Push-Pull Chromophores Based on the Naphthalene Scaffold: Potential Candidates for Optoelectronic Applications

A series of ten push-pull chromophores comprising 1H-cyclopenta[b]naphthalene-1,3(2H)-dione as the electron-withdrawing group have been designed, synthesized, and characterized by UV-visible absorption and fluorescence spectroscopy, cyclic voltammetry and theoretical calculations. The solvatochromic behavior of the different dyes has been examined in 23 solvents and a positive solvatochromism has been found for all dyes using the Kamlet-Taft solvatochromic relationship, demonstrating the polar form to be stabilized in polar solvents. To establish the interest of this polyaromatic electron acceptor only synthesizable in a multistep procedure, a comparison with the analog series based on the benchmark indane-1,3-dione (1H-indene-1,3(2H)-dione) has been done. A significant red-shift of the intramolecular charge transfer band has been found for all dyes, at a comparable electron-donating group. Parallel to the examination of the photophysical properties of the different chromophores, a major improvement of the synthetic procedure giving access to 1H-cyclopenta[b]naphthalene-1,3(2H)-dione has been achieved.


Introduction
During the past decades, push-pull molecules have attracted much attention due to their numerous applications ranging from non-linear optical (NLO) applications [1,2] to organic photovoltaics (OPVs) [3,4], organic field effect transistors (OFETs) [5], organic light-emitting diodes (OLEDs) [6], photorefractive applications [7], colorimetric pH sensors [8], ions detection [9], biosensors [10], gas sensors [11], or photoinitiators of polymerization [12][13][14][15][16]. Typically, push-pull chromophores are based on an electron donor and an electron acceptor connected to each other by mean of a saturated or a conjugated spacer [17]. As the first manifestation of the mutual interaction between the two partners, a broad absorption band corresponding to the intramolecular charge transfer (ICT) interaction can be detected in the visible to the near-infrared region. Position of this absorption band typically depends on the electron-donating ability of the donor and the electron-releasing ability of the acceptor, and a bathochromic shift of this transition is observed upon improvement of the strength of the donor and/or the acceptor. This ICT band can even be detected in the near and far infrared region for electron-acceptors based on poly(nitrofluorenes) [18]. Among electron acceptors, indane-1,3-dione (1H-indene-1,3(2H)-dione) EA1 has been extensively studied due to its commercial of one of the simplest reaction of Organic Chemistry, namely the Knoevenagel reaction [19][20][21][22][23]. By chemical engineering, its electron-withdrawing ability can be drastically improved by condensation of one or two malononitrile units under basic conditions, furnishing 2-(3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile EA2 [24] and 2,2'-(1H-indene-1,3 (2H)-diylidene)dimalononitrile EA3 [25] (see Figure 1). Based on the strong electron-withdrawing abilities of EA2 and EA3, push-pull dyes with non-fullerene acceptors and small energy gaps have been developed [26,27]. As a possible alternative to improve the electron-accepting ability, extension of the aromaticity of the acceptor can be envisioned. Using this strategy, a red-shift of the ICT band combined with an enhancement of the molar extinction coefficient can be both obtained for these new push-pull derivatives, relative to that obtained with the parent electron acceptor. In this article, an extended version of the well-known indane-1,3-dione, i.e., 1Hcyclopenta[b]naphthalene-1,3(2H)-dione EA4, has been used for the design of ten push-pull dyes. It has to be noticed that even if the synthesis of EA4 is reported in the literature since 2006 [28], only one report mentions the development of anion sensors with EA4 in 2013 [29] and the first push-pull dyes have been developed in 2017 for photovoltaics applications [30,31]. Considering the scarcity of studies devoted to this electron acceptor, a series of 10 push-pull chromophores PP1-PP10 have been developed with EA4. To evidence the contribution of the additional aromatic ring in EA4 relative to that of EA1, 10 dyes PP11-PP20 comprising EA1 as the electron acceptor have been synthesized for comparison (see Figure 2). In this article, an extended version of the well-known indane-1,3-dione, i.e., 1H-cyclopenta[b] naphthalene-1,3(2H)-dione EA4, has been used for the design of ten push-pull dyes. It has to be noticed that even if the synthesis of EA4 is reported in the literature since 2006 [28], only one report mentions the development of anion sensors with EA4 in 2013 [29] and the first push-pull dyes have been developed in 2017 for photovoltaics applications [30,31]. Considering the scarcity of studies devoted to this electron acceptor, a series of 10 push-pull chromophores PP1-PP10 have been developed with EA4. To evidence the contribution of the additional aromatic ring in EA4 relative to that of EA1, 10 dyes PP11-PP20 comprising EA1 as the electron acceptor have been synthesized for comparison (see Figure 2). of one of the simplest reaction of Organic Chemistry, namely the Knoevenagel reaction [19][20][21][22][23]. By chemical engineering, its electron-withdrawing ability can be drastically improved by condensation of one or two malononitrile units under basic conditions, furnishing 2-(3-oxo-2,3-dihydro-1H-inden-1-ylidene)malononitrile EA2 [24] and 2,2'-(1H-indene-1,3 (2H)-diylidene)dimalononitrile EA3 [25] (see Figure 1). Based on the strong electron-withdrawing abilities of EA2 and EA3, push-pull dyes with non-fullerene acceptors and small energy gaps have been developed [26,27]. As a possible alternative to improve the electron-accepting ability, extension of the aromaticity of the acceptor can be envisioned. Using this strategy, a red-shift of the ICT band combined with an enhancement of the molar extinction coefficient can be both obtained for these new push-pull derivatives, relative to that obtained with the parent electron acceptor. In this article, an extended version of the well-known indane-1,3-dione, i.e., 1Hcyclopenta[b]naphthalene-1,3(2H)-dione EA4, has been used for the design of ten push-pull dyes. It has to be noticed that even if the synthesis of EA4 is reported in the literature since 2006 [28], only one report mentions the development of anion sensors with EA4 in 2013 [29] and the first push-pull dyes have been developed in 2017 for photovoltaics applications [30,31]. Considering the scarcity of studies devoted to this electron acceptor, a series of 10 push-pull chromophores PP1-PP10 have been developed with EA4. To evidence the contribution of the additional aromatic ring in EA4 relative to that of EA1, 10 dyes PP11-PP20 comprising EA1 as the electron acceptor have been synthesized for comparison (see Figure 2). The photophysical properties of the series of 20 dyes as well as their electrochemical properties have been investigated. To support the experimental results, theoretical calculations have been carried out. Finally, the solvatochromic properties have also been examined.

Materials and Methods
All reagents and solvents were purchased from Aldrich, Alfa Aesar or TCI Europe and used as received without further purification. Mass spectroscopy was performed by the Spectropole of Aix-Marseille University. Electron spray ionization (ESI) mass spectral analyses were recorded with a 3200 QTRAP (Applied Biosystems SCIEX) mass spectrometer. The HRMS mass spectral analysis was performed with a QStar Elite (Applied Biosystems SCIEX) mass spectrometer. Elemental analyses were recorded with a Thermo Finnigan EA 1112 elemental analysis apparatus driven by the Eager 300 software. 1 H and 13 C nuclear magnetic resonance (NMR) spectra were determined at room temperature in 5 mm outer diameter (o.d.) tubes on a Bruker Avance 400 spectrometer of the Spectropole: 1 H (400 MHz) and 13 C (100 MHz). The 1 H chemical shifts were referenced to the solvent peak CDCl 3 (7.26 ppm) and the 13 C chemical shifts were referenced to the solvent peak CDCl 3 (77 ppm). UV-visible absorption spectra were recorded on a Varian Cary 50 Scan UV Visible Spectrophotometer, with concentration of 5 × 10 −3 M, corresponding to diluted solutions. Fluorescence spectra were recorded using a Jasco FP 6200 spectrometer. The electrochemical properties of the investigated compounds were measured in acetonitrile by cyclic voltammetry, scan rate 100 mV·s −1 , with tetrabutylammonium tetrafluoroborate (0.1 M) as a supporting electrolyte in a standard one-compartment, three-electrode electrochemical cell under an argon stream using a VSP BioLogic potentiostat. The working, pseudo-reference and counter electrodes were platinum disk (Ø = 1 mm), Ag wire, and Au wire gauze, respectively. Ferrocene was used as an internal standard, and the potentials are referred to the reversible formal potential of this compound. Computational details: All quantum mechanical calculations were computed using Gaussian Package [32]. All geometry optimizations were performed using the density functional theory (DFT) with the global hybrid exchange-correlation functional B3LYP [33] and all minima on the potential energy surface were verified via a calculation of vibrational frequencies, ensuring no imaginary frequencies were present. The Pople double-zeta basis set with a double set of polarization functions on non-hydrogen atoms (6-3111G(d,p)) [34,35] was used throughout. This computational approach was chosen in consistency with previous works, as it provides good agreement with experimental data. Excited states were probed using the time dependent density functional theory (TD-DFT) using the same function. All transitions (singlet-singlet) were calculated vertically with respect to the singlet ground state geometry. Solvent effects were taken into account by using the implicit polarizable continuum model (PCM) [36,37]. Dichloromethane (DCM) where chosen in analogy with the experiments. Computed spectra were simulated by convoluting each transition with Gaussians functions-centered on each absorption maximum using a constant full width at half maximum (FWHM) value of 0.2 eV. The assignment of electronic transitions for λ max has been determined with GaussSum 3.0 software [38]. 4-(Dodecyloxy)benzaldehyde D1 [39], 3,4-dibutoxy-benzaldehyde D2 [40], 2,4-dibutoxy-benzaldehyde D3 [41], 4-(diphenylamino)benzaldehyde D6 [42], 4-(bis(4-bromophenyl)-amino)benzaldehyde D7 [43], and 9-methyl-9H-carbazole-3-carbaldehyde D8 [44], 3-(4-(dimethylamino)phenyl)acrylaldehyde D9 [45], and 3,3-bis(4-(dimethylamino)phenyl)-acrylaldehyde) D10 [46] were synthesized, as previously reported, without modifications and in similar yields.

General Procedure for the Synthesis of PP1-PP10:
1H-Cyclopenta[b]naphthalene-1,3(2H)-dione EA4 (0.5 g, 2.55 mmol) and the appropriate substituted benzaldehyde (2.55 mmol, 1. eq.) were dissolved in absolute ethanol (50 mL) and a few drops of piperidine were added. The reaction mixture was refluxed and progress of the reaction was followed by thin layer chromatography (TLC). After cooling, a precipitate formed. It was filtered off, washed several times with ethanol and dried under vacuum.    13

General Procedure for the Synthesis of PP11-PP20
Indane-1,3-dione (1H-indene-1,3(2H)-dione) EA1 (0.37 g, 2.55 mmol) and the appropriate substituted benzaldehyde (2.55 mmol, 1 eq.) were dissolved in absolute ethanol (50 mL) and a few drops of piperidine were added. The reaction mixture was refluxed and progress of the reaction was followed by TLC. After cooling, a precipitate formed in most of the case. This latter was filtered off, washed several times with ethanol and dried under vacuum. For several chromophores (PP11, PP12, PP13 or PP20), a purification by column chromatography on SiO 2 was required.

Synthesis of the Dyes and Electron Acceptors
All dyes PP1-PP20 presented in this work have been synthesized by a Knoevenagel reaction involving the aldehydes D1-D10 and the two electron acceptors EA4 and EA1, respectively (See Figure 3). The different reactions were performed in ethanol using piperidine as the catalyst. PP1-PP20 were obtained with reaction yields ranging from 74% yield for PP4 and PP11 to 94% for PP5 (see Table 1). All compounds were obtained as solids and they were characterized by 1 H, 13 C nmR spectroscopies, and HRMS spectrometry (see Supplementary Materials). It has to be noticed that the synthetic procedure to EA4 has been greatly improved compared to that reported in the literature [47], enabling to reach a reaction yield of 91%. Indeed, a common method to synthesize 1,3-indanedione derivatives consists in a Claisen condensation of the corresponding diesters with ethyl acetate under basic conditions (generally sodium hydride).

Synthesis of the Dyes and Electron Acceptors
All dyes PP1-PP20 presented in this work have been synthesized by a Knoevenagel reaction involving the aldehydes D1-D10 and the two electron acceptors EA4 and EA1, respectively (See Figure 3). The different reactions were performed in ethanol using piperidine as the catalyst. PP1-PP20 were obtained with reaction yields ranging from 74% yield for PP4 and PP11 to 94% for PP5 (see Table 1). All compounds were obtained as solids and they were characterized by 1 H, 13 C NMR spectroscopies, and HRMS spectrometry (see Supplementary Materials). It has to be noticed that the synthetic procedure to EA4 has been greatly improved compared to that reported in the literature [47], enabling to reach a reaction yield of 91%. Indeed, a common method to synthesize 1,3indanedione derivatives consists in a Claisen condensation of the corresponding diesters with ethyl acetate under basic conditions (generally sodium hydride).  The final product is obtained after the condensation step by decarboxylation of the intermediate salt under hot acidic conditions (see Figure 4). To access the starting compound, i.e., diethyl naphthalene-2,3-dicarboxylate, two different reaction pathways were examined, by esterification of naphthalene-2,3-dicarboxylic acid in ethanol in the presence of an excess of thionyl chloride, or by the classical esterification conditions consisting in refluxing the acid in ethanol in the presence of a  The final product is obtained after the condensation step by decarboxylation of the intermediate salt under hot acidic conditions (see Figure 4). To access the starting compound, i.e., diethyl naphthalene-2,3-dicarboxylate, two different reaction pathways were examined, by esterification of naphthalene-2,3-dicarboxylic acid in ethanol in the presence of an excess of thionyl chloride, or by the classical esterification conditions consisting in refluxing the acid in ethanol in the presence of a catalytic amount of H 2 SO 4 (see Figure 4). If diethyl naphthalene-2,3-dicarboxylate could be obtained in almost quantitative yields with the two procedures, our attempt to convert the diester obtained by the first procedure were unfruitful to form EA4. This is attributable to remaining traces of SOCl 2 in the diester, despites the numerous washings in basic conditions. Due to the presence of water traces in ethyl acetate, SOCl 2 could react with water to form HCl, neutralizing part of the NaH introduced. Conversely, in the second procedure, H 2 SO 4 can be easily removed from the diester due to its catalytic use, avoiding this drawback. Using the oil obtained by this second procedure, a major improvement was obtained by replacing NaH 60% dispersion in oil by NaH 95% dispersion in oil for the first reaction step.
Materials 2018, 11, x FOR PEER REVIEW 9 of 24 catalytic amount of H2SO4 (see Figure 4). If diethyl naphthalene-2,3-dicarboxylate could be obtained in almost quantitative yields with the two procedures, our attempt to convert the diester obtained by the first procedure were unfruitful to form EA4. This is attributable to remaining traces of SOCl2 in the diester, despites the numerous washings in basic conditions. Due to the presence of water traces in ethyl acetate, SOCl2 could react with water to form HCl, neutralizing part of the NaH introduced. Conversely, in the second procedure, H2SO4 can be easily removed from the diester due to its catalytic use, avoiding this drawback. Using the oil obtained by this second procedure, a major improvement was obtained by replacing NaH 60% dispersion in oil by NaH 95% dispersion in oil for the first reaction step. By using this more concentrated dispersion, the reaction yield of the crude materials for the two steps (the intermediate salt was not isolated) could be greatly increased. By modifying the decarboxylation time (increased from 1.5 h instead of 20 min), the quantity of NaH (2.5 eq. instead of 1.45 eq.) and the recrystallization solvent [47] (benzene replaced by the less toxic toluene) compared to that reported in the literature, the overall reaction yield after purification could be increased from 65% (literature) up to 91% in our optimized conditions.

Theoretical Calculations
Theoretical studies were realized in order to investigate the energy levels as well as the molecular orbitals (M.O.) compositions of the different dyes. DFT calculations of all synthetized compounds were performed by the B3LYP/6-311G(d,p) level of theory using Gaussian 09 programs. Dichloromethane as the solvent and the polarizable continuum model (PCM) as the solvent model were used for the TD-DFT calculations. The optimized geometries as well as the highest occupied molecular orbitals (HOMO) and the lowest unoccupied molecular orbital (LUMO), i.e., the frontier orbitals' electronic distributions of selected compounds (PP6 and PP16) are given in Figure 5. The data for all compounds are supplied in the Supplementary Materials. The simulated absorption spectra are shown in the Figure 6 while the Table 2 summarizes all HOMO and LUMO energy levels and electronic transitions associated with the intramolecular charge transfer (ITC) absorption peaks. The assignment of the electronic transitions for λmax reported in the Table 2 has been determined with GaussSum 3.0 software, and especially, the contribution of the different transitions. By using this more concentrated dispersion, the reaction yield of the crude materials for the two steps (the intermediate salt was not isolated) could be greatly increased. By modifying the decarboxylation time (increased from 1.5 h instead of 20 min), the quantity of NaH (2.5 eq. instead of 1.45 eq.) and the recrystallization solvent [47] (benzene replaced by the less toxic toluene) compared to that reported in the literature, the overall reaction yield after purification could be increased from 65% (literature) up to 91% in our optimized conditions.

Theoretical Calculations
Theoretical studies were realized in order to investigate the energy levels as well as the molecular orbitals (M.O.) compositions of the different dyes. DFT calculations of all synthetized compounds were performed by the B3LYP/6-311G(d,p) level of theory using Gaussian 09 programs. Dichloromethane as the solvent and the polarizable continuum model (PCM) as the solvent model were used for the TD-DFT calculations. The optimized geometries as well as the highest occupied molecular orbitals (HOMO) and the lowest unoccupied molecular orbital (LUMO), i.e., the frontier orbitals' electronic distributions of selected compounds (PP6 and PP16) are given in Figure 5. The data for all compounds are supplied in the Supplementary Materials. The simulated absorption spectra are shown in the Figure 6 while the Table 2 summarizes all HOMO and LUMO energy levels and electronic transitions associated with the intramolecular charge transfer (ITC) absorption peaks. The assignment of the electronic transitions for λ max reported in the Table 2 has been determined with GaussSum 3.0 software, and especially, the contribution of the different transitions. As can be observed for all compounds, clear common trends are well observed for the two series: • ICT Absorption: The absorption peaks corresponded to the ICT are well observed in the Figure 6 and the values of maximal ICT absorption wavelengths are given in the Table 2. A clear trend is observed in both series: The ICT transition is mainly associated with the HOMO-LUMO transition. When the electron donor properties of the donor part increase, a clear redshift is obtained for the ICT transition. When we compare the two molecules in the two series with the same donor part, we observe that the ICT of the EA4-based molecule is redshift compared to that of EA1-based counterpart (for instance: λmax (nm) = 518 and 490 for PP6 and PP16, respectively, both compounds have triphenylamine as donor part).
• HOMO and LUMO Energy Levels: Within the same series, the values of EHOMO importantly vary as a function of the electron donating capacity of the donor part while the values of ELUMO slightly change. When we compare cross the two series: two molecules with the same donor part, we observe that they have nearly the same EHOMO while the ELUMO of the EA4-based molecule is deeper compared to that of EA1-based counterpart, which is in good accordance with the more electron deficient nature of EA4 (for instance: EHOMO (eV) = −5.633 and −5.635 for PP6 and PP16, respectively, while ELUMO (eV) = −2.836 and −2.694, respectively. Both compounds have triphenylamine as donor part).
• HOMO and LUMO Orbitals Distribution: The HOMO orbitals are well developed over the electron donor part while the LUMO orbitals are on the electron deficient moiety. Only a small overlap between the HOMO and LUMO orbitals is detected (see Figure 5). As can be observed for all compounds, clear common trends are well observed for the two series: ICT Absorption: The absorption peaks corresponded to the ICT are well observed in the Figure 6 and the values of maximal ICT absorption wavelengths are given in the Table 2. A clear trend is observed in both series: The ICT transition is mainly associated with the HOMO-LUMO transition. When the electron donor properties of the donor part increase, a clear redshift is obtained for the ICT transition. When we compare the two molecules in the two series with the same donor part, we observe that the ICT of the EA4-based molecule is redshift compared to that of EA1-based counterpart (for instance: λ max (nm) = 518 and 490 for PP6 and PP16, respectively, both compounds have triphenylamine as donor part).

HOMO and LUMO Energy Levels:
Within the same series, the values of E HOMO importantly vary as a function of the electron donating capacity of the donor part while the values of E LUMO slightly change. When we compare cross the two series: two molecules with the same donor part, we observe that they have nearly the same E HOMO while the E LUMO of the EA4-based molecule is deeper compared to that of EA1-based counterpart, which is in good accordance with the more electron deficient nature of EA4 (for instance: E HOMO (eV) = −5.633 and −5.635 for PP6 and PP16, respectively, while E LUMO (eV) = −2.836 and −2.694, respectively. Both compounds have triphenylamine as donor part).

HOMO and LUMO Orbitals Distribution:
The HOMO orbitals are well developed over the electron donor part while the LUMO orbitals are on the electron deficient moiety. Only a small overlap between the HOMO and LUMO orbitals is detected (see Figure 5).

Optical Properties
All compounds were characterized by UV-visible spectroscopy and their absorption spectra in dichloromethane are provided in the Figure 7. PP1-PP20 are push-pull compounds that possess an electron donating group and an electron accepting group which interact by mean of a π-conjugated system. The position of the charge transfer band will depend of parameters such as the strength of the electron donating/accepting groups, but also of the length of the π-conjugated system. By combining both effects, a decrease in the energy difference between the HOMO and the LUMO can be obtained, resulting in a shift of the absorption spectrum towards longer wavelengths.
While examining the first series PP1-PP10, absorption maxima ranging from 412 nm for PP1 to 524 nm for PP5 were found in dichloromethane for dyes exhibiting a short spacer. Elongation of the π-conjugated system resulted in a significant red-shift of the absorption maximum, shifting from 471 nm for PP1 to 571 nm for PP9. The most red-shifted absorption was found for PP10, peaking at 603 nm. While comparing the PP1-PP10 series with PP11-PP20 based on 1,3-indanedione, absorption spectra of PP11-PP20 were found to follow the same order than that of the PP1-PP10 series, but with an absorption blue-shifted by about 30 nm (see Figure 7). Examination of the molar extinction coefficients for the two series also revealed the PP1-PP10 series to exhibit higher molar extinction coefficients than that of the PP11-PP20 series, consistent with an improvement of the molar absorptivity with the oscillator strength and the conjugation extension (see Figure 8) [48].
The experimental absorption spectra recorded in dichloromethane of all compounds are in good accordance with predicted properties obtained by DFT calculation. Notably, a good accordance between the theoretical absorption maxima can be found for all dyes (See Tables 2-4). Second, considering that the main absorption band was theoretically determined for all dyes originating from a HOMO->LUMO transition, this latter can thus be confidently assigned to the intramolecular charge transfer bands for all dyes. A contribution of the HOMO->LUMO transition to the ICT band ranging between 85% (for PP20) to 100% for PP9 and PP19 could be calculated. The experimental absorption spectra recorded in dichloromethane of all compounds are in good accordance with predicted properties obtained by DFT calculation. Notably, a good accordance between the theoretical absorption maxima can be found for all dyes (See Tables 2-4). Second, considering that the main absorption band was theoretically determined for all dyes originating from a HOMO->LUMO transition, this latter can thus be confidently assigned to the intramolecular charge transfer bands for all dyes. A contribution of the HOMO->LUMO transition to the ICT band ranging between 85% (for PP20) to 100% for PP9 and PP19 could be calculated.

Solvatochromism
All dyes PP1-PP20 exhibited a good solubility in most of the common organic solvents so that examination of the solvatochromism could be carried out in 23 solvents of different polarities. It has to be noticed that alcohols such as methanol, ethanol, propan-2-ol, butan-1-ol and pentan-1-ol were initially considered as solvents for the solvatochromic study, but the absorption maxima obtained with these solvents were irregular compared to that obtained with the 23 other solvents. This specific behavior can be assigned to the fact that all dyes precipitated in alcohols such as in ethanol which was the solvent of reaction. Even if the absorption spectra could be recorded in alcohols for all dyes, presence of free molecules and aggregates in solution certainly modify the position of the absorption maxima. A summary of the absorption maxima for the twenty dyes are provided in Tables 3 and 4. Table 3. Summary of the optical properties of PP1-PP10 in twenty-three solvents and Kamlet and Taft parameters π*.

Solvatochromism
All dyes PP1-PP20 exhibited a good solubility in most of the common organic solvents so that examination of the solvatochromism could be carried out in 23 solvents of different polarities. It has to be noticed that alcohols such as methanol, ethanol, propan-2-ol, butan-1-ol and pentan-1-ol were initially considered as solvents for the solvatochromic study, but the absorption maxima obtained with these solvents were irregular compared to that obtained with the 23 other solvents. This specific behavior can be assigned to the fact that all dyes precipitated in alcohols such as in ethanol which was the solvent of reaction. Even if the absorption spectra could be recorded in alcohols for all dyes, presence of free molecules and aggregates in solution certainly modify the position of the absorption maxima. A summary of the absorption maxima for the twenty dyes are provided in Tables 3 and 4.
As evidenced in the Tables 3 and 4 [55], and Bakhshiev's [56] scales can be cited as the most popular ones. Among all scales, the Kamlet-Taft solvent polarity scale proved to be the most adapted one, linear correlations being obtained for all dyes by plotting the absorption maximum vs. The empirical Taft parameters (see linear regression in Supplementary Materials). For all the other polarity scales based on the dielectric constant or the refractive index of solvents, no reasonable correlations could be established. The Kamlet and Taft equation is also a multiparametric equation that can take into account the dipolarity-polarizability (π*), the hydrogen-donating and accepting ability (α and β) of the solvents, modelizing more precisely the interactions between the solvent and the solute. However, multiple linear regression analyses carried out on the triparametric Kamlet-Taft equation using the three solvent descriptors (α, β, π*) did not improve the correlation coefficients, as demonstrated in the Table in SI. It can therefore be concluded that the dipolarity-polarizability of the solvent is the primary cause influencing the position of the ICT band.
As evidenced in the Figure 9 and Figures in Supplementary Materials, PP1-PP20 show negative slopes with good linear correlations, indicative of a positive solvatochromism. Excepted for PP1, PP2, PP11, and PP12 that possess weak electron donors, all dyes displayed strong negative slopes, indicative of a significant charge redistribution upon excitation. The most important solvatochromism was found for PP9, PP10, PP19, and PP20 that exhibit the longest conjugated spacers. account the dipolarity-polarizability (π*), the hydrogen-donating and accepting ability (α and β) of the solvents, modelizing more precisely the interactions between the solvent and the solute. However, multiple linear regression analyses carried out on the triparametric Kamlet-Taft equation using the three solvent descriptors (α, β, π*) did not improve the correlation coefficients, as demonstrated in the Table in SI. It can therefore be concluded that the dipolarity-polarizability of the solvent is the primary cause influencing the position of the ICT band.
As evidenced in the Figure 9 and Figures in Supplementary Materials, PP1-PP20 show negative slopes with good linear correlations, indicative of a positive solvatochromism. Excepted for PP1, PP2, PP11, and PP12 that possess weak electron donors, all dyes displayed strong negative slopes, indicative of a significant charge redistribution upon excitation. The most important solvatochromism was found for PP9, PP10, PP19, and PP20 that exhibit the longest conjugated spacers.  Table 4. Summary of the optical properties of PP11-PP20 in twenty-three solvents and Kamlet and Taft parameters π*.

Fluorescence Spectroscopy
The fluorescent properties of all compounds were investigated in dichloromethane and toluene, in diluted solutions and the results are summarized in the Figure 10 and the Table 5. Interestingly, most of the chromophores were not emissive whatever the solvent is and this behavior is consistent with results classically reported in the literature for indane-1,3-dione derivatives [57][58][59][60][61]. However, it has to be noticed that a series of indane-1,3-dione derivatives was reported as being highly emissive, in a specific context, by use of oligo(phenylene)vinylene as electron donors [62]. Only eleven of these compounds displayed a weak emission in toluene so that the luminescence lifetime as well as the photoluminescence quantum yield were not determined for these compounds. While examining the emission maxima, the most red-shifted emissions were found for PP9 and PP10 displaying the most extended conjugated spacer but also the most extended electron acceptor. As attended, a blue-shifted emission was found for PP20 (λ em = 604 nm) relative to that of PP10 (λ em = 626 nm), directly resulting from its blue-shifted absorption maxima. While comparing the results obtained in dichloromethane and toluene, absorption maxima were found to be blue-shifted by ca. 20 nm for all dyes in toluene relative to that determined in dichloromethane. Based on the Taft parameters for both solvents (0.54 for toluene and 0.82 for dichloromethane), it can be concluded that a positive solvatochromism can also be observed in emission. This point is notably confirmed by the Stokes shifts determined in both dichloromethane and toluene, which are almost identical. examining the emission maxima, the most red-shifted emissions were found for PP9 and PP10 displaying the most extended conjugated spacer but also the most extended electron acceptor. As attended, a blue-shifted emission was found for PP20 (λem = 604 nm) relative to that of PP10 (λem = 626 nm), directly resulting from its blue-shifted absorption maxima. While comparing the results obtained in dichloromethane and toluene, absorption maxima were found to be blue-shifted by ca. 20 nm for all dyes in toluene relative to that determined in dichloromethane. Based on the Taft parameters for both solvents (0.54 for toluene and 0.82 for dichloromethane), it can be concluded that a positive solvatochromism can also be observed in emission. This point is notably confirmed by the Stokes shifts determined in both dichloromethane and toluene, which are almost identical.

Electrochemical Properties
The electrochemical properties of all compounds have been investigated by cyclic voltammetry (CV) in dilute solutions, either in acetonitrile or in dichloromethane. The selected voltammograms are shown in the Figure 11 and CV curves of all compounds are given in the Supplementary Materials. The redox potentials of all compounds are summarized in the Table 6 in which redox potentials are given against the half wave oxidation potential of the ferrocene/ferrocenium cation couple.

Electrochemical Properties
The electrochemical properties of all compounds have been investigated by cyclic voltammetry (CV) in dilute solutions, either in acetonitrile or in dichloromethane. The selected voltammograms are shown in the Figure 11 and CV curves of all compounds are given in the Supplementary Materials. The redox potentials of all compounds are summarized in the Table 6 in which redox potentials are given against the half wave oxidation potential of the ferrocene/ferrocenium cation couple. As shown in Figure 11, PP9 and PP19 differ from each other only in the nature of the accepting moiety. As expected, both compounds have quasi-reversible oxidation processes with identical oxidation potentials ( Figure 11, Table 6). Indeed, the oxidation process for the two chromophores is centered onto the electron-donating part and this latter is the same for the two dyes. Conversely, the reduction potential of PP9 comprising 1H-cyclopenta[b]naphthalene-1,3(2H)-dione (EA4) as the electron-accepting moiety is slightly lower than that of PP19 (comprising EA1 as the acceptor), leading to narrower electrochemical bandgap. This is in good accordance with the optical bandgap determined by UV-visible absorption spectroscopy, where PP9 showed a red-shifted ICT band in comparison to PP19 (See Figures 7 and 8).
The redox potentials of all other compounds are gathered in the Table 6. As can be seen in the   As shown in Figure 11, PP9 and PP19 differ from each other only in the nature of the accepting moiety. As expected, both compounds have quasi-reversible oxidation processes with identical oxidation potentials ( Figure 11, Table 6). Indeed, the oxidation process for the two chromophores is centered onto the electron-donating part and this latter is the same for the two dyes. Conversely, the reduction potential of PP9 comprising 1H-cyclopenta[b]naphthalene-1,3(2H)-dione (EA4) as the electron-accepting moiety is slightly lower than that of PP19 (comprising EA1 as the acceptor), leading to narrower electrochemical bandgap. This is in good accordance with the optical bandgap determined by UV-visible absorption spectroscopy, where PP9 showed a red-shifted ICT band in comparison to PP19 (See Figures 7 and 8).
The redox potentials of all other compounds are gathered in the Table 6. As can be seen in the same series (PP1-PP10 and PP11-PP20) where the nature of the acceptor moiety is identical, their reduction potentials changed very slightly while the oxidation potential importantly vary as a function of the donor moiety.
The number and the substitution position of the alkoxy chains on the phenyl ring slightly influence the redox potentials of the targeted molecules (PP1-3 and PP11-13). However, important variations were determined when the electron-donating ability of the electron donor was increased by the presence of diakylamino groups on the phenyl ring (see PP4-5 and PP14-15). This phenomenon was even much more pronounced when a double bond was inserted between the donor and the acceptor moiety leading to more conjugated push pull molecules (see PP9-10 & PP19-20). The presence of two 4-(N,N-methylamino)phenyl groups such as in PP10 and PP20 has only a negligible impact on the electrochemical property. In fact, while the second 4-(N,N-methylamino)phenyl group could increase the electron donating ability, examination of the mesomeric forms in PP10 and PP20 clearly evidences the two groups not to be able to contribute simultaneously to the electronic delocalization, as previously mentioned in the literature [46]. While comparing the dyes at identical electron donating groups, push pull compounds prepared with EA4 (PP1-PP10) have lower reduction potentials than their counterparts PP11-PP20 comprising EA1 as the electron withdrawing groups. This is in perfect accordance with the higher electron accepting capacity of EA4.
The redox behaviors of synthesized molecules were then used to estimate their HOMO and LUMO energy levels by using the ferrocene (Fc) ionization potential value (4.8 eV vs. vacuum) as the standard. The correcting factor of 4.8 eV is based on calculations obtained by Pommerehne et al. [63]. It is also important to note that some other correcting factors have also been used in the literature [64]. The obtained values of E HOMO and E LUMO issued from electrochemical characterizations are summarized in the Table 6 for comparison, the optical bandgaps of all dyes in acetonitrile have been added. The Figure 12 shows a comparative presentation of the frontier orbitals' energy levels experimentally and theoretically obtained. We can see that the experimental findings fit well with the trend predicted by DFT calculation.

Conclusions
In this work, a series of twenty dyes comprising indane-1,3-dione or 1Hcyclopenta[b]naphthalene-1,3(2H)-dione have been synthesized and examined for their photophysical properties. Introduction of a naphthalene moiety in EA4 greatly contributed to improve the electron-withdrawing ability of the group. Notably, a red-shift of the intramolecular charge transfer band of ca. 30 nm was observed for all dyes prepared with this acceptor, compared to the parent series comprising EA1. A linear and positive solvatochromism was found for all dyes, demonstrating an important charge redistribution upon excitation. A good correlation between the experimental HOMO-LUMO gaps and the theoretical ones could be found. Examination of their electrochemical properties revealed these dyes to be reversibly oxidized whereas an irreversible reduction monoelectronic process was determined for all dyes. By extending the aromaticity of EA4, a significant red-shift of the absorption maximum could be obtained for all dyes compared their analogues comprising EA1. Future works will consist of further improving the electron accepting ability of EA4, which is achievable by functionalizing EA4 with malononitrile. Based on the present work, the tetracyano derivatives of EA4 will certainly allow the design of near-infrared dyes that can find applications in research fields such as telecommunications and defense.
Funding: This research was funded by Aix Marseille University and The Centre National de la Recherche (CNRS) for financial supports. This research was also funded by the Agence Nationale de la Recherche (ANR agency) through the PhD grants of Corentin Pigot (ANR-17-CE08-0010 DUALITY project) and Guillaume Noirbent (ANR-17-CE08-0054 VISICAT project).

Conflicts of Interest:
The authors declare no conflict of interest.