Polycyclic aromatic hydrocarbon-substituted push–pull chromophores: an investigation of optoelectronic and nonlinear optical properties using experimental and theoretical approaches

A series of new push–pull chromophores were synthesized in moderate to very high yields (65%–97%) by treating TCNE and TCNQ with alkynes substituted by electron-rich diethylaniline and polycyclic aromatic hydrocarbons. Some of the chromophores exhibit strong intramolecular charge-transfer bands in the near-IR region with λmax values between 695 and 749 nm. With the help of experimental and theoretical analysis, it is concluded that the trend in λ max values is affected by PAH substituents sterically, not electronically. Steric constraints led to the increased dihedral angles, reducing conjugation efficiencies. The absorption properties of push-pull compounds have been investigated in solvents possessing different polarities. All chromophores exhibited positive solvatochromism. As an additional proof of efficient charge-transfer in push–pull chromophores, quinoid character (dr) values were predicted using calculated bond lengths. Remarkably, substantial dr values (0.045–0.049) were predicted for donor diethylaniline rings in all compounds. The effects of various polycyclic aromatic hydrocarbons on optical and nonlinear optical properties were also studied by computational methods. Several parameters, such as band gaps, Mulliken electronegativity, chemical hardness and softness, dipole moments, average polarizability, first hyperpolarizability, were predicted for chromophores at the B3LYP/6-31++G(d,p) level of theory. The predicted first hyperpolarizability β(tot) values vary between 198 to 538 × 10–30 esu for the reported push–pull chromophores in this study. The highest predicted β(tot) value in this study is 537.842 × 10–30 esu, 8150 times larger than the predicted β(tot) value of benchmark NLO material urea, suggests possible utilization of these chromophores in NLO devices. The charge-transfer character of the synthesized structures was further confirmed by HOMO-LUMO depictions and electrostatic potential maps.


Synthesis of donor-substituted substrates
It was the initial goal of the project to prepare PAH-substituted alkynes by considering several key points such as solubility of the substrates, sufficient donor-activation for the subsequent CA-RE transformations, and commercial availability of the reagents. The reaction sequence to access substrates 4, 8, and 12, required for the synthesis of target push-pull structures initiated with the conversion of the aryl bromides 1, 5, and 9 to TMS-protected alkynes 2, 6, and 10 via Sonogashira crosscoupling reactions (Scheme 1). All three compounds 2, 6, and 10 were successfully synthesized according to the literature procedure reported by Ouyang and co-workers [39]. Upon subsequent deprotection using K 2 CO 3 in MeOH/THF mixture, aryl alkynes 3, 7, and 11 were obtained in 86, 73, and 93 yields, respectively [39]. The final Sonogashira cross-coupling of PAH-substituted alkynes 3, 7, and 11 and N,N-diethyl-4-iodoaniline provided target donor-substituted substrates 4, 8, and 12 in 71, 90, and 64 yields, respectively. The reported synthesis could also be carried out between aryl halides 1, 5, 9, and N,N-diethyl-4-ethynylaniline. However, the stability of all three PAH-substituted alkynes under ambient conditions, in contrast to N,N-diethyl-4-ethynylaniline, guided us to follow the proposed synthetic strategy.
PAH (naphthene and phenanthrene) and donor group (diethylaniline)-substituted acetylenes 4, 8, and 12 were treated with strong acceptors TCNE/TCNQ for the final [2+2] CA-RE step under ambient conditions, which successfully afforded target push-pull chromophores in high yields. Even with the bulky TCNQ, reactions proceeded smoothly at room temperature (Scheme 2). No significant difference was observed regarding the effect of substituent positions on isolated yields. The yields were slightly lower in CA-RE transformations of TCNQ compared to that of TCNE, mainly due to solubility and purification difficulties that were encountered during column chromatography.
Although there are two possible regioisomers 13 and 19 that can be formed theoretically during the reaction of TCNQ and alkyne 4, the reaction was fully regioselective and afforded only 13 (Scheme 3). The reason behind this well-studied selectivity can easily be deducted from resonance structures of 13 and 19. Intramolecular charge transfer breaks the aromaticity of the diethylaniline ring (I) while forming a new one (II), as in the case of 13 [1,26].

UV/vis absorption spectra
The formal [2+2] CA-RE of 4, 8, 12 with TCNE/TCNQ provided dark green and dark orange chromophores, respectively. The UV/vis absorption data was in full agreement with this observation (Figure 1a  As it will be discussed in more detail in the theoretical section, the trend in λ max values is affected by PAH substituents sterically, not electronically. The steric restrictions from the substitution pattern of the PAHs 15 and 17 lead to a substantial deconjugation, as indicated by the large dihedral angle between the acceptor and the donor groups (15: 33 ° and 17: 34 °). The dihedral angle in chromophore 13 was relatively smaller (28 °) compared to that of 15 and 17. The large dihedral angle between the diethyl aniline donor and the cyano-based acceptor moiety prevents efficient linear π-conjugation in 15 and 17, which leads to a bathochromic shift [41,42]. A similar but completely opposite trend was observed for the compounds 14 (18 o  65% in yield 97% in yield 83% in yield 97% in yield 81% in yield 95% in yield All push-pull systems 13-18 feature strong CT bands that are assigned to intramolecular charge transfer from the donor diethylaniline group to acceptor polycyano units accessed by the CA-RE of TCNE and TCNQ. To further support this claim, protonation-reneutralization experiments were conducted by using CF 3 COOH (TFA) and NEt 3 (Figure 2a and b). Treatment of selected chromophores with TFA resulted in the disappearance of the CT bands via protonation of the diethylaniline unit. Upon treatment with NEt 3 , re-neutralization occurred, and CT bands were successfully recovered. With these experiments, the CT nature of the low energy bands has been confirmed [40,43,44].
All push-pull systems 13-18 feature strong CT bands that are assigned to intramolecular charge transfer from the donor diethylaniline group to acceptor polycyano units accessed by the CA-RE of TCNE and TCNQ. To further support this claim, protonation-reneutralization experiments were conducted by using CF 3 COOH (TFA) and NEt 3 (Figure 2a   diethylaniline unit. Upon treatment with NEt 3 , re-neutralization occurred, and CT bands were successfully recovered.
With these experiments, the CT nature of the low energy bands has been confirmed [40,43,44].

Theoretical studies
All calculations (DFT and TD-DFT) were performed using Gaussian 09 program package [45]. The effect of different functionals and basis sets on chromophore 14 has been evaluated. Since optimization results were obtained for B3LYP   optimizations were obtained at the functional using basis set B3LYP/6-31G(d). As a solvation model, a conductor-like polarizable continuum model (CPCM) in CH 2 Cl 2 was utilized. Initially, the highest occupied molecular orbital (HOMO) -lowest unoccupied molecular orbital (LUMO) and electrostatic potential (ESP) map analyses were used to explain ICT behavior in push-pull systems 13-18. As can be seen in Table 1, HOMO is mainly located on the diethylaniline region of the molecules in all chromophores 13-18. On the other hand, LUMO covers the area where cyano groups are present. Together with the partial overlap of HOMO and LUMO frontier orbitals, CT from diethylaniline groups to cyano groups can be claimed for the designed push-pull systems. In addition to FMO analysis, electrostatic potential maps were also analyzed to further evaluate CT interactions in push-pull chromophores. ESP representations describe the total charge density and the molecular polarity in push-pull systems [46]. Red and blue color codes have been used to identify the most negative and positive areas, respectively. As expected, red regions were located around cyano groups; on the other hand, blue regions were identified around the donor diethylaniline donor group.
The energy level diagrams of push-pull dyes are depicted in Figure 5. The estimated frontier orbital energy levels showed that there is a substantial increase in the HOMO-LUMO band-gap of TCNE adducts 14, 16, and 18 compared to TCNQ adducts 13, 15, and 17 due to the shortened π-conjugation pathway in the former case. The theoretical band gaps for push-pull dyes are 1.87, 2.57, 1.72, 2.30, 1.73, 2.28 eV, respectively. The theoretically obtained band gap values are slightly higher compared to the optical band gaps 1.44, 2.28, 1.31, 1.82, 1.31 1.77 eV [47]. No significant improvement in band gaps were obtained from the TD-DFT band gaps 2.01, 2.77, 1.88, 2.52, 1.89, 2.51 eV for compounds 13, 14, 15, 16, 17, and 18 respectively. However, the trend in theoretical band-gap values can also be seen in experimental band-gap values for both TCNE and TCNQ adducts. E red,1 values were reported for the benchmark cyano acceptors, such as TCNE (-0.32 V), and TCNQ (-0.25 V). A key property of this class of compounds is their electron-accepting power increasing from TCNE to TCNQ. It is also well-known that CA-RE products from TCNQ possess smaller band-gap compared to the products from TCNE. This can also be explained by the increased conjugation length in between donor and acceptor cyano groups in TCNQ products [1]. Figure 6 shows both theoretical and experimental UV/vis spectra for the selected compounds 13 and 14. Timedependent density theory calculations (CAM-B3LYP/6-31G(d)) were applied on optimized push-pull structures of 13-18 with CPCM solvation in CH 2 Cl 2 . In order to match the theoretical UV/vis spectra with the corresponding theoretical spectra, wavelengths were red-shifted by 0.2 eV and 0.6 eV for compounds 13 and 14, respectively. Scaling of extinction  coefficients was also required (scaled by 2.29 for 13 and by 1.87 for 14) since these values are slightly overestimated by TD-DFT calculations. However, the overall shapes of the theoretical spectra are relatively well-estimated (see Figures 6a and  b). Although the reason for the overestimation in molar absorption coefficients in TD-DFT calculations is still unclear, the choice of basis set and solvation method may be responsible from a slight inaccuracy in dihedral angles that may facilitate donor-acceptor conjugation efficiency and increase the molar absorption coefficients. All push-pull chromophores 13-18 possess charge-transfer bands originated from HOMO to LUMO transitions. The observed error in the calculated transitions are in the expected range for similar push-pull systems reported in the literature [48].
Quinoid character values (dr) are commonly used to predict the amount of charge transferred in push-pull systems [11,19]. If there is no charge-transfer, dr value equals 0 and represent a perfect benzene structure. On the other hand, fully quinoidal structures possess dr values in between 0.08-0.1. Bond lengths from optimized geometries of compounds 13-18 have been used for the dr value calculations (Table 2). A substantial quinoid character values (dr = 0.045-0.049) were predicted for diethylaniline donor groups in all chromophores. A slight increase in dr values of TCNQ adducts was observed compared to TCNE adducts. The predicted quinoid characters were comparable to earlier reports on CA-RE products [11,19]. Inspired by the literature on strong D-A systems [49,50] these results were further confirmed by total average atomic charges (δ) by ESP fitting on donor, acceptor and PAH parts of push-pull chromophores. Atomic charges were  calculated based on the ESP fitting scheme of Merz-Singh-Kollman (MK) [51]. All chromophores were divided into three parts: diethylaniline donor, PAH substituents, and cyano-based acceptor groups (Table 2) The successful utilization of organic molecules in nonlinear optics motivates researchers to develop rational design strategies [52,53]. The fast and inexpensive theoretical calculations compared to experimental measurements provide a great advantage for the design of NLOphores with tailor-made properties. The most common NLOphore structures are generally containing D-π-A-type molecular frameworks [54,55].   (2) μ = [(μ x ) 2 + (μ y ) 2 +( μ z ) 2 ] 1/2 (3) χ = -1/2 (E HOMO + E LUMO ) (4) η = -1/2 (E HOMO -E LUMO ) (5) σ = 1/ η (6) Overall results are summarized in Table 3. Molecular geometry is crucial as it dictates the NLO properties of the compounds. There are several strategies to modulate NLO responses, such as changing solvent choice, altering donoracceptor strength, or π-linker length. In this study, we mainly focused on the effect of PAH rings on average polarizability [α (tot) ] and first hyperpolarizability [β (tot) ] values. When we compare TCNE and TCNQ adducts, the lowest [α (tot) ] and [β (tot) ] values were predicted for chromophores 13 and 14, which possess naphthalene groups substituted at two positions. For the rest of the compounds, differences in [α (tot) ] and [β (tot) ] values are not very significant. Compounds with the bulky phenanthrene or naphthene group substituted at one position exhibited almost similar NLO responses. This prediction can be supported by the calculated and optically measured band gaps for 13 and 14. The HOMO-LUMO energy gap for 13 is significantly larger compared to 15 and 17. A similar trend can also be seen in TCNQ adducts. We found a general trend by which β (tot) increases with the size of the spacer between donor and acceptor groups as can be seen by higher β (tot) values in TCNQ products compared to those of TCNE products. We presume that smaller band-gap results in more efficient charge transfer and, as a result, larger NLO responses, as can be seen in Table 2. Optimized structures also displayed significant deviation from planarity in the case of compounds 15 (33 o (18 o ). Accordingly, we have shown that substituent groups can play an important role in modulating NLO properties of push-pull type chromophores, although they are mainly utilized as solubilizing groups or side groups to improve the physical properties of chromophores. The highest predicted β (tot) value in this study is 537.842 × 10 -30 esu for chromophore 17. That value is 8150 times larger than the benchmark NLO material urea, β (tot) value of 0.066 × 10 -30 esu, calculated at the CAM-B3LYP/6-31++G(d,p) [22]. In the final part, the chemical properties of chromophores will be discussed using equations (4)(5)(6). The results are very promising when compared with the literature. For example, pushpull system, p-nitroaniline, possesses β (tot) value of 9. Koopmans' theorem states that HOMO and LUMO energies are related to ionization potential and electron affinity, respectively. Accordingly, the Mulliken electronegativity (χ) can be estimated by equation 4. Besides electronegativity, chemical hardness (η) is another term to be used for chemical behavior predictions of materials. Global hardness is directly related to the HOMO-LUMO gap and can be defined as the resistance of an atom to charge-transfer. Equations 4 and 5 show that larger HOMO-LUMO gaps are required to improve hardness values. Compounds 14, 16, 18 are predicted to have higher global hardness values with their larger band gaps compared to 13, 15, and 17. An opposite trend can be seen in global softness (σ), values as expected from Equation 6. In summary, compounds 13, 15, and 17 are expected to be more reactive compared to 14, 16, and 18.

Conclusions
In this study, 6 new polycyclic aromatic hydrocarbon-substituted push-pull chromophores have been synthesized using click-type [2+2] cycloaddition-retroelectrocyclization. The synthetic approach worked smoothly under ambient laboratory conditions without requiring heat or catalyst. Optoelectronic properties of the highly-colored chromophores were investigated by using both experimental (UV/vis spectroscopy) and computational methods. Solvatochromism and pH studies were performed for all 6 chromophores. Computational studies (TD-DFT, electrostatic potential maps, HOMO-LUMO orbital depictions) were further confirmed that all chromophores undergo intramolecular charge transfer. The HOMO-LUMO energy gap of dye 13 is found to have the largest energy gap resulting in worse charge-transfer properties as compared to 15 and 17. A similar trend was also observed for chromophores 14, 16, and 18. PAH-substituted push-pull chromophores are predicted to have significant NLO properties, and these properties can simply be tuned by changing substituent PAH structures. The present study provides valuable knowledge for the design and synthesis of new NLOphore systems in the near future.