Stable Radical Trianions from Reversibly Formed Sigma-Dimers of

Article pubs.acs.org/JPCA

Stable Radical Trianions from Reversibly Formed Sigma-Dimers of Selenadiazoloquinolones Studied by In Situ EPR/UV−vis Spectroelectrochemistry and Quantum Chemical Calculations Andrej Staško,† Karol Lušpai,† Zuzana Barbieriková,† Ján Rimarčík,† Adam Vagánek,† Vladimír Lukeš,† Maroš Bella,‡ Viktor Milata,§ Michal Zalibera,†,∥ Peter Rapta,*,† and Vlasta Brezová*,† †

Institute of Physical Chemistry and Chemical Physics, Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava, Radlinského 9, SK-812 37 Bratislava, Slovak Republic ‡ Institute of Chemistry, Slovak Academy of Sciences, Dúbravská cesta 9, SK-845 38 Bratislava, Slovak Republic § Institute of Organic Chemistry, Catalysis and Petrochemistry, Faculty of Chemical and Food Technology, Slovak University of Technology in Bratislava, Radlinského 9, SK-812 37 Bratislava, Slovak Republic ∥ Institute of Physical and Theoretical Chemistry, Graz University of Technology, Stremayrgasse 9, A-8010 Graz, Austria S Supporting Information *

ABSTRACT: The redox behavior of the series of 7substituted 6-oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-h]quinolines and 8-substituted 9-oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-f ]quinolines with R7, R8 = H, COOC2H5, COOCH3, COOH, COCH3, and CN has been studied by in situ EPR and EPR/UV−vis spectroelectrochemistry in dimethylsulfoxide. All selenadiazoloquinolones undergo a one-electron reduction process to form the corresponding radical anions. Their stability strongly depends on substitution at the nitrogen atom of the 4-pyridone ring. The primary generated radical anions from N-ethyl-substituted quinolones are stable, whereas for the quinolones with imino hydrogen, the initial radical anions rapidly dimerize to produce unusually stable sigmadimer (σ-dimer) dianions. These are reversibly oxidized to the initial compounds at potentials considerably less negative than the original reduction process in the back voltammetric scan. The dimer dianion can be further reduced to the stable paramagnetic dimer radical trianion in the region of the second reversible reduction step. The proposed complex reaction mechanism was confirmed by in situ EPR/UV−vis cyclovoltammetric experiments. The site of the dimerization in the σ-dimer and the mapping of the unpaired spin density both for radical anions and σ-dimer radical trianions with unusual unpaired spin distribution have been assigned by means of density functional theory calculations.

■

demonstrated biological impact, as previously reviewed.13−15 Comprehensive studies of potential biological/photobiological activities of selenadiazoloquinolones, as well as their capacity to promote the formation of photoinduced reactive oxygen species were recently performed in our group.11,16,17 The ability of selenadiazoloquinolones to generate paramagnetic intermediates and singlet oxygen upon photoexcitation was studied by means of EPR spectroscopy using the spin-trapping technique and oxidation of sterically hindered amines. Ultraviolet A photoexcitation of the selenadiazoloquinolones in dimethylsulfoxide (DMSO) or acetonitrile resulted in the activation of molecular oxygen generating the superoxide radical anion and singlet oxygen. We demonstrated the cytotoxic/photocytotoxic impact of selenadiazoloquinolones on murine and human

INTRODUCTION

The four-generation family of antibacterial agents with 4oxoquinoline structure represents an important tool in the treatment of bacterial infections.1−3 In particular, the latest fourth-generation fluoroquinolones reveal a broad-spectrum bactericidal activity against gram-positive and gram-negative pathogens coupled to improved pharmacokinetic properties.1,4 However, because of the increased bacterial resistance to quinolone antibiotics, there is a permanent interest in the synthesis of novel 4-oxoquinoline derivatives possessing an extended number of biological activities.5,6 Recently, quinoloneannulated derivatives with an additional heterocyclic ring containing nitrogen, sulfur, or selenium atoms were synthesized, and their biological effects were examined.7−10 Our effort to prepare selenaheterocyclic compounds resulted in a successful utilization of the Gould−Jacobs reaction in the efficient synthesis of selenadiazoloquinolones.11,12 The presence of selenium in a molecule can provide new attributes, and some of the synthesized selenaheterocyclic compounds © XXXX American Chemical Society

Received: July 23, 2012 Revised: September 13, 2012

A

dx.doi.org/10.1021/jp307270b | J. Phys. Chem. A XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry A

Article

formation of paramagnetic σ-dimeric trianions available so far. In this work, the proposed complex reaction mechanism for reduction of selenadiazoloquinolones, including reversible dimerization of the corresponding radical anions and the formation of stable σ-dimer radical trianions, is studied by a series of in situ EPR/UV−vis cyclovoltammetric experiments. The site of dimerization in the σ-dimer and the mapping of the unpaired spin density both for the radical anions and the σdimer radical trianions are interpreted with the help of quantum chemical calculations.

cancer cell lines using 7-acetyl-6-oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-h]quinoline16 (derivative 5h in Scheme 1). Scheme 1. Schematic Structure of the Investigated 6-Oxo6,9-dihydro[1,2,5]selenadiazolo[3,4-h]quinolines (1h−E5h) and 9-Oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-f ]quinolines (1f−E1f)

■

EXPERIMENTAL AND THEORETICAL METHODS The investigated selenadiazoloquinolones (Scheme 1) were synthesized in our laboratories as described in refs 11 and 12. All solutions were prepared in DMSO (SeccoSolv, Merck); tetrabutylammonium hexafluorophosphate (TBAPF6) from Fluka, dried in a vacuum oven for 16 h at 70 °C, served as the supporting electrolyte. Cyclovoltammetric experiments were performed at room temperature under an inert argon atmosphere. A standard three-electrode arrangement with a platinum wire as the working electrode, a platinum wire as the counter electrode, and a silver wire pseudoreference electrode was used. The selenadiazoloquinolone concentration in DMSO was 10−4 and 10−3 M in cyclic voltammetry (CV) and spectroelectrochemical experiments, respectively. A Heka PG285 (Lambrecht, Germany) potentiostat served for potential control in both voltammetric and spectroelectrochemical studies. The spectroelectrochemical in situ EPR/UV−vis cyclovoltammetric experiments were carried out under an argon atmosphere in a special flat spectroelectrochemical cell (0.1 mm path length), suitable for an optical transmission EPR resonator (ER 4104OP) of an X-band EPR spectrometer EMX (Bruker, Germany). The working electrode was a laminated Pt mesh with a small hole in the foil coincident with the light beam and limiting the active surface area of the electrode. A Pt wire auxiliary (counter) electrode and an Ag wire pseudoreference electrode were used. Ferrocene (Sigma-Aldrich) served as the internal potential standard, and all potentials were referred versus a ferrocenium/ferrocene (Fc+/Fc) redox couple. The optical EPR resonator cavity was connected to the diode-array UV−vis spectrometer Sentronic S2000 by optical fibers. A deuterium-halogen lamp DH 2000 (Sentronic, Germany) was used as a light source. To obtain a higher yield of paramagnetic ions and therewith a better quality of EPR spectra, we performed the amperostatic electrochemical in situ EPR experiments in DMSO solutions containing 10−3 M selenadiazoloquinolone and 0.05 M TBAPF6. The solutions, prepared immediately before the EPR experiments, were purged with argon and inserted in a Varian electrolytic cell equipped with a platinum mesh. The electrolytic cell was polarized in an amperostatic mode directly in the cylindrical EPR cavity TM-110 (ER 4103 TM), and the EPR spectra were measured in situ. The g values of the radical species were determined using a 4-hydroxy-2,2,6,6-tetramethylpiperidine N-oxyl (Tempol; Aldrich) reference or using a builtin magnetometer. The experimental EPR spectra were analyzed and simulated by the Bruker software WinEPR and SimFonia and the Winsim2002 software freely available from the website of the National Institute of Environmental Health Sciences (NIEHS).27 All quantum chemical calculations were performed using the Gaussian 03 program package.28 The optimal geometries of the

In the frame of our systematic investigations on novel synthesized quinolones, we also followed the cathodic reduction of selenadiazoloquinolones by in situ spectroelectrochemistry. This method combines standard electrochemistry with different spectral methods; that is, the products of electrochemical reactions on a working electrode are simultaneously detected by spectroscopic techniques.18−20 Spectroelectrochemistry is thus a very useful tool for the investigation of complex reaction mechanisms with a variety of consecutive reactions. An understanding of the electrochemical reduction mechanism of selenadiazoloquinolones can shed more light on their reactions coupled to electron-transfer processes. Here we report on the redox behavior of 7-R-6-oxo-6,9dihydro[1,2,5]selenadiazolo[3,4-h]quinolines (1h−6h) and 8R-9-oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-f ]quinolines (1f− 6f; R = H, COOC2H5, COOCH3, COOH, COCH3 and CN) and their N-ethyl derivatives (E1h−E5h and E1f), summarized in Scheme 1, investigated by in situ EPR and EPR/UV−vis spectroelectrochemistry in DMSO solutions. Selenadiazoloquinolones undergo a one-electron reduction to form the corresponding radical anions. Their stability strongly depends on the substitution at the nitrogen atom of the 4-pyridone ring. The consecutive reactions of the primary generated radical anions include various reaction pathways. Among them, the formation of covalently bonded σ-dimers of radical anions represents a generally accepted alternative, and there are already several examples of reports on the reversibility of the dimerization reactions.21−25 Such a dimerization process obviously exhibits a large equilibrium constant, and for some cases the reduction of the dimer to its radical trianion was proposed to achieve a good match with the experimental cyclovoltammetric data.26 However, to our knowledge there are no direct spectroelectrochemical studies concerning the B



Article

electrically neutral and anionic structures were obtained using the density functional theory (DFT) method with B3LYP (Becke’s three-parameter Lee−Yang−Parr)29 functional without any constraints (energy cutoff of 10−5 kJ mol−1, final rootmean-square (rms) energy gradient under 0.01 kJ mol−1 Å−1). The solvent effect of DMSO was included using the IEFPCM (polarizable continuum model based on the integral equation formalism variant) approach.30 The atomic orbitals for the C and H atoms were 6-31G* and for the Se, N and O atoms the augmented 6-31+G* basis sets were used.31 The optimized structures were confirmed to be real minima by frequency analysis (no imaginary frequencies). The spin contaminations of radicals were found in the range of 0.759 to 0.765. After the annihilation of the first spin contaminant, they dropped to the correct value of 0.750. On the basis of optimized geometries, the vertical electronic transition energies and oscillator strengths between the initial (S0) and final excited (Sn) states were computed by the time-dependent (TD)-B3LYP method.32

Table 1. Electrochemical Data from Cyclic Voltammetry of Investigated Selenadiazoloquinolones in Dimethylsulfoxide (in volts versus Fc+/Fc, Scan Rate 0.1 V s−1)

■

RESULTS AND DISCUSSION The CV of the selenadiazoloquinolones possessing an ethyl substitution at the nitrogen atom of the 4-pyridone ring (samples E1h−E5h and E1f) shows only one reversible peak in the cathodic part, as illustratively shown for E1h in Figure 1 (red line).

a

selenadiazoloquinolone

E1/2(red1)

E1h E2h E3h E4h E5h E1f 1h 2h 3h 4h 5h 6h 1f 2f 3f 4f 5f 6f

−1.59a −1.49a −1.50a −1.38a −1.48a −1.68a −1.53b −1.45b −1.42b −1.34b −1.41b −1.36b −1.53b −1.57b −1.52b −1.40b −1.51b −1.40b

E1/2(red2)

−1.58c

−2.02c −1.89c −1.89c −1.75c −1.87c −1.85c −2.08c −1.84c −1.86c −1.80c −1.87c −1.83c

Reversible. bIrreversible. cQuasireversible.

reduction of the samples E1h−E5h. The EPR signals, with a g factor in the range of 2.0045 to 2.0050, reveal resolved hyperfine splittings coming from the two nitrogen nuclei of the selenadiazole ring and several hydrogens. (See Table 2.) The high stability of the corresponding radical anions can be attributed to the steric effect of the bulky ethyl substituent at the nitrogen atom of the 4-pyridone ring. The values of the hyperfine coupling constants (hfcc) elucidated from experimental EPR spectra are in good correlation with theoretical data calculated by DFT for the corresponding radical anions, as will be stated in more detail below. A completely different redox behavior was observed for selenadiazoloquinolones possessing the imino hydrogen at the 4-pyridone ring (samples 1h−6h and 1f−6f). The characteristic cyclic voltammogram for this series of selenadiazoloquinolones is shown in Figure 1 (black line) for 1h reduction in TBAPF6/ DMSO. The first reduction step is irreversible and is followed by a quasireversible reduction step at more negative potentials. All samples from series 1h−6h and 1f−6f show a similar redox behavior upon reduction. However, in the region of the first irreversible voltammetric peak, no EPR signal was observed in the EPR spectroelectrochemical experiment. Only at the higher potentials of the second-reversible or quasireversible reduction voltammetric peak were well-defined EPR spectra of cathodically generated radicals detected. Interestingly, the hyperfine pattern of these radicals formed at the second voltammetric peak is very close to the hfcc of the corresponding radical anions of the N-ethyl derivatives (Figure 2S in the Supporting Information, Table 2). To solve this puzzle, we chose a representative quinolone derivative 1f to study its redox mechanism in detail by in situ EPR/UV−vis cyclovoltammetry. Additionally, DFT studies were performed to clarify an unusual spin density distribution in paramagnetic species formed at the second voltammetric peak. We propose that for all of the selenadiazoloquinolones investigated the first reduction step is associated with the formation of a radical anion of the parent compound. However, the stability of this radical anion for samples 1h−6h and 1f−6f

Figure 1. Cyclic voltammograms of E1h (red line) and 1h (black line) in 0.2 M TBAPF6 in DMSO (scan rate 0.1 V s−1).

The electrochemical data derived from CV of all the selenadiazoloquinolones under study are summarized in Table 1. Figure 2 shows the cyclic voltammogram (Figure 2a) along with selected UV−vis spectra (Figure 2b) detected simultaneously during the in situ reduction of E1f in the region of the first reversible voltammetric peak. Two new absorption bands with maxima at 415 and 573 nm increase during the reduction of E1f in 0.2 M TBAPF6/DMSO solution. (The isosbestic point is observable at 392 nm.) All UV−vis spectra detected in situ during the one-electron reduction of E1f in both the forward and back scans are presented in Figure 1S in the Supporting Information. Both optical bands can be clearly assigned to the generated radical anion because their intensity correlates well with the EPR intensity of the simultaneously taken EPR spectra. (See the representative EPR spectrum in the inset of Figure 2b.) Figure 3 shows the experimental and simulated EPR spectra detected during the amperostatic C



Article

with the imino group is too low, in comparison with their Nethyl substituted analogues; therefore, its EPR spectrum was not observed in EPR/UV−vis spectroelectrochemistry at the first reduction peak. During the second and the third cyclovoltammetric cycles for the same sample the current of the first irreversible CV peak decreases (for an illustration, see the reduction of sample 1h in Figure 3S in the Supporting Information). This indicates consecutive chemical reactions of the radical anions previously formed near the electrode surface. We assume that upon the cathodic reduction the initially generated radical anions from the 1h−6h and 1f−6f samples rapidly dimerize to the corresponding EPR-silent stable σ-dimer dianions (Scheme 2). Similar indications have already been reported in the literature by other systems.21,24,26,33,34 More information was obtained from in situ EPR/UV−vis electrochemistry of 1f quinolone in 0.2 M TBAPF6/DMSO solution (scan rate 4 mV s−1) performed in four consecutive voltammetric cycles (Figures 4 and 5 and Figure 4S in the Supporting Information). Results from the first cycle are presented in Figure 4a,b. Here the first irreversible cathodic peak (from −1.46 to −1.82 V) is coupled to the appearance of a new dominating optical band at 450 nm and a decrease in the pristine quinolone band at 375 nm (Figure 4a). No radical formation was observed in simultaneous EPR measurements. In the region of the second quasireversible voltammetric peak (from −2.06 to −2.42 V, Figure 4b), the band at 450 nm decreases, and two new bands at 335 and 380 nm are evident. This is simultaneously coupled to the development of the EPR signal (inset in Figure 4b). The first voltammetric cycle (inset in Figure 4a) thus represents two reduction processes, namely, the intermediate formation of unstable radical anions and their conversion to the stable EPR silent σ-dimer dianions (the band at 450 nm), at the first peak, and the consecutive reduction of the dimer to its stable σ-dimer radical trianion, at the second peak (the bands at 335 nm, 380 nm) (Scheme 2 and Figures 4a,b, respectively). Significantly different behavior was observed in the second voltammetric cycle. In comparison with the first one, the first peak is missing here and only the second peak is obtained in the voltammogram (Figure 5). The absence of the first peak

Figure 2. (a) Cyclic voltammogram of E1f in 0.2 M TBAPF6 in DMSO (scan rate 4 mV s−1). (b) Selected UV−vis spectra simultaneously detected during the in situ reduction of E1f in the region of the first reversible voltammetric peak. (The corresponding potentials in V versus Fc+/Fc are marked in the Figure.) Inset: representative EPR spectrum of E1f•− radical anion observed simultaneously upon reduction (magnetic field sweep, SW = 4 mT).

Figure 3. Experimental (solid black lines) and simulated (solid red lines) EPR spectra of (a) E1h•− (SW = 5 mT); (b) E2h•−, E3h•−; (c) E4h•−; and (d) E5h•− (SW = 4 mT) radical anions observed upon cathodic reduction of the parent compounds in the region of the first reversible voltammetric peak in 0.2 M TBAPF6 in DMSO. D



Article

Table 2. Hyperfine Coupling Constants (hfcc) Elucidated from EPR Spectra of Paramagnetic Species (g ∈ 2.0045 to 2.0050) Observed upon Cathodic Reduction of Selenadiazoloquinolones in DMSO hyperfine coupling constants, mT aN

selenadiazoloquinolone Radical Anion E1h E2h E3h E4h E5h E1f σ-Dimer Radical Trianion 1h 2h 3h 4h 5h 6h 1f 2f or 3f 4f 5f 6f

aH

0.650, 0.629, 0.628, 0.630, 0.616, 0.611,

0.528 0.500 0.492 0.439 0.481 0.408, 0.0141

0.107, 0.189, 0.190, 0.191 0.196 0.382,

0.099 0.068 0.069

0.619, 0.626, 0.624, 0.620, 0.616, 0.627, 0.643, 0.602, 0.604, 0.582, 0.596,

0.608 0.562 0.559 0.542 0.541 0.559 0.498 0.461 0.472 0.446 0.464

0.205, 0.165, 0.157, 0.141, 0.146, 0.163, 0.288, 0.232, 0.255, 0.272, 0.266,

0.092, 0.122 0.135 0.136, 0.143, 0.126 0.173, 0.246 0.216, 0.213, 0.218

0.209, 0.111; 0.088, 0.031 (CH2), 0.031 (CH3) 0.085

0.048 0.060 (CH3) 0.128, 0.053 0.062 0.066 (CH3)

the potential up to +0.4 V versus Fc+/Fc. Interestingly, the σdimer dianion can be reversibly reoxidized, back to the neutral initial compound at strongly anodically shifted potentials (Figure 4c), which clearly indicates that the proposed dimer dianion is of the σ-type dimer.26 As seen from Figure 4c, upon oxidation, a decrease in the σ-dimer optical band at 450 nm was accompanied by the recovery of the optical band of the initial compound at 375 nm, confirming the reversible reoxidation of the σ-dimer dianion to the initial molecules. This was also unambiguously confirmed in the fourth voltammetric cycle, performed immediately after the third scan, where a cyclovoltammetric response analogous (Figure 4Sd in the Supporting Information) to that observed in the first voltammetric scan (Figure 4Sa in the Supporting Information) was found. It should be noted that these processes are not fully reversible under the experimental conditions used and consequently the voltammetric peaks in the fourth scan are slightly lower compared with the first one. Ideally, the generation of the σ-dimer dianion at the first reduction step and its single-electron reduction to a σ-dimer radical trianion at the second step should result in a smaller current at the second reduction peak (two times smaller than the current of the first reduction peak, see inset (a) in Scheme 1S in the Supporting Information). However, a different current ratio was observed in our experimental cyclic voltammograms (Figures 1, 4, and 5 and Figures 3S and 4S in the Supporting Information). Therefore, we performed a simulation analysis of the experimental cyclic voltammograms taking into account the complex reaction mechanism as is shown in Scheme 1S in the Supporting Information. Importantly, the adsorption phenomena for initial compound should be taken into account as well. In many cases, for investigated selenadiazoloquinolones we observed upon cycling such phenomena, particularly going back into the anodic region. Taking into consideration the overall reaction mechanism and the partial adsorption of the initial compound on the electrode, we obtained very similar response, as observed experimentally for obvious scan rates (e.g., for 0.1

Scheme 2. Reaction Mechanism Proposed for the Electrochemical Reduction in Aprotic Media of Selenadiazoloquinolones Possessing Imino Hydrogen at the 4-Pyridone Ring

reflects the previous reduction of pristine selenadiazoloquinolone to its very stable σ-dimer dianion in the entire layer close to the electrode. The appearance of the second peak documents the presence of the σ-dimer dianion, remaining from the previous cycle, and its reduction to the σ-dimer radical trianion (Figure 5b). Formation of the radical trianion was confirmed by EPR spectroscopy and is additionally reflected in the changes in UV−vis spectra. The spectrum of pristine quinolone evident at the beginning in Figure 5a is missing in Figure 5b, in agreement with the absence of the first reduction peak of pristine quinolone in the voltammogram. The spectra evident in Figure 5b represent the transformation of σ-dimer dianion to its σdimer radical trianion. In the third voltammetric cycle, the products formed upon reduction in the previous scans have been reoxidized, increasing E



Article

Figure 5. UV−vis spectra of the 1f sample taken during the spectroelectrochemical experiment (scan rate 4 mV s−1, Pt-mesh working electrode) along with the cyclic voltammograms (insets) during (a) the first and (b) immediately afterward, the second voltammetric scan.

solvent or electrolyte impurities, although at negligible concentrations. The B3LYP calculations of the selected neutral monomers E1h, E1f, 1h, and 1f showed that the selenadiazoloquinolone part is planar, and the longest bond lengths are between the selenium and neighboring nitrogen atoms (Figure 5S in the Supporting Information). The perpendicular orientation of the ethyl group with respect to the aromatic skeleton was found for the most stable conformation of E1h and E1f. The first ten TDB3LYP vertical singlet excitation energies for E1h, E1f, 1h, and 1f are summarized in Table 1S in the Supporting Information. The nonsubstituted structure 1f exhibits the largest excitation energy for the lowest S0→S1 transition at 3.15 eV (394 nm), whereas the 1h molecule has for the lowest transition the energy of 2.82 eV (440 nm). The ethyl substitution at the nitrogen atom of the 4-pyridone ring (position 9 or 6) leads to the small oscillator strength increase (2.6% for E1f and 17.1% for E1h) and to the bathochromic (red) shift of the corresponding optical band. These results are in good agreement with the spectral measurements performed in DMSO solution16 where, for example, the absorption maximum for the lowest experimental optical absorption band for neutral 1f is at 375 nm and for its N-ethyl substituted analogue E1f the maximum was observed at 415 nm. (See Figures 2b and 4a.) The next electronic transitions with significant oscillator strengths are S0→S4 (at 4.00 eV for E1h and 1h) and S0→S3 (at 3.69 eV for E1f and at 3.74 eV for 1f). As illustrated in Figure 6S in the Supporting Information, the excitation from the highest occupied molecular orbital (HOMO) to the lowest unoccupied orbital (LUMO)

Figure 4. In situ EPR/UV−vis spectroelectrochemistry for the sample 1f in 0.2 M TBAPF6 in DMSO (scan rate 4 mV s−1). (a) UV−vis spectra detected simultaneously during the in situ reduction in the region of the first irreversible voltammetric peak. (The corresponding potentials in V versus Fc+/Fc are marked in the Figure.) Inset: corresponding voltammogram with potential regions of the first and second peaks color highlighted. (b) UV−vis spectra measured in the region of the second quasireversible voltammetric peak. (The corresponding potentials in V versus Fc+/Fc are marked in the Figure; for voltammogram, see the inset in panel a.) Inset: representative EPR spectrum (SW = 5 mT) of paramagnetic species observed simultaneously upon reduction at the second peak. (c) UV− vis spectra detected during the back reoxidation in the selected potential region. (The corresponding potentials in V versus Fc+/Fc are marked in the Figure.) Inset: corresponding voltammogram with color-highlighted potential region where spectra were taken.

or 0.5 V s−1). Also, the broad shape of the first reduction peak obtained by the simulation correlates very well with our experimental CVs (see inset (b) in Scheme 1S and Figure 3S in the Supporting Information). At very low scan rates, the voltammetric peaks are less reversible, which probably reflects the additional side-reactions of generated charged species with F



Article

Figure 6. Mulliken B3LYP(IEFPCM = DMSO) spin densities in atomic units for (a) 1f•−, (b) 1h•− radical anions and stable σ-dimer radical trianion, (c) (1f)2•3−, and (d) (1h)2•3−, respectively.

dominates in the S0→S1 transition. The evaluated percentage is higher than 85%. The HOMOs of E1h, 1h, and 1f are uniformly delocalized over the carbon−carbon bonds between the atoms nos. 9b−9a−5a−6−7 and 5−5a−9a−9−8 and the CO bonds, respectively. In the case of 1f, the lobes of the HOMO are also over the selenium atom. The N-ethyl substitution dramatically changes the electron delocalization in the vicinity of the ethyl carbon atoms only for the E1f molecule. The shapes of the LUMOs differ considerably from the HOMOs. They are delocalized mostly along the bonds of the selenadiazole moiety and confirm that this part of the molecule represents the redox site upon cathodic reduction. Interestingly, the significant role of the ethyl group’s presence with respect to the changes in electronic structure is clear from the comparison of E1f and 1f orbitals (Figure 6S in the Supporting Information). Only for the E1f molecule is the electron delocalized over the CO group and C(8)−C(7)− N(6) atoms. The shape of the LUMOs indicates also the possible distribution of the unpaired spin over the selenadiazole moiety after the addition of the electron to the LUMO. Indeed, the Mulliken electron spin density distribution computed for the radical anions 1f and 1h is expected on the N(1)−Se(2)−N(3) atoms, as depicted in Figures 6a,b. The same holds for N-ethyl substituted selenadiazoloquinolones, and this finding corresponds very well to the B3LYP and experimental EPR splitting constants found for the E1f derivative (Figure 7). The absolute values of the theoretical EPR hfcc’s are linearly dependent with the experimental data evaluated by simulation of the experimental EPR spectra (Figure 7). As already discussed above, the cathodic reduction of 1h and 1f monomers leads to the formation of the stable diamagnetic σ-dimer dianions (1h)22− and (1f)22−, respectively. (See Scheme 2.) In this context, we have used the set of twelve possible starting geometries of σ-dimers for each monomer. However, only two (the most stable) conformations were obtained. These preferable conformations are bonded through the carbon atom numbers 7 and 8 at the 4-pyridone ring, and

the geometries of the selenadiazoloquinolones are quasisymmetric. The conformations of the indicated σ-dimers are distinguished according to the mutual orientation of the CO groups. Our calculations showed that the cis orientation of the σ-dimer dianions for 1h and 1f is more stable than the trans orientation, with energy differences of about 7.7 and 2.8 kJ mol−1, respectively. With respect to this fact, we next performed calculations of the σ-dimer dianions and the σ-dimer radical trianions for the most stable cis conformation. The LUMOs depicted for the σ-dimers’ dianions, consisting of 1h or 1f monomer units, indicated that the addition of the next electron upon the cathodic reduction is possible for both monomer moieties (Figure 7S in the Supporting Information). However, the optimal geometry of the σ-dimer radical trianion, based on the IEFPCM approach, is not quasisymmetric. Consequently, the substantial differences in the lengths between the relevant bonds within the selenadiazoloquinolones were found. This nonsymmetry has a direct influence on the Mulliken spin densities. As is shown in Figure 6c,d, the presence of unpaired electron spin density is expected mainly on the N(1)−Se(2)−N(3) atoms of one selenadiazoloquinolone moiety. This situation is identical to the spin densities obtained for radical monoanion, in good agreement with the experimental observations discussed in EPR spectroelectrochemical experiments. The relevance of the interpretation and identification of the experimental bands in the UV−vis spectra observed upon 1f reduction (see Figure 4) is also supported qualitatively by the theoretical spectra given in Figure 8S in the Supporting Information. For example, a new intense TDB3LYP vertical transition for the (1f)22− σ-dimer in the UV−vis region at 516 nm was found in accordance with experiment. This band corresponds to the HOMO−1 to LUMO transition (Figure 9S in the Supporting Information). Interestingly, this transition can be attributed to the charge transfer from the central part of the dimer to the N(3)−C(3a)−C(4)−C(5) atoms on benzoselenadiazole moieties. For the radical trianion (1f)2•3−, an intense TD-B3LYP vertical transition appears at 385 nm, which matches well the experimental observation of G



Article

radical anions that rapidly dimerize to give the unusually stable σ-dimer dianions. These dianions can be reversibly reoxidized back to the initial neutral monomers at a potential strongly positively shifted with respect to the monoanion reduction peak. Stable σ-dimer dianion can be further reduced to the stable σ-dimer radical trianion in the region of the second reversible or quasireversible reduction step. The site of the dimerization in the σ-dimer (at carbon atoms 7 and 8 of the 4pyridone moieties) as well as the distribution of the unpaired electron spin density have been elucidated by DFT calculations. Even though the LUMOs estimated for the optimized model σdimer dianions indicate that the addition of the next electron is possible for both monomeric moieties, the optimal geometry of paramagnetic σ-dimer radical trianions based on the IEFPCM approach is not symmetric. Consequently, the presence of unpaired electron spin density was unambiguously confirmed to be mainly on the N(1)−Se(2)−N(3) atoms of a single selenadiazoloquinolone unit within the dimer. The situation is identical to the spin densities obtained for the corresponding monomer radical anions. This finding explained adequately the unusual spin distribution in radical ions formed in the region of the second reduction voltammetric peak for selenadiazoloquinolones possessing the imino hydrogen at the 4-pyridone ring. In this region, the stable paramagnetic σ-dimer radical trianions are formed, as confirmed for the first time by in situ spectroelectrochemical EPR and EPR/UV−vis studies.

■

ASSOCIATED CONTENT

S Supporting Information *

Potential dependences of UV−vis spectra, experimental and simulated EPR spectra, cyclic voltammograms, B3LYP bond lengths, and plots of frontier orbitals and electronic absorption spectra for both neutral and charged selenadiazoloquinolones. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: ++4212 5932 5537; Fax: ++4212 5932 5751; E-mail: [email protected] (P.R.). Tel: ++4212 5932 5666; Fax: + +4212 5932 5751; E-mail: [email protected] (V.B.) Notes

The authors declare no competing financial interest.

■

Figure 7. (a) Experimental (solid black line) and simulated (solid red line) EPR spectra of E1f•− (SW = 4 mT) radical anion. (b) Experimental (in red) and B3LYP (in green italic) hyperfine EPR coupling constants found for E1f•− radical anion along with (c) their linear correlation (R2 = 0.97).

ACKNOWLEDGMENTS This study was financially supported by the Scientific Grant Agency (Projects VEGA 1/0289/12, 1/0660/11, 1/1072/11) and the Research and Development Agency of the Slovak Republic (contract no. APVV-0339-10). We thank Stanislav Biskupič and Ján Šajbidor for their support and Philip Grier for helpful discussion.

the new optical band in the region 300−400 nm arising at the second voltammetric peak.

■

■

CONCLUSIONS On the basis of the EPR spectroelectrochemical and in situ EPR/UV−vis cyclovoltammetric experiments of the series of 7substituted 6-oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-h]quinolines and 8-substituted 9-oxo-6,9-dihydro[1,2,5]selenadiazolo[3,4-f ]quinolines, a reaction mechanism for their reduction was suggested. The substitution at the nitrogen atom of the 4-pyridone ring plays the decisive role here. The N-ethyl substituted derivatives reversibly form stable radical anions with a spin-density distribution on the N(1)−Se(2)−N(3) moiety. The cathodic reduction of imino derivatives leads to unstable

REFERENCES

(1) King, D.; Malone, R.; Lilley, S. Am. Fam. Physician 2000, 61, 2741−2748. (2) Oliphant, C.; Green, G. Am. Fam. Physician 2002, 65, 455−464. (3) Boteva, A.; Krasnykh, O. Chem. Heterocycl. Compd. 2009, 45, 757−785. (4) Koba, M.; Baczek, T.; Macur, K.; Bober, L.; Frackowiak, T.; Bucinski, A.; Rystok-Grabska, D.; Stasiak, J.; Koba, K. J. Mol. Model. 2010, 16, 327−335. (5) Drlica, K.; Hiasa, H.; Kerns, R.; Malik, M.; Mustaev, A.; Zhao, X. Curr. Top. Med. Chem. 2009, 9, 981−998. H



Article

(6) Darque, A.; Dumetre, A.; Hutter, S.; Casano, G.; Robin, M.; Pannecouque, C.; Azas, N. Bioorg. Med. Chem. Lett. 2009, 19, 5962− 5964. (7) Ramesh, E.; Manian, R.; Raghunathan, R.; Sainath, S.; Raghunathan, M. Bioorg. Med. Chem. 2009, 17, 660−666. (8) Majumdar, K.; Chattopadhyay, B.; Pal, A. Lett. Org. Chem. 2008, 5, 276−281. (9) Majumdar, K.; Mondal, S. Tetrahedron Lett. 2008, 49, 2418− 2420. (10) Milata, V.; Claramunt, R. M.; Elguero, J.; Zálupský, P. Targets Heterocycl. Syst. 2000, 4, 167−203. (11) Bella, M.; Schultz, M.; Milata, V.; Koňariková, K.; Breza, M. Tetrahedron 2010, 66, 8169−8174. (12) Bella, M.; Schultz, M.; Milata, V. ARKIVOC 2012, 2012, 242− 251. (13) Naik, H.; Ramesha, M.; Swetha, B.; Roopa, T. Phosphorus, Sulfur Silicon Relat. Elem. 2006, 181, 533−541. (14) Mlochowski, J.; Giurg, M. Aromacity in Heterocyclic Compounds; Springer: Berlin, 2009; pp 288−340. (15) Mlochowski, J.; Kloc, K.; Lisiak, R.; Potaczek, P.; Wojtowicz, H. ARKIVOC 2007, 2007, 14−46. (16) Barbieriková, Z.; Bella, M.; Kučerák, J.; Milata, V.; Jantová, S.; Dvoranová, D.; Veselá, M.; Staško, A.; Brezová, V. Photochem. Photobiol. 2011, 87, 32−44. (17) Staško, A.; Bella, M.; Rimarčík, J.; Barbieriková, Z.; Milata, V.; Lukeš, V.; Brezová, V. J. Phys. Org. Chem. 2012, 25, 643−648. (18) Gale, R. J. Spectroelectrochemistry: Theory and Practice; Plenum Press: New York, 1988. (19) Kaim, W.; Klein, A. Spectroelectrochemistry; RSC Publishing: Cambridge, U.K., 2008. (20) Dunsch, L. J. Solid State Electrochem. 2011, 15, 1631−1646. (21) Hammerich, O.; Parker, V. D. Acta Chem. Scand., Ser. B 1981, 35, 341−347. (22) Amatore, C.; Garreau, D.; Hammi, M.; Pinson, J.; Saveant, J. J. Electroanal. Chem. 1985, 184, 1−24. (23) (a) Mazine, V.; Heinze, J. J. Phys. Chem. A 2004, 108, 230−235. (b) Heinze, J.; Willmann, C.; Bäuerle, P. Angew. Chem., Int. Ed. 2001, 40, 2861−2864. (24) Macias-Ruvalcaba, N.; Telo, J.; Evans, D. J. Electroanal. Chem. 2007, 600, 294−302. (25) Gallardo, I.; Guirado, G.; Marquet, J.; Vilà, N. Angew. Chem., Int. Ed. 2007, 46, 1321−1325. (26) Macias-Ruvalcaba, N.; Felton, G.; Evans, D. J. Phys. Chem. C 2009, 113, 338−345. (27) Duling, D. R. J. Magn. Reson. 1994, 104, 105−110. (28) Frisch, M. J. et al. Gaussian 03, revision A.1; Gaussian, Inc.: Pittsburgh, PA, 2003. (29) Becke, A. J. Chem. Phys. 1993, 98, 5648−5652. (30) Tomasi, J.; Mennucci, B.; Cancès, E. J. Mol. Struct.: THEOCHEM 1999, 464, 211−226. (31) Binkley, J. S.; Pople, J. A.; Hehre, W. J. J. Am. Chem. Soc. 1980, 102, 939−947. (32) Furche, F.; Ahlrichs, R. J. Chem. Phys. 2002, 117, 7433−7447. (33) Sanchez, P.; Evans, D. J. Electroanal. Chem. 2011, 660, 91−96. (34) Macias-Ruvalcaba, N.; Evans, D. J. Electroanal. Chem. 2011, 660, 243−246.

I


Stable Radical Trianions from Reversibly Formed Sigma-Dimers of

Recommend Documents