Introduction

Quinones are biologically important class of compounds occurring in plants, living organism, and in inanimate world. Many derivatives containing a quinone fragment in their structure exhibit a significant biological activity [1, 2]. For instance, in the respiratory enzymes quinone derivatives constitute a part of the electron-transfer chain [3]. Moreover, they are used as defensive compounds by some leaf beetles and also exhibit antitermitic activity [4, 5].

A number of previous studies have shown that the redox properties of anthraquinones are directly connected to their acid–base properties [6]. It is well known that under protic condition, the reduction of quinones produces corresponding hydroquinones [7, 8].

$$ {\text{Q}} + 2 {\text{H}}^{ + } + 2\bar{e} = {\text{H}}_{ 2} {\text{Q}} $$
(1)

In an aqueous solution, the individual one-electron processes comprised in reaction (1) usually occur as a single two-electron reduction process. However, in aprotic media, the reduction of quinones takes place in two well-resolved steps: the first one corresponds to the formation of the Q−• radical anion, and at the more negative potential, the closed-shell dianion, Q2−, is formed (see eqs. 2 and 3):

$$ {\text{Q}} + \bar{e} = {\text{Q}}^{ - \bullet } $$
(2)
$$ {\text{Q}}^{ - \bullet } + \bar{e} = {\text{Q}}^{ 2- } $$
(3)

The occurrence of two well-resolved reduction waves on a cyclic voltammogram is dubbed the EE mechanism [9, 10]. However, in the literature reports on the electrochemical reduction of quinones in polar, aprotic solvents, the EE mechanism is very often claimed to be not operative [11].

While studying various methyl derivatives of aminoanthraquinones in a methanol solution, Peters and Sumner proved that the difference in their basicity amounts to as much as around four pK units [12]. Such a significant difference in basicity was attributed to the possibility of the formation of a hydrogen bond between the proton of the amino group and the oxygen of the quinone carbonyl function only in some of the studied derivatives [12, 13].

The protonation of 1-amino substituent changes electron density on the nitrogen atom, which leads to the reduction of the molar absorption coefficient of the band appearing at about 500 nm for the neutral anthraquinone [13, 14]. This phenomenon allows the dissociation constants of aminoanthraquinone derivatives to be determined spectrophotometrically.

In this article, the acid–base properties of 1-aminoanthraquinone (AQNH2), 1-(ethylamino)anthraquinone (AQNHEt), and 1-(diethylamino)anthraquinone (AQNEt2) (for structures see Fig. 1) in an acetonitrile solution were studied experimentally by electrochemical and UV–Vis spectroscopic methods.

Fig. 1
figure 1

The structures of studied anthraquinone derivatives: AQNH2, AQNHEt, and AQNEt2

Full size image

The measured pKa constants were compared with the quantum-chemically derived differences in the stability of the neutral and protonated forms in acetonitrile for all the studied systems. Finally, differences in the behaviors of AQNH2 and AQNHEt and that of AQNEt2 in cyclic voltammetry, performed in the acetonitrile solution, have been interpreted in terms of the reaction of the fully reduced Q2− form of the latter compound with traces of water. This supposition has been conformed in cyclic voltammetry experiments for acetonitrile solutions containing the controlled concentrations of water, and with the quantum chemical (QM) thermodynamic characteristics of the reaction of anthraquinone dianions with water.

Materials and methods

Materials

All commercially available reagents and solvents were purchased from Sigma Aldrich and used without further purification. Only acetonitrile (HPLC grade, >99.9 %) was dried with molecular sieves (4 Å) before usage. 1H and 13C NMR spectra were recorded on a Varian, Mercury-400 spectrometer 400 and 100 MHz, respectively, with TMS as an internal reference and CDCl3 as a solvent. IR spectra were recorded on a Bruker IFS66 spectrometer as KBr pellets. MALDI mass spectra were recorded on a Biflex III MALDI-TOF mass spectrometer. Elemental analyses were recorded on a Carlo-Erba CHNS–O EA1108 elemental analyzer. The purity of the obtained compounds was checked by Shimadzu HPLC system LC-20A Series using a Phenomenex Luna C8 column (100 Å, 3 μm, 150 × 4.60 mm). The solvent system contained 0.1 % TFA in water (solvent A) and 100 % acetonitrile (solvent B). A linear gradient was applied from 1 % A to 100 % B for 30 min at a flow rate of 1.0 mL/min. Detection of the peaks was achieved by a UV detector at 254 nm.

AQNH2 was purchased from Sigma Aldrich and used without further purification, while AQNHEt and AQNEt2 were synthesized as described below.

Synthesis

AQNHEt

Ethylamine (185 mg, 4.121 mmol) was added to a mixture of 1-chloroanthraquinone (500 mg, 2.060 mmol) and cesium carbonate (1.342 g, 4.121 mmol) in 200 mL of toluene (Scheme 1). The reaction mixture was stirred at 80 °C for 48 h under argon. The reaction mixture was cooled to room temperature (RT), and the solvent was removed under reduced pressure. The residue was dissolved in dichloromethane (200 mL) and washed with water (2 × 100 mL). The organic phase was dried over anhydrous magnesium sulfate, and the solvent was removed in vacuo. The crude product was purified by flash column chromatography using silica gel and mixture of dichloromethane and methanol (95:5) as eluent to give 360 mg of desired compound (69 %) as red solid. mp 165 °C; IR (KBr) 3424, 2925, 1677,1630, 1593, 1575, 1510,1313, 1269, 1150, 1071, 1009, 801, 735, 725, 703,646; 1H NMR (CDCl3, 400 MHz) δ 1.39–1.43 (t, 3H, NH–CH2–CH 3 , J = 7.2 Hz), 4.02–4.04 (p, 2H, NH–CH 2 –CH3, J 1 = 6.2 Hz, J 1 = 7.2 Hz, J 1 = 4.8 Hz, J 2 = 6.7 Hz, J 2 = 7.2 Hz, J 2 = 6.0 Hz, J 3 = 6.6 Hz, J 3 = 6.9 Hz, J 4 = 6.7),7.54–7.56 (d, 1H, H-2 Ar, J = 7.6 Hz), 7.58–7.61 (dd, 1H, H-4 Ar, J 1 = 1.2 Hz, J = 7.2 Hz), 7.68–7.72 (dt, 1H, H-3 Ar, J 1 = 1.2 Hz, J 2 = 7.2 Hz, J 2 = 7.8 Hz, J 3 = 7.5 Hz), 7.74–7.78 (dt, 1H, H-6 Ar, J 1 = 1.2 Hz, J 1 = 2.0 Hz, J 2 = 7.4 Hz), 7.78–7.82 (dt, 1H, H-7 Ar, J 1 = 1.2 Hz, J 1 = 2.0 Hz, J 2 = 7.6 Hz),8.23–8.25 (dd, 1H, H-5 Ar, J 1 = 1.2 Hz, J 1 = 1.6 Hz, J 2 = 7.4 Hz), 8.29–8.30 (dd, 1H, H-8 Ar, J 1 = 1.6 Hz, J 2 = 7.6 Hz),9.66 (s, 1H, NH–CH2–CH3); 13C NMR (CDCl3, 100 MHz) δ 14.5; 36.5 (CH2); 115.7, 118.03, 126.8, 126.9, 127.01, 127.8, 133.0, 133.8, 134.13, 134.7, 135.5, 135.5 (CH, Ar); 182.1, 183.5 (2C=O, Ar); Maldi–Tof m/z 252.2 [M+H]+ (MW = 251.280); Anal. (C16H13NO2): Calcd. C, 76.48; H, 5.21; N, 5.57; found C, 76.44; H, 5.22; N, 5.60; Purity (HPLC): 98.9 %. t R = 17.95 min.

Scheme 1
scheme 1

Synthesis of 1-(ethylamino)anthraquinone (AQNHEt)

Full size image

AQNEt2

Diethylamine (605 mg, 8.264 mmol) was added to a solution of 1-chloroanthraquinone (500 mg, 2.060 mmol) in 150 mL of toluene (Scheme 2). The reaction mixture was thoroughly degassed by passing a stream of argon. The reaction mixture was heated at 100 °C for 24 h. After cooling to RT, the solvent was evaporated at reduced pressure. The residue was dissolved in dichloromethane (200 mL), washed with water (2 × 100 mL), and dried over anhydrous magnesium sulfate. The organic phase was evaporated at reduced pressure. The crude product was purified by flash column chromatography using silica gel and dichloromethane as eluent to give 520 mg of product (90 %) as red solid. mp 98–100 °C; IR (KBr) 3434, 2967, 1662, 1643, 1577, 1452, 1426, 1316, 1253, 1157, 1051, 985, 896, 797, 736, 718, 663620;1H NMR (CDCl3, 400 MHz) δ 1.13–1.17 (t, 6H, NH–CH2–CH 3 , J 1 = 7.2 Hz, J 1 = 6.8 Hz, J 2 = 7.0 Hz), 3.35–3.41 (q, 4H, NH–CH 2 –CH3, J 1 = 7.2 Hz, J 1 = 6.8 Hz, J 2 = 7.2 Hz, J 2 = 7.0 Hz, J 3 = 7.1 Hz), 7.54–7.56 (dd, 1H, H-2 Ar, J 1 = 0.8 Hz, J = 8.4 Hz), 7.52–7.56 (t, 1H, H-3 Ar, J 1 = 7.8 Hz, J 1 = 8.2 Hz, J 2 = 8.0 Hz), 7.67–7.71 (dt, 1H, H-6 Ar, J 1 = 1.2 Hz, J 1 = 1.6 Hz, J 2 = 7.2 Hz, J 2 = 7.4 Hz, J 3 = 7.6 Hz), 7.73–7.77 (dt, 1H, H-7 Ar, J 1 = 1.2 Hz, J 1 = 2.0 Hz, J 1 = 1.6 Hz, J 2 = 7.4 Hz, J 2 = 7.6 Hz, J 2 = 7.5 Hz), 7.81–7.83 (dt, 1H, H-4 Ar, J 1 = 1.2 Hz, J 2 = 7.6 Hz), 8.20–8.23 (dd, 1H, H-5 Ar, J 1 = 1.2 Hz, J 2 = 7.6 Hz), 8.23–8.25 (dd, 1H, H-8 Ar, J 1 = 1.2 Hz, J 2 = 7.6 Hz),13C NMR (CDCl3, 100 MHz) δ 14.1; 48.3 (CH2); 116.8, 120.40, 126.9, 127.47, 127.82, 128.5, 133.2, 133.8, 134.0, 134.5, 135.3, 135.4 (CH, Ar); 182.6, 184.7 (2C=O, Ar); Maldi–Tof m/z 280.1 [M+H]+, (MW = 279.333); Anal. (C18H17NO2): Calcd. C, 77.40; H, 6.13; N, 5.01; found C, 77.42; H, 6.15; N, 5.01; Purity (HPLC): 99.8 %. t R = 13.48 min.

Scheme 2
scheme 2

Synthesis of 1-(diethylamino)anthraquinone (AQNEt2)

Full size image

UV–Vis spectrophotometry

All UV–Vis spectra were recorded in 290–650-nm range on a Perkin Elmer Lambda 650 UV–Vis double beam spectrophotometer with automatic stirrer. 1-cm quartz microcells were used. The concentrations of compounds used for all spectrophotometric measurements were around 1.7 × 10−4–2.2 × 10−4 M. All measurements were performed at 298 K in acetonitrile.

To obtain the values of dissociation constants of measured compounds (pK a), spectrophotometric titration was used: the compounds were dissolved in acetonitrile and titrated with methanesulfonic acid solution. For each point of known pH, the absorption spectrum was recorded.

pH measurements were made with a CercoLab system with a pH-meter using a combined glass electrode. The electrode was calibrated in the buffer system of 2,6-dinitrophenol/tetrabutylammonium 2,6-dinitrophenolate in acetonitrile. The resolution of the voltage measurement was <0.1 mV.

To define the number of equilibria present in the studied system, the A-diagrams, which show a relationship between absorbances at two different wavelengths, were analyzed. The theoretical model was fitted to the experimental data presented as a plot of absorbance versus pH.

All calculations were performed with the OriginLab software using the Henderson–Hasselbach equation, based on change in absorption as a function of pH of the solution [15].

$$ pK_{\text{a}} = {\text{pH}} - { \log }\frac{{[{\text{B}}]}}{{[{\text{BH}}]}} = {\text{pH}} - { \log }\frac{{{\text{A}}_{\lambda } - {\text{A}}_{{\lambda {\text{BH}}}} }}{{{\text{A}}_{{\lambda {\text{B}}}} - {\text{A}}_{\lambda } }} $$
(4)

where B and BH stand for the compound in its basic and acidic forms, respectively, while A λB and A λBH denote the absorbance of these forms.

For the evaluation of electrode parameters, the STOICHIO version of the CVEQUID software based on the nonlinear least-squares Gauss–Newton–Marquardt algorithm was used [16–18].

Cyclic voltammetry

The electrochemical measurements were carried out in a single-compartment, three-electrode cell. The potential was applied with an Autolab potentiostat/galvanostat PGSTAT30 (Eco Chemie B.V., The Netherlands) controlled with the General Purpose Electrochemical System (GPES 4.9) software. All potentials were measured against an Ag/AgCl reference electrode with the aqueous silver–silver chloride (0.1 M NaCl) solution. This reference electrode was separated from the measuring cell by a salt bridge. In all experiments, the solution was deaerated by passing argon. The working electrode was a 0.2-cm-diameter glassy carbon electrode, and a platinum wire served as an auxiliary electrode. The investigated solutions consisted of a 0.1 M tetrabutylammoniumperchlorate (TBAP) in acetonitrile as a base electrolyte. All cyclic voltammetry measurements were performed at RT (~ 20 °C).

Computational details

We applied the density functional theory method with Becke’s three-parameter hybrid functional (B3LYP) [19–21], the 6-31 ++G** [22, 23] basis set, and the Polarizable Continuum Model (PCM) [24–26] to the acetonitrile solution calculations. All the geometries were fully optimized without any geometric constraints, and the analysis of harmonic frequencies demonstrated that all these were geometrically stable (all force constants were positive). The Gibbs free energies of particular reactions (ΔGs) were electronic energy change (ΔEs) corrected for zero-point vibration terms, thermal contributions to energy, the pV term, and the entropy term. These terms were calculated in the rigid rotor–harmonic oscillator approximation for T = 298 K and p = 1 atm [27]. Relative free energy change ΔΔG was defined as the difference between ΔG of particular aminoanthraquinone derivative and the lowest ΔG for the same reaction type.

All calculations have been carried out with the GAUSSIAN09 [28] code. The images of the molecules were plotted using the GaussView package [29].

Results and discussion

UV–Vis Spectrophotometry

Two absorption bands, in the range of 290–650 nm, occur in the UV–Vis spectra of anthraquinone derivatives containing the amino substituent at position 1 [13, 14]. The band at around 320 nm is associated with the benzenoid character of the molecule [13, 14]. The second absorption maximum occurs around 500 nm, which is due to the presence of amine nitrogen atom at position 1 of anthraquinone (see Fig. 2).

Fig. 2
figure 2

The absorption spectra of anthraquinone derivatives AQNH2, AQNHEt, and AQNEt2 in acetonitrile

Full size image

In methanol, Peter and Sumner observed a bathochromic shift of benzenoid band with the increase of its intensity for the AQ derivatives with an electron-donating amino substituent [12]. A similar effect has been found in our experiments, where the replacement of the amino group with the ethylamino or diethylamino one causes the bathochromic shift of the benzenoid absorption band with the concomitant increase of its molar absorption coefficient. An increase in the intensity of the band is associated with an increase in the share of benzenoid electronic transitions at the expense of the quinonoid ones [13]. Likewise, the band in the visible region of the spectrum moves to longer wavelength region and decreases its intensity with the increasing electron-donating character of the amino substituent (see Fig. 1), which is evidenced by the values of absorption maximum and the logarithm of molar absorption coefficient, which vary from λ = 466 nm, logε = 3.79 for compound (AQNH2) to λ = 504 nm, logε = 3.64; and λ = 514 nm, logε = 3.58 for compounds AQNHEt and AQNEt2, respectively (see Table 1).

Table 1 Logarithm of molar absorption coefficient (logε), negative of logarithm of acid dissociation constant (pK a) with uncertainty, and experimental (∆∆G exp) and theoretical (∆∆G theor) relative Gibbs free energy changes for the reaction (5) for anthraquinone derivatives in acetonitrile
Full size table

The addition of the excess of methanesulfonic acid to the acetonitrile solution of the investigated compounds causes disappearance of the long wavelength band and small shifts of the short wavelength absorption (~ 300 nm) [(AQNH2) 20 nm, (AQNHEt) 15 nm, and (AQNEt2) 7 nm]. In cases of AQNH2 and AQNHEt, however, it was necessary to use an acid solution of 0.1 M to observe a complete disappearance of the long wavelength band, while for AQNEt2, the equimolar amount of acid was sufficient.

In order to determine the number of acid–base equilibria present in acetonitrile solution and to measure the acid dissociation constants of the investigated derivatives, pH-spectroscopic titrations were performed. The example of titration curve is presented for AQNHEt (see Fig. 3). The respective data for the remaining compounds are shown in Supporting Information (SI; Figures S1 and S3).

Fig. 3
figure 3

Spectrophotometric titration of AQNHEt (concentration of 1.79 × 10−4 M) in the pH range of 7.5–13.5 in acetonitrile. The arrow indicates the change in absorbance during titration

Full size image

The change of absorbance with pH shows one equivalence point (see Fig. 4a), and the A-diagram represents a straight line (see Fig. 4b).

Fig. 4
figure 4

Fitting of the Henderson–Hasselbach equation to the absorption changes (λ = 504 nm) with pH for AQNHEt (a), and the respective A-diagram (b)

Full size image

In addition, there are two isosbestic points (for AQNHEt at 323 and 377 nm, see Fig. 3). The analysis of the obtained spectroscopic data shows that in the acetonitrile solution of AQNHEt, only one acid–base equilibrium is present. Hence, the following reaction model has been used to obtain the dissociation constants:

$$ {\text{HB}}^{ + } \rightleftarrows {\text{B}} + {\text{H}}^{ + } $$
(5)
$$ K_{\text{a}} = \frac{{\left[ {\text{B}} \right][{\text{H}}^{ + } ]}}{{[{\text{HB}}^{ + } ]}} $$
(6)

where B stands for a compound in the base form and HB+ denotes its acidic form.

The pK a values for the AQNH2 and AQNHEt derivatives differ slightly, while that for AQNEt2 is much higher (see Table 1), indicating, thus, a higher basicity of AQNEt2 compared with the other two. These values vary from 8.01 for AQNH2 and 8.65 for AQNHEt to 14.93 in the case of AQNEt2. One can wonder why AQNEt2 is a much stronger base then monoethylamino- or unsubstituted aminoanthraquinones. Basic organic chemistry suggests that the increase in the number of alkyl substituents on amine nitrogen is associated with the increase of its basicity. Indeed, the pK a of AQNHEt is larger, by 0.7 pK unit, than that of AQNH2 (see Table 1). Introducing, the second ethyl substituent to the amine function, however, leads to a dramatic increase in the aminoanthraquinone basicity. In other words, the pK a of AQNEt2 is by ca 6 pK units larger than that of AQNHEt. The results of our QM calculations account well for the experimental findings. In Table 1, the relative differences between the free energies of the basic and acidic forms of the studied aminoanthraquinones are compared. The experimental and B3LYP data remain in a qualitative accordance, and quantitatively they differ by no more than 1.3–6.1 kcal/mol (see Table 1).

Moreover, in Fig. 5, the geometries of the neutral aminoanthraquinones and their protonated counterparts are displayed. These structures demonstrate why only small difference in basicity is observed for AQNH2 and AQNHEt, whereas the substituent effect is much more pronounced for AQNEt2. First, note that for both AQNH2 and AQNHEt, the addition of proton does not change the topology of the internal hydrogen bond present in the neutral molecule (cf. structures A and B in Fig. 5). On the other hand, the presence of two bulky ethyl substituents in AQNEt2 brings about rotation of the amine group by ca 40 degrees (see C in Fig. 5), which prevents the coupling of the lone pair on the amine nitrogen with the aromatic system of anthraquinone π-bonds that make the AQNEt2 lone pair better accessible to protonation than that in AQNH2 or AQNHEt. Moreover, due to the lack of amine hydrogens, there is no stabilizing hydrogen bond in the neutral AQNEt2 (see C in Fig. 5), while such a hydrogen bond is formed in the protonated species (see C in Fig. 5), which greatly stabilizes the latter form and significantly increases the basicity of the considered derivative (see Table 1).

Fig. 5
figure 5

Neutral and protonated forms of AQNH2 (a), AQNHEt (b), and AQNEt2 (c), calculated at the B3LYP/6-31 ++G(d, p) level in acetonitrile. Θ 1-2-3-4 denotes the dihedral angle defined by the atoms no 1, 2, 3, and 4

Full size image

Cyclic voltammetry

The examined anthraquinone derivatives are electrochemically reduced within a typical two-step reduction process as is usually observed for quinones [3]. For AQNH2 and AQNHEt, the same value of the first reduction peak was observed, E 1pa  = −1.088 V, while the second one shows slight variations as E 2pc  = −1.581 and −1.553 V, respectively (see Table 2).

Table 2 The values of the reduction potential peak (E pc), oxidation potentials peak (E pa), and the difference between the potentials of reduction and oxidation peaks of the studied electrode processes
Full size table

Again in case of AQNEt2, significant differences in electrochemical behaviors with respect to the remaining compounds are observed. E 1pc reduction peak is slightly shifted toward more negative potential values (E 1pc  = −1.134 V), and the second reduction peak E 2pc is clearly moved (about 0.1 V) in the positive potential direction (Table 2). The main differences arise in the anodic curve (Fig. 6).

Fig. 6
figure 6

Cyclic voltammograms of anthraquinone derivatives (concentration of 1.0 mM) in acetonitrile with 0.1 M Bu4NClO4, ν = 0.1 V/s. Arrow indicates initial potential and initial scan direction

Full size image

In the cyclic voltammograms of AQNEt2, the second anodic process almost vanishes. In the anodic part related to the first reduction process, E 1pc , a dual peak appears, with the E 1pa and E 3pa potentials (see Fig. 6; Table 2). Moreover, the anodic response does not significantly change regardless of whether the potential of the electrode reaches the value at which the formation of Q− or Q2− occurs.

In order to explain why the three relatively similar compounds differ in the mechanisms of reduction and oxidation processes in the acetonitrile solution, additional cyclic voltammetry experiments were performed. All the above described facts emphasize the importance of basicity. On the other hand, different acid–base characteristics should not be relevant for polar aprotic medium such as acetonitrile. However, one should remember that even a dry acetonitrile contains traces of water which can react with a substance of acidic or basic character [30]. Therefore, these additional cyclic voltammetry experiments were carried out in the presence of controlled concentration of water.

The addition of water to the acetonitrile solution of the investigated compounds resulted in a substantial change in cyclic voltammograms. An example of cyclic voltammograms for AQNHEt in the presence of increasing concentrations of water is shown in Fig. 7.

Fig. 7
figure 7

Cyclic voltammograms of 0.95 mM AQNHEt in the presence of increasing concentrations of water (0.0–0.126 M) in acetonitrile with 0.1 M Bu4NClO4, ν = 0.1 V/s. Arrow indicates initial potential and initial scan direction

Full size image

With the increase of water concentration, the reduction and oxidation potentials for the first process shifted toward more positive values. However, the reversibility of the process expressed as ΔE = E 1pc −E 1pa and the cathode and anode currents height, i 1pc , i 1pa , did not change with water content. The second step reduction/oxidation potential (E 2pc , E 2pa ) moves stronger to positive potentials than the first one.

A completely different situation occurs in case of AQNEt2. Increasing the concentration of water results in the disappearance of reduction peak E 2pc (see Fig. 8).

Fig. 8
figure 8

Cyclic voltammograms of 1.01 mM AQNEt2 in the presence of increasing concentrations of water (0.0–0.087 M) in acetonitrile with 0.1 M Bu4NClO4, ν = 0.1 V/s. Arrow indicates initial potential and initial scan direction

Full size image

A gradual increase of water concentration causes a strong increase in the reduction peak E 1pc current and shifts it toward more positive potential values. Moreover, the presence of water brings about a gradual increase of the peak E 3pa and total disappearance of the peak E 1pa in the anodic part of the voltammogram curve. Finally, under high concentration of water (over 18 mM), a significant change in the reduction process occurs, which suggests the product of the reduction undergoes a further chemical transformation.

The differences observed for the electroreduction of the investigated quinones might be related to the difference in the basicity of their reduced form. Indeed, the electrochemical experiments with the controlled amount of water indicate clearly that the usually observed EE mechanism (eq. 2 and 3) is modified, when the content of water is increased. This effect is conspicuously visible for AQNEt2 and suggests that the EE mechanism is converted to the ECE one. The dianion Q2− formed in the second reduction process is probably capable of reacting even with traces of water (eq. 7). The hydroxyl ion formed in the latter reaction can subsequently interact with the quinone molecule (eq. 8) resulting in the third oxidation peak E 3pa .

$$ {\text{Q}}^{2 - } {\text{ + H}}_{ 2} {\text{O}} \rightleftarrows {\text{HQ}}^{ - } {\text{ + OH}}^{ - } $$
(7)
$$ {\text{Q + OH}}^{ - } \to {\text{C}^{ - }} $$
(8)
$$ 2 {\text{Q}} + 2\bar{e} + {\text{H}}_{ 2} {\text{O}} \to {\text{HQ}}^{ - } {\text{ + C}^{ - }} $$
(9)

Interestingly, the third oxidation peak for AQNEt2 is registered even before its dianion is electrochemically generated. In order to observe E 3pa , it is sufficient to polarize the working electrode to a potential characteristic for the first reduction process. The explanation of this phenomenon can be a disproportionation reaction of anion radical created at the E 1pc potential (eq. 10).

$$ 2 {\text{Q}}^{ - \bullet } \to {\text{Q}}^{ 2- } + {\text{Q}} $$
(10)

Q2− will then react according to the scheme comprising reactions (7–9), forming ultimately a product C that gives rise to the E 3pa signal during the oxidation phase of cyclic voltammetry. Hence, the formation of the product C without producing dianion electrochemically can be considered as an indirect proof of the postulated ECE mechanism.

The results of our B3LYP calculations fully support the conclusions formulated in the previous paragraph. Table 3 gathers thermodynamic characteristics for steps 7–9. In ∆G calculations, entropy was limited to its vibrational contribution since in the condensed phase, translations and rotations are almost completely hindered. Although one can note only tiny differences between the thermodynamic stimuli for the total process (step 9) involving particular aminoanthraquinones, the characteristics for steps 7 and 8 are pretty much in line with the experimental observations. Indeed, the energy and free energy of proton transfer (PT) between water and an aminoanthraquinone dianion are negative only for AQNEt2, while they are substantially positive for the remaining derivatives. Thus, only the AQNEt2 dianion is able to react spontaneously with water. In a subsequent step, the hydroxyl anions may react with the neutral AQNEt2 molecules forming a gemdiol type species (see Fig. 9).

Table 3 Thermodynamic characteristics of ECE mechanism in the presence of water, calculated at the B3LYP/6-31 ++G(d, p) level, in an acetonitrile solution
Full size table
Fig. 9
figure 9

Gemdiol-type products originated from AQNH2 (a), AQNHEt (b), and AQNEt2 (c)

Full size image

In fact, the electrochemical reduction of quinones has already been a subject of Lehman and Evans studies [10] that suggested the Q2− dianion formed in the second step (eq. 3) may react with even a trace amount of water present in acetonitrile, leading to the hydroxyl ion, OH− (eq. 7). According to those researchers, the reaction product (C) (eq. 8) is a gemdiol which results from the hydroxyl anion addition to one of the carbonyl carbons of benzoquinone (In Fig. 9, the most stable product of OH− addition to the carbonyl group of the studied aminoanthraquinones are depicted). As demonstrated by the energetic characteristics gathered in Table 3, the thermodynamic stimulus for the formation of a gemdiol product is indeed negative, which makes this reaction probable as far as thermodynamics is concerned.

According to the proposed model, the product of the electrochemical reaction AQNEt2 transformed by a chemical reaction (C; eq. 8) is then further oxidized at the E 3ac potential. Hence, the proposed mechanism can be defined as the ECE one. In cases of the AQNH2 and AQNHEt compounds, the basicity values of their neutral (Table 1) and reduced forms are substantially lower than that of the AQNEt2 derivative. Indeed, the second reduction process is quasi-reversible, which suggests that there is no chemical reaction of dianion with water in solutions containing AQNH2 or AQNHEt. Interactions between AQNH2/AQNHEt and the water molecules lead to the formation of molecular complexes (associates) only and the electrochemical processes proceed for these two systems according to the EE mechanism. These water complexes probably facilitate the reduction of anion radical to dianion, which results in shifting the second reduction peak, E 2pc , toward the first one (see Fig. 7).

Conclusions

Spectroscopic, electrochemical, and quantum-chemical studies were carried out to determine basicity and electrochemical behavior of 1-aminoanthraquinone and its ethyl derivatives in acetonitrile solutions. We found out that the basicity of the studied 1-aminoanthraquinones increases in the following order: AQNH2 ≈ AQNHEt < AQNEt2. The difference in basicity values between AQNH2/AQNHEt and AQNEt2 amounts to as much as ca 6 pK units. Such significantly larger pKa for AQNEt2 results from the presence of two ethyl group at the 1-amino substituent, which prevents the formation of an intramolecular hydrogen bond involving the carbonyl group of anthraquinone moiety in the neutral form of the derivative, while making it possible for the protonated species.

A consequence of the increased basicity of AQNEt2 is the increased basicity of its dianion which manifests itself in altering the mechanism of electrochemical processes. Indeed, the AQNHEt derivative, even in the presence of large excess of water, does not produce a reaction product, oxidized during cyclic voltammetry, since the basicity of its dianion is too small to react with water. On the other hand, the dianion of aminoanthraquinone containing tertiary amine group (AQNEt2) is a relatively strong base which, reacting with water, produces the hydroxyl radical that in a subsequent chemical process reacts with the neutral AQNEt2 giving a gemdiol-type derivative. The latter species gives rise to an additional oxidation peak on the cyclic voltammogram of AQNEt2.

Supporting Information

Spectrophotometric titrations, fitting curves and A-diagrams for AQNH2 and AQNEt2