Importantly, no significant modifications of the relative reducing power of the above mentioned dianions was observed.

To evaluate the competition between the reducing and basic properties of these diorganometals, we investigated their reactivity toward 2-bromo- and 4-bromophenylacetic acids, 5e and 5f, under closely related reaction conditions (Scheme 4). The outcome of these reactions, alternatively quenched with H₂O or D₂O, is summarized in Table 4, together with selected results of comparable reactions run with dianions 3cNa₂ and 3dNa₂.¹⁹ Once again, these results confirm the same scale of reducing power of the diorganometals, i.e., 3aNa₂ < 3bNa₂ < 3cNa₂ < 3dNa₂. They also show that these dianions are all sufficiently basic to promote the α-metalation of the bromophenylacetic acids under investigation, albeit by different degrees.³⁶

Table 4 Reactions of 1,2-diaryl-1,2-disodioethane, 3Na₂, with bromophenylacetic acids 5e,f^a

Entry	Diorganometal	Substrate	Recovered substrate (% D)^b	5g (% D)^b
a All reactions were run in dry THF, during 12 h, with a diorganometal/carboxylic acid molar ratio = 2 : 1.b As determined by ^1H NMR spectroscopic analyses of crude reaction mixtures; the numbers in parenthesis refer to the percentage of hydrogen substitution for deuterium.c Ref. 19.
1	3aNa2	5e	5e, 82 (75)	18
2	3aNa2	5f	5f, 90 (91)	10 (66)
3	3bNa2	5f	5f, 68 (76)	32 (65)
4	3cNa2	5e	5e, 10 (90)^c	90 (89)^c
5	3dNa2	5e	5e, <5^c	>95 (78)^c

---

Through an appropriate balance between its reducing and basic properties, 3aNa₂ appears to be the ideal candidate to achieve the chemoselective generation of the enediolates of bromophenylacetic acids. As a matter of fact, by employing 3aNa₂ as the reagent of choice, we were able to accomplish the α-alkylation of bromophenylacetic acids 5e and 5f under relatively mild reaction conditions, as illustrated in Scheme 5.

Scheme 5 Reactions of 1-(2-pyridyl)-2-(4-pyridyl)-1,2-disodioethane 3aNa₂ with bromoarylacetic acids and reactions with alkyl iodides.

Conclusions

Theoretical calculations and electrochemical analysis of the voltammetric reduction of the parent alkenes allowed us to set up a relative scale for the reducing strength of delocalized dianions. Inclusion of the solvent effect in the calculations improved the correlation between calculations and electrochemical results very significantly. These results nicely confirm and extend our original hypothesis that within a family of compounds the less delocalized dianions are the most powerful reducing agents. This analysis was validated by studying the reactivity of 1,2-dipyridyl-1,2-disodioethanes towards halogenated benzoic and arylacetic acids. In full agreement with the predictions, diorganometal 3aNa₂ displays an appropriate balance between its reducing and basic properties, which allowed us to use it as a selective base for the successful α-alkylation of bromo-substituted arylacetic acids.

Acknowledgements

U. A., L. P. and F. M. gratefully acknowledge financial support from the Regione Autonoma della Sardegna, through the Legge Regionale 07/09/2007 (code CRP-59740).

Notes and references

1. B. W. Knettle and R. A. Flowers II, Org. Lett., 2001, 3, 2321.
2. D. J. Procter, R. A. Flowers II and T. Skrydstrup, Organic Synthesis Using Samarium Diiodide A Practical Guide, Royal Society of Chemistry, Cambridge, 2010.
3. J. M. Yates, J. S. Fell, J. A. Miranda and B. F. Gherman, J. Electrochem. Soc., 2013, 160, G3080.
4. S. B. Bateni, K. R. England, A. T. Galatti, H. Kaur, V. A. Mendiola, A. R. Mitchell, M. H. Vu, B. F. Gherman and J. A. Miranda, Beilstein J. Org. Chem., 2009, 5, 82.
5. S. S. Hanson, E. Doni, K. T. Traboulsee, G. Coulthard, J. A. Murphy and C. A. Dyker, Angew. Chem., Int. Ed., 2015, 54, 11236.
6. H.-B. Pan and C. M. Wai, J. Phys. Chem. C, 2010, 114, 11364.
7. L. Cosimbescu, X. Wei, M. Vijayakumar, W. Xu, M. L. Helm, S. D. Burton, C. M. Sorensen, J. Liu, V. Sprenkle and W. Wang, Sci. Rep., 2015, 5, 14117.
8. These sodium derivatives are delocalized species that can be described either as dianions of the corresponding alkenes or as diorganometals, to better illustrate their reactivity as dinucleophiles towards, e.g., 1,3- and 1,4-dihaloalkanes (see ref. 9 and references cited therein). In a general context, we find it practical to depict them as 1,2-diaryl-1,2-disodioethanes. A similar notation was recently employed to depict, e.g., dianion 1d in D. Seyferth, Organometallics, 2009, 28, 2–33.
9. U. Azzena, G. Dettori, C. Lubinu, A. Mannu and L. Pisano, Tetrahedron, 2005, 61, 8663.
10. H. Braunschweig, C. Claes, A. Alexander Damme, A. Deißenberger, A. R. D. Dewhurst, C. Hörl and T. Kramer, Chem. Commun., 2015, 51, 1627.
11. U. Azzena, M. Pittalis, G. Dettori, S. Madeddu, L. Pisano and E. Azara, Tetrahedron Lett., 2006, 47, 1055.
12. U. Azzena, M. Pittalis, G. Dettori, L. Pisano and E. Azara, J. Organomet. Chem., 2007, 692, 3892.
13. U. Azzena, G. Dettori, S. Mocci, L. Pisano, G. Cerioni and F. Mocci, Tetrahedron, 2010, 66, 9171.
14. U. Azzena and M. Pittalis, Tetrahedron, 2011, 67, 3360.
15. U. Azzena, G. Dettori, L. Pisano, M. Pittalis, G. Mangano, G. Petretto and G. Pintore, Monatsh. Chem., 2012, 143, 601.
16. M. Pittalis, U. Azzena and L. Pisano, Tetrahedron, 2013, 69, 207.
17. U. Azzena, F. Kondrot, L. Pisano and M. Pittalis, Appl. Organomet. Chem., 2012, 26, 180.
18. U. Azzena, L. Pisano, S. Antonello and F. Maran, J. Org. Chem., 2009, 74, 8064.
19. U. Azzena, G. Dettori, M. Pittalis and L. Pisano, Tetrahedron, 2011, 67, 3470.
20. C. Adamo and V. Barone, J. Chem. Phys., 1999, 110, 6158–6179.
21. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865.
22. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian 09, Revision A.02, Gaussian, Inc., Wallingford, CT, 2009.
23. G. Scalmani and M. J. Frisch, J. Chem. Phys., 2010, 132, 114110.
24. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999.
25. S. Antonello, N. V. Perera, M. Ruzzi, J. A. Gascón and F. Maran, J. Am. Chem. Soc., 2013, 135, 15585.
26. C. E. Stivala and A. Zakarian, J. Am. Chem. Soc., 2011, 133, 11936.
27. S. G. Lias and J. E. Bartmess, Gas Ion Thermochemistry, in NIST Chemistry WebBook, NIST Standard Reference Database Number 69, ed. P. J. Linstrom and W. G. Mallard, National Institute of Standards and Technology, Gaithersburg MD, 20899, http://www.webbook.nist.gov/chemistry/ion/#EA, (retrieved May 6, 2016).
28. B. S. Jensen and V. D. Parker, J. Am. Chem. Soc., 1975, 97, 5211.
29. R. S. Nicholson and I. Shain, Anal. Chem., 1964, 36, 706.
30. B. S. Jensen, R. Lines, P. Pagsberg and V. D. Parker, Acta Chem. Scand., Ser. B, 1977, 31, 707.
31. (a) P. Kebarle and S. Chowdhury, Chem. Rev., 1987, 87, 513; (b) H. Shalev and D. H. Evans, J. Am. Chem. Soc., 1989, 111, 2667; (c) R. S. Ruoff, K. M. Kadish, P. Boulas and E. C. M. Chen, J. Phys. Chem., 1995, 99, 8843; (d) D. H. Evans, Chem. Rev., 2008, 108, 2113.
32. A. P. Davis and A. J. Fry, J. Phys. Chem. A, 2010, 114, 12299.
33. A good linear correlation was also found for calculated LUMO energies of compound 1a–d (Table S2, ESI†) and the respective E^°₁ with r² values of 0.931 for the in vacuo calculations, and 0.999 for those with the PCM solvent, in agreement with what already reported for different classes of organic compounds.³⁴ However, extension of such an approach to the SOMO energies (Table S2, ESI†) do not show a good correlation with respect to the difference E^°₂ − E^°₁, with r² values for both in vacuo and PCM solvent calculations of 0.385 and 0.678, respectively. Accordingly, these data strongly suggest that the electron affinity is a better index for reduction potential predictions involving dianions.
34. R. S. Assary, F. R. Brushett and L. A. Curtiss, RSC Adv., 2014, 4, 57442 , and references therein.
35. F. Alonso, I. P. Beletskaya and M. Yus, Chem. Rev., 2002, 102, 4009.
36. Substituents located on the aromatic positions of 1,2-diaryl-1,2-disodioethanes affect their reducing power but should conceivably affect also their basicity. Our results do not highlight such an effect probably because of the relatively high acidity of arylacetic acids employed. From a qualitative point of view, it is worth noting that whereas the pK_a of the 2^nd ionization of phenylacetic acid (as Li carboxylate in THF: HMPA = 80 : 20) is 30.2,³⁷ that of halo-substituted arylacetic acids should be somewhat more acidic. Additionally, we observed that in THF 2dNa₂ deprotonates fluorene (pK_a in THF = 22.4 ³⁸ or 24.4 ³⁹), 4-benzylpyridine (pK_a in THF = 25.0)³⁹ and xanthene (pK_a in THF = 31.4),³⁹ but does not deprotonate diphenylmethane (pK_a in THF = 33.0).³⁸ On the other side, 2aNa₂ deprotonates fluorene, but does not deprotonate xanthene, thus supporting the expected parallelism between the relative basicities and reducing powers of our dianions.
37. P. Renaud and P. M. A. Fox, J. Am. Chem. Soc., 1988, 110, 5705.
38. A. Streitweiser Jr, D. A. Bors and M. J. Kaufman, J. Chem. Soc., Chem. Commun., 1983, 1394.
39. R. R. Fraser, T. S. Mansour and S. Savard, J. Org. Chem., 1985, 50, 3232.

---

Footnote

† Electronic supplementary information (ESI) available: Sum of the electronic and zero point energy for minimum energy structures of compounds 1a–d, 2a–d, 3a–d. Structure of global minima for compounds 1a–d. Spin density for compounds 2a–d. ¹H NMR and ¹³C NMR spectra of compounds 4a, 4a_d2, 6a and 6b. See DOI: 10.1039/c6ra03303b

---