CHEMPHYSCHEM ARTICLES DOI: 10.1002/cphc.201402123
BODIPY-Bridged Push–Pull Chromophores for Nonlinear Optical Applications Gilles Ulrich,*[a] Alberto Barsella,[b] Alex Boeglin,[b] Songlin Niu,[a] and Raymond Ziessel*[a] A set of linear and dissymmetric BODIPY-bridged push–pull dyes are synthesized. The electron-donating substituents are anisole and dialkylanilino groups. The strongly electron-accepting moiety, a 1,1,4,4-tetracyanobuta-1,3-diene (TCBD) group, is obtained by insertion of an electron-rich ethyne into tetracyanoethylene. A nonlinear push–pull system is developed with a donor at the 5-position of the BODIPY core and the acceptor at the 2-position. All dyes are fully characterized and their electrochemical, linear and nonlinear optical properties are discussed. The linear optical properties of dialkylamino compounds show strong solvatochromic behavior and undergo
drastic changes upon protonation. The strong push–pull systems are non-fluorescent and the TCBD-BODIPY dyes show diverse photochemistry and electrochemistry, with several reversible reduction waves for the tetracyanobutadiene moiety. The hyperpolarizability mb of selected compounds is evaluated using the electric-field-induced second-harmonic generation technique. Two of the TCBD-BODIPY dyes show particularly high mb (1.907 mm) values of 2050 1048 and 5900 1048 esu. In addition, one of these dyes shows a high NLO contrast upon protonation–deprotonation of the donor residue.
1. Introduction Boron dipyrromethene (BODIPY) dyes are well known for their strong absorption bands in the UV/Vis region, relatively sharp emission bands, high solubility in common organic solvents, high fluorescence quantum yields, relatively long excitation lifetimes (1–5 ns), and the high chemical and photochemical stability of their core p system.[1] These properties have made BODIPY fluorophores candidates for optical applications. Moreover, the construction of a “push–pull” system upon the BODIPY scaffold has led to new properties, such as photochromism, enhanced charge transport, and recombination of p-conjugated electronic states. These properties could provide the basis of new organic photovoltaic devices.[2] Several instances of the use of two-photon excitation of BODIPY chromophores have been reported in the fields of cell imaging and telecommunications,[3–7] but the study of BODIPY-based push– pull systems in these areas has been somewhat limited.[8, 9] Nonlinear optical (NLO) responses are among the most intensively investigated properties of push–pull chromophores.[10] Push–pull molecules typically possess large groundstate dipole moments (m) and their second-order susceptibility
[a] Dr. G. Ulrich, Dr. S. Niu, Dr. R. Ziessel ICPEES–LCOSA, UMR 7515 CNRS/Universit de Strasbourg, ECPM 25 rue Becquerel, 67087 Strasbourg Cedex 02 (France) E-mail: [email protected] [email protected] [b] Dr. A. Barsella, Dr. A. Boeglin IPCMS, UMR 7504 CNRS/Universit de Strasbourg 23 rue du Loess, BP 43, 67034 Strasbourg Cedex 02 (France) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201402123.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
(also referred to as the first hyperpolarizability, b)[11] plays a dominant role in their contributions to observed NLO responses.[12] Polar organic chromophores with quadratic NLO properties are valued for their ability to undergo two-photon absorption processes or to double the frequency of incident radiation—useful properties for advanced microscopy techniques.[13] They are also proposed as new materials in the field of telecommunications for their ability to be oriented in external electric fields, which allows, for example, the switching or modulation of optical signals.[14] Extensive research, both theoretical and experimental, has thus been devoted to understanding the properties of push–pull chromophores in order to obtain molecules with large first hyperpolarizabilities for NLO applications.[3, 4, 10, 15–18] From first principles in molecular physics, it has been concluded that an asymmetric distribution of the p electrons within a conjugated molecule will lead to an NLO response. Oudar’s two-state model has been widely used as the basis for designing organic NLO molecules having a large b.[19, 20] According to this model, b is proportional to the change in dipole moment (Dm) between the ground (mgg) and excited states (mee) as well as to the square of the transition dipole moment (mge). Also, b is inversely proportional to the square of the energy gap (Ege) between these two states.[21] Therefore, b is intrinsically related to the molecular charge-transfer excited states.[11, 22] However, it was also concluded that simply increasing the donor/acceptor strength does not necessarily enhance the hyperpolarizability, regardless of whether the two states are fully delocalized or localized,[16, 23] whereas an optimal response should be achieved at an intermediate point where there is an appreciable overlap of the wave functions that define the two states.[11] Thus, an overlap between the HOMO ChemPhysChem 2014, 15, 2693 – 2700
2693
CHEMPHYSCHEM ARTICLES and the LUMO in the bridging region is necessary to obtain large hyperpolarizability in a push–pull system.[24, 25] Attributing the quadratic NLO response of a chromophore to its lowest (charge-transfer) transition is only an approximate way to explain the phenomenon; in practice, the contributions of all the excited states,[11] as well as multipolar transitions[26] should be included. Nevertheless, Marder and co-workers have initiated a successful approach for the design of push–pull chromophores by considering their states as combinations of two limiting resonance forms.[10, 16, 27–30] These studies recognize the importance of the bridges linking the donor and acceptor groups in determining the NLO response of the molecule. Various spacers (e.g. ethynyl, phenylene, small heteroaromatic rings, styrenes, and a combination of two or more of these simple units) have been used together with large chromophores, such as porphyrins, in the synthesis of NLO materials.[10, 16, 21, 30–33] Whereas much progress has been achieved in the fundamental understanding of the structure–property relationships of push–pull systems, as well as in the engineering of molecular materials with large NLO responses, issues of (photo-)stability of both the chromophore and the host material still necessitate further synthesis research in the field. Recently, in two fields of particular interest for us, chromophores based on functionalized merocyanines have been studied as push–pull dyes for organic photovoltaics,[34–36] and fluorene molecular probes for two-photon absorption and NLO cell imaging,[37–39] . In our investigations, we have worked with pentamethyl-BODIPY dyes, the general molecular structure and numbering of which is given in Scheme 1.
Scheme 1. General formula and numbering of the pentamethyl-BODIPY framework.
In this work, we studied a series of BODIPY-based push–pull chromophores.[40, 41] The main strategy involved introduction of an electron-donating group and a strongly electron-accepting group through ethynyl linkers to opposite sides of the BODIPY core (Figure 1). A similar strategy has been previously developed, with the classical nitrophenyl group as an acceptor linked to the 2-position of a BODIPY.[8] The linear and nonlinear optical properties of these new compounds have been evaluated.
2. Results and Discussion The dyes of interest (Scheme 2) were prepared and purified as previously described.[40, 41] The design was dictated by the need to construct molecules with strong electron-donor residues (e.g. anisole or dibutylamino) and strong electron-withdrawing groups [e.g. malonitrile or tetracyanobuta-1,3-diene (TCBD)], 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
Figure 1. A push–pull chromophore in its ground (top) and excited (bottom) states with dipole moments mgg and mee.
and also by the easy access to the respective carbaldehyde or alkyne derivatives. The main strategy for preparing the compounds of interest (Scheme 2) was based on the following steps: 1) halogenation of the 2-position and/or the 6-position; 2) substituting these positions with a phenylalkyne function carrying either the electron-donor residues (MeOPh or nBu2NPh) or an aldehyde suitable for reaction with malonitrile; 3) selective Knoevenagel reaction at the 5-methyl group, after grafting a moiety at the 2-position, for tuning the linear optical properties between 500– 600 nm; 4) selective insertion of the alkynes into tetracyanoethylene (TCNE) to afford the powerful electron acceptor. In most cases, excellent yields were obtained and the selectivity for single insertion into TCNE relative to the reaction with a second molecule of TCNE is remarkable. Furthermore, the formation of a single vinyl derivative at the 5-position (the opposite side of the BODIPY core from the alkyne at the 2-position) as well as the selective reaction of TCNE with the triple bond versus the double bond have been carefully addressed using NMR tools. This is a general synthetic strategy that allows the preparation of libraries of BODIPY dyes. 2.1. Linear Optical Properties The optical properties of the push–pull chromophores were evaluated in different solvents and are summarized in Table 1. In solution, compounds 1 and 2 (Scheme 2) show a strong S0–S1 transition between l 540–570 nm, with an absorption coefficient in the 73 000–94 000 m1 cm1 range, which is unambiguously assigned to the boradiazaindacene chromophore.[42] At higher energy, two overlapping, weak, broad bands centered around 410 nm and 350 nm are attributed to the chargetransfer absorption[43] and the S0–S2 transition of the BODIPY moiety, respectively (Figure 2).[44] In addition, p–p* transitions with vibronic structure are observed at 230–320 nm for the phenyl groups. In CH2Cl2, the BODIPY dyes show emission maxima in the 561–616 nm range. The fluorescence quantum yields (F) are in the range 28–53 %. This fluorescence is independent of the excitation wavelength, and the excitation spectra almost match the absorption spectra over the range of wavelengths measured. The influence of the solvent on the fluorescence is a more important factor, with a decrease of F and the lifetime (t) as the dipole moment of the solvent inChemPhysChem 2014, 15, 2693 – 2700
2694
CHEMPHYSCHEM ARTICLES
Scheme 2. Push–pull dyes used in the present studies.
creases (Table 1). Conversely, the Stokes shifts increase slightly in more polar solvents. For compound 1, F and t decrease when the solvent changes from dioxane (0.46, 2.4 ns) to EtOH (0.16, 1.1 ns). The rate of non-radiative decay (knr) increases (approximately ninefold), whereas that of radiative decay (kr) remains similar. The small changes in Stokes shift in various solvents might indicate that the difference in dipole moments between the excited states and the ground states is relatively modest.[44] The absorption spectra of the TCBD derivatives 4–7 (Scheme 2) in dioxane (Figures 3–5) are much different than those of compounds 1 and 2, and show intense absorption around 475–560 nm with weaker molar absorption coefficients in the 36 000–77 000 m1 cm1 range and larger peak shape. For compound 3 (Figure 3), a pronounced hypsochromic shift of 60 nm is found compared with 2. This shift results from a desta 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org bilization of the HOMO and an increase of the optical gap. The broad, lower-energy shoulder is likely due to a charge-transfer transition inherent to the push– pull effect. The absorption spectrum of 4 (Figure 3) is different due to the presence of a second anisole moiety (a push fragment). The main absorption band, centered at 540 nm, is broad, not solvent dependent, and likely originates from the configuration of the dye in solution and from absorption transitions that are due to the presence of several charge-transfer bands. The absorption coefficient of this band is lower (36 000 m1 cm1) compared to that of 3. The absorption around 410 nm can be attributed to the S0–S2 transition with an absorption coefficient of approximately 15 000 m1 cm1 (Figure 3). Surprisingly, no significant solvent effect is observed in the absorption spectra of these push–pull dyes, with a variation of less than 3 nm from a solvent with a weak dipole moment (dioxane) to one with a stronger dipole (ethanol). Neither 3 nor 4 are fluorescent in solution. The electronic absorption spectra of 5 and 6 also have in common a similar broad, structureless band between 450– 700 nm, with a molar absorption coefficient in the 60 000– 77 000 m1 cm1 range (Figure 4). In comparison with the parent compound bearing two p-dibutylaminophenylacetylene residues at the 2- and 6-positions, the lowest energy absorption maxima of 5 and 6 show bathochromic shifts of 25 and 40 nm, respectively. The broad absorption band can arise from an inhomogeneous distribution of the solvation sphere and charge transfer that is due to the various conformations taken up by each TCBD group. Compared to that of 5, the spectrum of the symmetric 6 is significantly hypsochromically shifted by 16 nm and has a 30 % larger extinction coefficient. At shorter wavelengths, 5 has two overlapping absorptions between 260–400 nm, which correspond to the p–p* transitions of the free alkylaminophenylethylene group and, more specifically around 370 nm, of the TCBD fragment. Interestingly, in dye 6 the latter transition is enhanced as might be expected by the presence of two TCBD moieties, whereas the low-energy transition is absent in the ChemPhysChem 2014, 15, 2693 – 2700
2695
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
spectrum (Figure 4). Again, little influence of the solvent dipole e [M1 cm1] lem [nm] F[a] t [ns] Compound Solvent labs [nm] moment and no fluorescence were observed in solution. 1 dioxane 565 73 000 610 46 2.4 The absorption spectra of CH2Cl2 569 82 000 616 28 1.7 EtOH 563 74 000 612 16 1.1 TCBD dye 7 (Figure 5) show 535 94 000 561 53 2.3 2 CH2Cl2 broad bands between 500– 800 nm, with a maximum cen3 CH2Cl2 475 46 000 – – – tered at 600 nm (e 45 000 m1 4 dioxane 543 36 000 – – – cm1). Compared with the specEtOH 540 35 000 – – – trum of the ethynyl precursor, 5 dioxane 574 60 000 – – – this represents a hypsochromic EtOH 575 57 000 – – – shift of 21 nm and a decrease of 6 dioxane 558 77 000 – – – EtOH 554 74 000 – – – the extinction coefficient by 599 45 500 – – – 7 dioxane about 50 %. Weak solvatochrom+ 579 39 300 – – – dioxane + H ism is observed in solvents with 602 44 200 – – – EtOH higher dipole moments. Upon 572 36 500 – – – EtOH + H + protonation of 7 with anhydrous 6 [a] Quantum yield determined for dilute solutions (1 10 m) using rhodamine 6G (FF = 0.78 in H2O, lexc = HCl gas, a hypsochromic shift of 488 nm), or cresyl violet as reference (FF = 0.51 in EtOH, lexc = 578 nm). Protonation was performed with HCl(g).[45] about 20 and 30 nm is observed in dioxane and ethanol, respectively (Table 1). In all cases, the fluorescence of the chromophore is completely quenched, probably due to an oxidative photoinduced electron transfer from the BODIPY* to the nitrile group.[46–48] A strong driving force of DG0 720 meV has been calculated using the Rehm–Weller equation, disregarding the electrostatic effects.[49, 50] The optical transition E00 2.08 eV has been estimated using the absorption profile and the steady-state emission of the 5-styryl-2-ethynylanisole precursor.[51] Table 1. Optical properties of the dyes.
2.2. Nonlinear Optical Properties
Figure 2. Absorption (105 m) and emission (106 m) spectra of compounds 1 and 2 in CH2Cl2 at room temperature. Excitation wavelengths were 490 and 480 nm for compounds 1 and 2, respectively.
Figure 3. Absorption spectra of 3 and 4 in dioxane (105 m, room temperature).
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The hyperpolarizability of the push–pull chromophores was investigated using the electric-field-induced second-harmonic generation (EFISHG) technique and the results are shown in Figure 6. The experiments were performed on chloroform solutions of the BODIPY dyes using an Nd:YAG laser source that was actively Q-switched at 10 Hz and emitted 7 ns pulses at
Figure 4. Absorption spectra of 5 and 6 in dioxane (105 m, room temperature).
ChemPhysChem 2014, 15, 2693 – 2700
2696
CHEMPHYSCHEM ARTICLES
Figure 5. Absorption spectra of 7 and 7 + H + in dioxane (105 m, room temperature).
the wavelength of 1.064 mm. These pulses were Raman-shifted to 1.907 mm through a high-pressure hydrogen cell so that their second harmonic at 953 nm was clearly separated from the absorption bands of the chromophores. Absolute values have been defined against a quartz wedge reference by taking its quadratic susceptibility to be d11 = 1.2 109 esu at 1.064 mm and by extrapolating it to 1.1 109 esu at 1.907 mm. In EFISHG, a fundamental-mode-locked laser beam is focused into a solution of the chromophore being analyzed, to which a strong static electric field is applied. The interaction of the field with the permanent dipole moments of the molecules causes a bias in the average orientation of the molecules. The partial removal of the isotropy allows second-harmonic genera-
www.chemphyschem.org tion to occur.[1, 5] The intensity of the detected second-harmonic light is proportional to mb, the scalar product of the permanent dipole moment with the vectorial part of the hyperpolarizability tensor.[3, 52, 53] The vectorial part can be extracted by separately measuring the permanent dipole moment,[1, 54] however, such measurements must be delicately performed and we have not made any attempts to do so with these BODIPY compounds.[1, 19] In EFISHG, a minimal dye concentration of 103 m is typically needed for a meaningful measurement. The poor solubility of compound 1 in CHCl3 therefore made these measurements difficult. Nevertheless, 2 was found to be highly soluble although the mb value is somewhat small.[2] Surprisingly, 3, with the strongly electron-withdrawing group TCBD, showed almost no NLO response. Dye 4 has a mb value comparable to that of 2. Interestingly, being conjugated with a strong electron-donating group, the mono- and di-TCBD derivatives 5 and 6 show high mb NLO responses of 1500 1048 and 2050 1048 esu, respectively. The p-dimethylaminostyryl 7 has the highest mb value of all the molecules in this study, 5900 1048 esu, which is almost four times that of 5. Another interesting property of 7 is that it shows a strong NLO contrast upon protonation–deprotonation. EFISHG measurements have been recently used to quantify this type of effect in diphenyl-(4-{2-[4-(2-pyridin-4-yl-vinyl)-phenyl]-vinyl}phenyl)-amine (DPVPA).[55] The NLO response of 7 can be reduced from 5900 1048 to 680 1048 esu by exposing the solution to HCl vapor. This causes protonation of the dimethylamino group, which results in an 80 % NLO contrast, significantly higher than the one reported in DPVPA molecules.[55] This process is completely reversible, as the high mb value can be
Figure 6. The mb values for the push–pull chromophores in CHCl3 solution, TEA = triethylamine.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2693 – 2700
2697
CHEMPHYSCHEM ARTICLES restored by adding a drop of triethylamine to the protonated solution, which restores the original dimethylamino donor group. 2.3. Calculations Geometry optimizations of BODIPY compounds were performed with the program MOPAC2009[56] using the recent PM6[57] semi-empirical Hamiltonian and the COSMO solvation model[58] for chloroform (see the Supporting Information for the effect of solvation model). MOPAC2009 has been used in a recent study of analogous chromophores.[8] The permanent ground-state dipole moment vector components and the zerofrequency hyperpolarizability tensor components were obtained using the same Hamiltonian for at least two stable geometries of each compound. In all cases, it was possible to converge to distinct geometries after starting from initial structures that essentially differed by a 1808 rotation of one of the substituting groups around the axis that was defined by a bond linking it to the BODIPY core (see Figures S3–S17 for proximal projections of geometries used). To uncover a rationale behind the trends in the experimental quadratic responses, the calculations were made only for the dyes that were characterized successfully by EFISHG (2–7). Unfortunately, the purely semi-empirical approach yielded erratic results. In particular, for the TCBD dyes, even the sign of the response depends on which one of two orientations was adopted by the electron-withdrawing group. Indeed, the calculated mb values for dye 4 change from 140 1048 to 220 1048 esu and for dye 7 from 235 1048 to 365 1048 esu as the TCBD unit is rotated by 1808. For dye 5, the corresponding values are 300 1048 and 80 1048 esu. In the case of dye 6, which has two TCBD moieties, the three distinct geometries they can adopt with respect to the BODIPY core give rise to the three values of 200 1048, 800 1048, and 1360 1048 esu. It can be concluded that the changes in sign merely reflect the fact that the values given by these semi-empirical calculations are so low as to be meaningless, with the possible exception for dye 6. With the development of the CAM–B3LYP functional,[59] into which the idea of long-range correction to the coulomb correlation into the popular hybrid functional B3LYP is implemented, the performance of density functional theory (DFT) has been greatly enhanced with respect to the description of charge-transfer systems. Because this functional has recently been shown to satisfactorily reproduce EFISHG results in push– pull phenyl-polyenes,[53] we have used a suitably patched version of the DALTON program[60] to evaluate the properties of dyes 3 and 7. Because geometry optimizations could not be carried out with this implementation, we could only evaluate CAM–B3LYP properties for PM6 geometries. For dye 3, the quadratic responses are still negative, but almost an order of magnitude lower than at the semi-empirical level, and are consistent with the fact that the experimental value is low and therefore difficult to measure. For the two geometries of dye 7, we obtained two positive values of 360 1048 and 900 1048 esu at the DFT level. Although an improvement 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org with respect to the PM6 value, the best result is still too small by a factor of six with respect to the experimental target value of 5900 1048 esu. We must conclude that the interpretation of the NLO properties of the BODIPY dyes cannot be based on an effective two-level charge-transfer model. This also implies that the extrapolation of the experimental (finite frequency) results to static responses, although useful as a first approximation when comparing dyes with different charge-transfer absorption bands, is questionable here also. We are thus led to attribute the large quadratic response to three-level contributions to b where only the three transition dipole moments come into play. Moreover, in the formal expression for b, one of the terms corresponding to such contributions becomes resonant when the transition between the two excited states occurs at twice the frequency of the excitation field. Unfortunately, when one of the transitions becomes resonant, it is necessary to include not only the vibrational structure but also the environment-induced broadening of the vibronic levels in the general expression for the first hyperpolarizability.[61] The situation is more complex in the case of TCBD-containing dyes. Because of the large amplitude motions these groups undergo in solution at room temperature, a representation in terms of vibronic states falls short in the description of dynamics that give rise to non-adiabatic couplings between electronic excited states. Under these circumstances, it is incorrect to expect realistic predictions from response calculations based on the electronic structure of a representative equilibrium structure. However, we might consider the large amplitude motions to be slow with respect to the three optical transitions that are involved in the generation of the second harmonic from the excitation field. Indeed, this process does not require the intermediate states to be populated, but only for transient coherences between these states to be created, as happens, for example, in stimulated Raman processes. For our purposes, we can thus view the absorption bands of the BODIPY dyes as being inhomogeneously broadened, and we can explain the surprisingly high quadratic responses of some of them as stemming from resonantly enhanced three-level contributions to b. To test this idea, excited states were computed for BODIPY dyes at the PM6 geometries in the INDO/S Hamiltonian of the ZINDO program.[62] Between 460–624 singly excited configurations were used in these computations, depending on the size of the molecule and the nature of the occupied and virtual orbitals. Although the excited state properties derived from such calculations are far from definitive, they provide a generally useful qualitative first assessment. For all the compounds, and all their geometries, the transition to the first excited state is the one that carries the largest oscillator strength, but always leads to a reduction in permanent dipole moment (Supporting Information). For a conventional push–pull chromophore for which the charge-transfer band rules the quadratic property, this would imply a negative first hyperpolarizability. Here, however, the change in permanent dipole moment is approximately orthogonal to the transition dipole moment, resulting in a vanishing contribution. For the BODIPY dyes studied here, this clearly implies that their properties derive from higher exChemPhysChem 2014, 15, 2693 – 2700
2698
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
cited states. Dye 6 stands out as being the compound in which some of the Sn S0 transitions show substantial alignment between their transition dipole moments and the corresponding difference (increase) in permanent dipole moment. Thus 6 is likely to derive at least part of its strong response from additive two-level contributions from higher excited states. Dye 7 on the other hand is remarkable because both the ground and first excited state have rather strong transition dipole moments of 6–8 Debye to the S5 state with an S5 S2 energy gap of 10 000 cm1, close to twice the photon energy of the 1.907 mm excitation source used in our EFISHG experiment (Figure 7). Dye 7 is therefore likely to derive its high response from resonance enhancement of a three-level contribution to b.
ior of its quadratic susceptibility as its first absorption band is scanned. Dye 7, on the other hand, has the potential for improved quadratic responses at specific wavelength ranges because of resonance effects between excited states, yet remains perfectly transparent. There is evidence for interesting NLO contrast upon protonation–deprotonation cycles. Future studies will investigate the use of fluorescent hyperpolarizable compounds in polymers and thin films.
!
!
Acknowledgements This work was financially supported by the CNRS providing research facilities and financial support. The authors thank Dr. Alexander Tarnovsky and his lab for assisting with the transient absorption measurements. Keywords: BODIPY · dyes/pigments · fluorescence · nonlinear optics · push–pull chromophores
!
Figure 7. ZINDO S1 S0 transition dipole moment for dye 7 (see also the Supporting Information).
3. Conclusions Several BODIPY bridge compounds have been synthesized and studied. Compounds with strong push–pull character were obtained by reaction of TCNE with an electron-rich ethyne bond. A nonlinear donor–BODIPY–acceptor system was obtained with a p-dimethylaminostyryl donor function at the 5-position of the BODIPY and formation of TCBD-group acceptor at the 2-position. As expected, push–pull compounds exhibit solvatochromic behavior and a reduced emission efficiency with increasing solvent dipole moment. The TCBD moiety strongly stabilizes the radical anion and the dianion, and moderately affects the reduction and oxidation of the BODIPY. The oxidation of the BODIPY core is reversible and much easier than the reversible oxidation of the dimethylamino function in 7 (Supporting Information). Evaluation of the quadratic responses (mb) of several of the compounds using EFISHG measurements has shown the TCBD-containing compounds 6 and 7 to have particularly interesting NLO responses, although probably for different microscopic reasons. Dye 6 is unconventional in the sense that its first absorption band probably does not contribute to its quadratic response, resulting in a monotonic behav 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[1] a) A. Loudet, K. Burgess, Chem. Rev. 2007, 107, 4891 – 4932; b) G. Ulrich, R. Ziessel, A. Harriman, Angew. Chem. Int. Ed. 2008, 47, 1184 – 1207; Angew. Chem. 2008, 120, 1202 – 1219; c) R. Ziessel, G. Ulrich, A. Harriman, New J. Chem. 2007, 31, 496 – 501; d) N. Boens, V. Leen, W. Dehaen, Chem. Soc. Rev. 2012, 41, 1130 – 1172. [2] a) S. Erten-Ela, M. D. Yilmaz, B. Icli, Y. Dede, S. Icli, E. U. Akkaya, Org. Lett. 2008, 10, 3299 – 3302; b) S. Kolemen, Y. Cakmak, S. Erten-Ela, Y. Altay, J. Brendel, M. Thelakkat, E. U. Akkaya, Org. Lett. 2010, 12, 3812 – 3815. [3] P. A. Bouit, K. Kamada, P. Feneyrou, G. Berginc, L. Toupet, O. Maury, C. Andraud, Adv. Mater. 2009, 21, 1151 – 1154. [4] P. Didier, G. Ulrich, Y. Mely, R. Ziessel, Org. Biomol. Chem. 2009, 7, 3639 – 3642. [5] Q. Zheng, G. Xu, P. Prasad, Chem. Eur. J. 2008, 14, 5812 – 5819. [6] D. Zhang, Y. Wang, Y. Xiao, S. Qian, X. Qian, Tetrahedron 2009, 65, 8099 – 8103. [7] Q. Zheng, G. S. He, P. N. Prasad, Chem. Phys. Lett. 2009, 475, 250 – 255. [8] W.-J. Shi, P.-C. Lo, A. Singh, I. Ledoux-Rak, D. K. P. Ng, Tetrahedron 2012, 68, 8712 – 8718. [9] a) D. Collado, J. Casado, S. R. Gonzlez, J. T. L. Navarrete, R. Suau, E. Perez-Inestrosa, T. M. Pappenfus, M. M. M. Raposo, Chem. Eur. J. 2011, 17, 498 – 507; b) R. Ziessel, P. Retailleau, K. J. Elliott, A. Harriman, Chem. Eur. J. 2009, 15, 10369 – 10374. [10] M. Blanchard-Desce, V. Alain, P. V. Bedworth, S. R. Marder, A. Fort, C. Runser, M. Barzoukas, S. Lebus, R. Wortmann, Chem. Eur. J. 1997, 3, 1091 – 1104. [11] D. R. Kanis, M. A. Ratner, T. J. Marks, Chem. Rev. 1994, 94, 195 – 242. [12] a) B. F. Levine, Chem. Phys. Lett. 1976, 37, 516 – 520; b) S. J. Lalama, A. F. Garito, Phys. Rev. A 1979, 20, 1179 – 1194. [13] a) P. T. C. So, C. Y. Dong, B. R. Masters, K. M. Berland, Annu. Rev. Biomed. Eng. 2000, 2, 399 – 429; b) F. M. Geiger, Ann. Rev. Phys. Chem. 2009, 60, 61 – 83. [14] L. R. Dalton, P. A. Sullivan, D. H. Bale, Chem. Rev. 2010, 110, 25 – 55. [15] a) T. Verbiest, S. Houbrechts, M. Kauranen, K. Clays, A. Persoons, J. Mater. Chem. 1997, 7, 2175 – 2189; b) F. Wrthner, R. Wortmann, K. Meerholz, ChemPhysChem 2002, 3, 17 – 31. [16] S. R. Marder, B. Kippelen, A. K. Y. Jen, N. Peyghambarian, Nature 1997, 388, 845 – 851. [17] A. J. Moore, A. Chesney, M. R. Bryce, A. S. Batsanov, J. F. Kelly, J. A. K. Howard, I. F. Perepichka, D. F. Perepichka, G. Meshulam, G. Berkovic, Z. Kotler, R. Mazor, V. Khodorkovsky, Eur. J. Org. Chem. 2001, 2001, 2671 – 2687. [18] N. Bloembergen, Selected Topics in Quantum Electronics, IEEE J. Sel. Top. Quantum Eletron. 2000, 6, 876 – 880. [19] a) J. F. Ward, Rev. Mod. Phys. 1965, 37, 1 – 18; b) J. L. Oudar, J. Chem. Phys. 1977, 67, 446 – 457. [20] J. L. Oudar, H. Le Person, Opt. Commun. 1975, 15, 258 – 262.
ChemPhysChem 2014, 15, 2693 – 2700
2699
CHEMPHYSCHEM ARTICLES [21] S. M. LeCours, H.-W. Guan, S. G. DiMagno, C. H. Wang, M. J. Therien, J. Am. Chem. Soc. 1996, 118, 1497 – 1503. [22] J. Zyss, J. Non-Cryst. Solids 1982, 47, 211 – 225. [23] M. C. Zerner, W. M. F. Fabian, R. Dworczak, D. W. Kieslinger, G. Kroner, H. Junek, M. E. Lippitsch, Int. J. Quantum Chem. 2000, 79, 253 – 266. [24] T. Yoshimura, Phys. Rev. B 1989, 40, 6292 – 6298. [25] a) T. Yoshimura, Mol. Cryst. Liq. Cryst. Inc. Nonlinear Opt. 1990, 182, 43 – 50; b) T. Yoshimura, Appl. Phys. Lett. 1989, 55, 534 – 536. [26] J. Zyss, I. Ledoux, Chem. Rev. 1994, 94, 77 – 105. [27] S. R. Marder, C. B. Gorman, B. G. Tiemann, J. W. Perry, G. Bourhill, K. Mansour, Science 1993, 261, 186 – 189. [28] F. Meyers, S. R. Marder, B. M. Pierce, J. L. Bredas, J. Am. Chem. Soc. 1994, 116, 10703 – 10714. [29] I. D. L. Albert, T. J. Marks, M. A. Ratner, J. Am. Chem. Soc. 1997, 119, 6575 – 6582. [30] I. D. L. Albert, T. J. Marks, M. A. Ratner, Chem. Mater. 1998, 10, 753 – 762. [31] R. A. Huijts, G. L. J. Hesselink, Chem. Phys. Lett. 1989, 156, 209 – 212. [32] P. R. Varanasi, A. K.-Y. Jen, J. Chandrasekhar, I. N. N. Namboothiri, A. Rathna, J. Am. Chem. Soc. 1996, 118, 12443 – 12448. [33] I. F. Perepichka, D. F. Perepichka, M. R. Bryce, A. Chesney, A. F. Popov, V. Khodorkovsky, G. Meshulam, Z. Kotler, Synth. Met. 1999, 102, 1558 – 1559. [34] W. Frank, M. Klaus, Chem. Eur. J. 2010, 16, 9366 – 9373. [35] A. V. Kulinich, A. A. Isschenko, Russ. Chem. Rev. 2009, 78, 141 – 164. [36] H. Burckstummer, N. M. Kronenberg, K. Meerholz, F. Wrthner, Org. Lett. 2010, 12, 3666 – 3669. [37] K. D. Belfield, D. J. Hagan, E. W. Van Stryland, K. J. Schafer, R. A. Negres, Org. Lett. 1999, 1, 1575 – 1578. [38] C. Barsu, R. Cheaib, S. Chambert, Y. Queneau, O. Maury, D. Cottet, H. Wege, J. Douady, Y. Bretonniere, C. Andraud, Org. Biomol. Chem. 2010, 8, 142 – 150. [39] K. D. Belfield, K. J. Schafer, W. Mourad, B. A. Reinhardt, J. Org. Chem. 2000, 65, 4475 – 4481. [40] S. Niu, G. Ulrich, P. Retailleau, R. Ziessel, Tetrahedron Lett. 2011, 52, 4848 – 4853. [41] S. Niu, G. Ulrich, P. Retailleau, R. Ziessel, Org. Lett. 2011, 13, 4996 – 4999. [42] W. Qin, M. Baruah, M. Van der Auweraer, F. C. De Schryver, N. Boens, J. Phys. Chem. A 2005, 109, 7371 – 7384.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org [43] W. Qin, T. Rohand, M. Baruah, A. Stefan, M. van der Auweraer, W. Dehaen, N. Boens, Chem. Phys. Lett. 2006, 420, 562 – 568. [44] J. Karolin, L. B. A. Johansson, L. Strandberg, T. Ny, J. Am. Chem. Soc. 1994, 116, 7801 – 7806. [45] J. Olmsted, J. Phys. Chem. 1979, 83, 2581 – 2584. [46] A. P. de Silva, H. Q. N. Gunaratne, C. P. McCoy, J. Am. Chem. Soc. 1997, 119, 7891 – 7892. [47] A. Coskun, E. Deniz, E. U. Akkaya, Org. Lett. 2005, 7, 5187 – 5189. [48] M. A. Fox, Photochem. Photobiol. 1990, 52, 617 – 627. [49] D. Rehm, A. Weller, Ber. Bunsen-Ges. 1969, 73, 834 – 839. [50] D. Rehm, A. Weller, Isr. J. Chem. 1970, 8, 259 – 271. [51] R. Ziessel, L. Bonardi, P. Retailleau, G. Ulrich, J. Org. Chem. 2006, 71, 3093 – 3102. [52] I. Maltey, J. A. Delaire, K. Nakatani, P. Wang, X. Shi, S. Wu, Adv. Mater. Opt. Elec. 1996, 6, 233 – 238. [53] L. Ferrighi, L. Frediani, C. Cappelli, P. Salek, H. Agren, T. Helgaker, K. Ruud, Chem. Phys. Lett. 2006, 425, 267 – 272. [54] F. Wrthner, F. Effenberger, R. Wortmann, P. Krmer, Chem. Phys. 1993, 173, 305 – 314. [55] E. Cariati, D. Dragonetti, E. Lucenti, F. Nisic, S. Righetto, R. E. Tordin, Chem. Commun. 2014, 50, 1608 – 1610. [56] J. P. J. Stewart, Computational Chemistry, Colorado Springs, CO, 2008, http://OpenMOPAC.net. [57] J. P. J. Stewart, J. Mol. Model. 2007, 13, 1173 – 1213. [58] A. Klamt, G. Schrmann, J. Chem. Soc. Perkin Trans. 2 1993, 799 – 805. [59] H. Iikura, T. Tsuneda, T. Yanai, K. Hirao, J. Chem. Phys. 2001, 115, 3540 – 3544. [60] DALTON, Release 2.0, a molecular electronic structure program (with CAM – B3LYP functional enabled) 2005, http://www.kjemi.uio.no/software/dalton/dalton.html. [61] C. H. Wang, J. Chem. Phys. 2000, 112, 1917 – 1924. [62] ZINDO version of the Cerius2 suite of programs by MSI under license from Accelrys.
Received: March 12, 2014 Revised: May 12, 2014 Published online on June 20, 2014
ChemPhysChem 2014, 15, 2693 – 2700
2700
BODIPY-Bridged Push–Pull Chromophores for Nonlinear Optical Applications Gilles Ulrich,*[a] Alberto Barsella,[b] Alex Boeglin,[b] Songlin Niu,[a] and Raymond Ziessel*[a] A set of linear and dissymmetric BODIPY-bridged push–pull dyes are synthesized. The electron-donating substituents are anisole and dialkylanilino groups. The strongly electron-accepting moiety, a 1,1,4,4-tetracyanobuta-1,3-diene (TCBD) group, is obtained by insertion of an electron-rich ethyne into tetracyanoethylene. A nonlinear push–pull system is developed with a donor at the 5-position of the BODIPY core and the acceptor at the 2-position. All dyes are fully characterized and their electrochemical, linear and nonlinear optical properties are discussed. The linear optical properties of dialkylamino compounds show strong solvatochromic behavior and undergo
drastic changes upon protonation. The strong push–pull systems are non-fluorescent and the TCBD-BODIPY dyes show diverse photochemistry and electrochemistry, with several reversible reduction waves for the tetracyanobutadiene moiety. The hyperpolarizability mb of selected compounds is evaluated using the electric-field-induced second-harmonic generation technique. Two of the TCBD-BODIPY dyes show particularly high mb (1.907 mm) values of 2050 1048 and 5900 1048 esu. In addition, one of these dyes shows a high NLO contrast upon protonation–deprotonation of the donor residue.
1. Introduction Boron dipyrromethene (BODIPY) dyes are well known for their strong absorption bands in the UV/Vis region, relatively sharp emission bands, high solubility in common organic solvents, high fluorescence quantum yields, relatively long excitation lifetimes (1–5 ns), and the high chemical and photochemical stability of their core p system.[1] These properties have made BODIPY fluorophores candidates for optical applications. Moreover, the construction of a “push–pull” system upon the BODIPY scaffold has led to new properties, such as photochromism, enhanced charge transport, and recombination of p-conjugated electronic states. These properties could provide the basis of new organic photovoltaic devices.[2] Several instances of the use of two-photon excitation of BODIPY chromophores have been reported in the fields of cell imaging and telecommunications,[3–7] but the study of BODIPY-based push– pull systems in these areas has been somewhat limited.[8, 9] Nonlinear optical (NLO) responses are among the most intensively investigated properties of push–pull chromophores.[10] Push–pull molecules typically possess large groundstate dipole moments (m) and their second-order susceptibility
[a] Dr. G. Ulrich, Dr. S. Niu, Dr. R. Ziessel ICPEES–LCOSA, UMR 7515 CNRS/Universit de Strasbourg, ECPM 25 rue Becquerel, 67087 Strasbourg Cedex 02 (France) E-mail: [email protected] [email protected] [b] Dr. A. Barsella, Dr. A. Boeglin IPCMS, UMR 7504 CNRS/Universit de Strasbourg 23 rue du Loess, BP 43, 67034 Strasbourg Cedex 02 (France) Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cphc.201402123.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
(also referred to as the first hyperpolarizability, b)[11] plays a dominant role in their contributions to observed NLO responses.[12] Polar organic chromophores with quadratic NLO properties are valued for their ability to undergo two-photon absorption processes or to double the frequency of incident radiation—useful properties for advanced microscopy techniques.[13] They are also proposed as new materials in the field of telecommunications for their ability to be oriented in external electric fields, which allows, for example, the switching or modulation of optical signals.[14] Extensive research, both theoretical and experimental, has thus been devoted to understanding the properties of push–pull chromophores in order to obtain molecules with large first hyperpolarizabilities for NLO applications.[3, 4, 10, 15–18] From first principles in molecular physics, it has been concluded that an asymmetric distribution of the p electrons within a conjugated molecule will lead to an NLO response. Oudar’s two-state model has been widely used as the basis for designing organic NLO molecules having a large b.[19, 20] According to this model, b is proportional to the change in dipole moment (Dm) between the ground (mgg) and excited states (mee) as well as to the square of the transition dipole moment (mge). Also, b is inversely proportional to the square of the energy gap (Ege) between these two states.[21] Therefore, b is intrinsically related to the molecular charge-transfer excited states.[11, 22] However, it was also concluded that simply increasing the donor/acceptor strength does not necessarily enhance the hyperpolarizability, regardless of whether the two states are fully delocalized or localized,[16, 23] whereas an optimal response should be achieved at an intermediate point where there is an appreciable overlap of the wave functions that define the two states.[11] Thus, an overlap between the HOMO ChemPhysChem 2014, 15, 2693 – 2700
2693
CHEMPHYSCHEM ARTICLES and the LUMO in the bridging region is necessary to obtain large hyperpolarizability in a push–pull system.[24, 25] Attributing the quadratic NLO response of a chromophore to its lowest (charge-transfer) transition is only an approximate way to explain the phenomenon; in practice, the contributions of all the excited states,[11] as well as multipolar transitions[26] should be included. Nevertheless, Marder and co-workers have initiated a successful approach for the design of push–pull chromophores by considering their states as combinations of two limiting resonance forms.[10, 16, 27–30] These studies recognize the importance of the bridges linking the donor and acceptor groups in determining the NLO response of the molecule. Various spacers (e.g. ethynyl, phenylene, small heteroaromatic rings, styrenes, and a combination of two or more of these simple units) have been used together with large chromophores, such as porphyrins, in the synthesis of NLO materials.[10, 16, 21, 30–33] Whereas much progress has been achieved in the fundamental understanding of the structure–property relationships of push–pull systems, as well as in the engineering of molecular materials with large NLO responses, issues of (photo-)stability of both the chromophore and the host material still necessitate further synthesis research in the field. Recently, in two fields of particular interest for us, chromophores based on functionalized merocyanines have been studied as push–pull dyes for organic photovoltaics,[34–36] and fluorene molecular probes for two-photon absorption and NLO cell imaging,[37–39] . In our investigations, we have worked with pentamethyl-BODIPY dyes, the general molecular structure and numbering of which is given in Scheme 1.
Scheme 1. General formula and numbering of the pentamethyl-BODIPY framework.
In this work, we studied a series of BODIPY-based push–pull chromophores.[40, 41] The main strategy involved introduction of an electron-donating group and a strongly electron-accepting group through ethynyl linkers to opposite sides of the BODIPY core (Figure 1). A similar strategy has been previously developed, with the classical nitrophenyl group as an acceptor linked to the 2-position of a BODIPY.[8] The linear and nonlinear optical properties of these new compounds have been evaluated.
2. Results and Discussion The dyes of interest (Scheme 2) were prepared and purified as previously described.[40, 41] The design was dictated by the need to construct molecules with strong electron-donor residues (e.g. anisole or dibutylamino) and strong electron-withdrawing groups [e.g. malonitrile or tetracyanobuta-1,3-diene (TCBD)], 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org
Figure 1. A push–pull chromophore in its ground (top) and excited (bottom) states with dipole moments mgg and mee.
and also by the easy access to the respective carbaldehyde or alkyne derivatives. The main strategy for preparing the compounds of interest (Scheme 2) was based on the following steps: 1) halogenation of the 2-position and/or the 6-position; 2) substituting these positions with a phenylalkyne function carrying either the electron-donor residues (MeOPh or nBu2NPh) or an aldehyde suitable for reaction with malonitrile; 3) selective Knoevenagel reaction at the 5-methyl group, after grafting a moiety at the 2-position, for tuning the linear optical properties between 500– 600 nm; 4) selective insertion of the alkynes into tetracyanoethylene (TCNE) to afford the powerful electron acceptor. In most cases, excellent yields were obtained and the selectivity for single insertion into TCNE relative to the reaction with a second molecule of TCNE is remarkable. Furthermore, the formation of a single vinyl derivative at the 5-position (the opposite side of the BODIPY core from the alkyne at the 2-position) as well as the selective reaction of TCNE with the triple bond versus the double bond have been carefully addressed using NMR tools. This is a general synthetic strategy that allows the preparation of libraries of BODIPY dyes. 2.1. Linear Optical Properties The optical properties of the push–pull chromophores were evaluated in different solvents and are summarized in Table 1. In solution, compounds 1 and 2 (Scheme 2) show a strong S0–S1 transition between l 540–570 nm, with an absorption coefficient in the 73 000–94 000 m1 cm1 range, which is unambiguously assigned to the boradiazaindacene chromophore.[42] At higher energy, two overlapping, weak, broad bands centered around 410 nm and 350 nm are attributed to the chargetransfer absorption[43] and the S0–S2 transition of the BODIPY moiety, respectively (Figure 2).[44] In addition, p–p* transitions with vibronic structure are observed at 230–320 nm for the phenyl groups. In CH2Cl2, the BODIPY dyes show emission maxima in the 561–616 nm range. The fluorescence quantum yields (F) are in the range 28–53 %. This fluorescence is independent of the excitation wavelength, and the excitation spectra almost match the absorption spectra over the range of wavelengths measured. The influence of the solvent on the fluorescence is a more important factor, with a decrease of F and the lifetime (t) as the dipole moment of the solvent inChemPhysChem 2014, 15, 2693 – 2700
2694
CHEMPHYSCHEM ARTICLES
Scheme 2. Push–pull dyes used in the present studies.
creases (Table 1). Conversely, the Stokes shifts increase slightly in more polar solvents. For compound 1, F and t decrease when the solvent changes from dioxane (0.46, 2.4 ns) to EtOH (0.16, 1.1 ns). The rate of non-radiative decay (knr) increases (approximately ninefold), whereas that of radiative decay (kr) remains similar. The small changes in Stokes shift in various solvents might indicate that the difference in dipole moments between the excited states and the ground states is relatively modest.[44] The absorption spectra of the TCBD derivatives 4–7 (Scheme 2) in dioxane (Figures 3–5) are much different than those of compounds 1 and 2, and show intense absorption around 475–560 nm with weaker molar absorption coefficients in the 36 000–77 000 m1 cm1 range and larger peak shape. For compound 3 (Figure 3), a pronounced hypsochromic shift of 60 nm is found compared with 2. This shift results from a desta 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org bilization of the HOMO and an increase of the optical gap. The broad, lower-energy shoulder is likely due to a charge-transfer transition inherent to the push– pull effect. The absorption spectrum of 4 (Figure 3) is different due to the presence of a second anisole moiety (a push fragment). The main absorption band, centered at 540 nm, is broad, not solvent dependent, and likely originates from the configuration of the dye in solution and from absorption transitions that are due to the presence of several charge-transfer bands. The absorption coefficient of this band is lower (36 000 m1 cm1) compared to that of 3. The absorption around 410 nm can be attributed to the S0–S2 transition with an absorption coefficient of approximately 15 000 m1 cm1 (Figure 3). Surprisingly, no significant solvent effect is observed in the absorption spectra of these push–pull dyes, with a variation of less than 3 nm from a solvent with a weak dipole moment (dioxane) to one with a stronger dipole (ethanol). Neither 3 nor 4 are fluorescent in solution. The electronic absorption spectra of 5 and 6 also have in common a similar broad, structureless band between 450– 700 nm, with a molar absorption coefficient in the 60 000– 77 000 m1 cm1 range (Figure 4). In comparison with the parent compound bearing two p-dibutylaminophenylacetylene residues at the 2- and 6-positions, the lowest energy absorption maxima of 5 and 6 show bathochromic shifts of 25 and 40 nm, respectively. The broad absorption band can arise from an inhomogeneous distribution of the solvation sphere and charge transfer that is due to the various conformations taken up by each TCBD group. Compared to that of 5, the spectrum of the symmetric 6 is significantly hypsochromically shifted by 16 nm and has a 30 % larger extinction coefficient. At shorter wavelengths, 5 has two overlapping absorptions between 260–400 nm, which correspond to the p–p* transitions of the free alkylaminophenylethylene group and, more specifically around 370 nm, of the TCBD fragment. Interestingly, in dye 6 the latter transition is enhanced as might be expected by the presence of two TCBD moieties, whereas the low-energy transition is absent in the ChemPhysChem 2014, 15, 2693 – 2700
2695
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
spectrum (Figure 4). Again, little influence of the solvent dipole e [M1 cm1] lem [nm] F[a] t [ns] Compound Solvent labs [nm] moment and no fluorescence were observed in solution. 1 dioxane 565 73 000 610 46 2.4 The absorption spectra of CH2Cl2 569 82 000 616 28 1.7 EtOH 563 74 000 612 16 1.1 TCBD dye 7 (Figure 5) show 535 94 000 561 53 2.3 2 CH2Cl2 broad bands between 500– 800 nm, with a maximum cen3 CH2Cl2 475 46 000 – – – tered at 600 nm (e 45 000 m1 4 dioxane 543 36 000 – – – cm1). Compared with the specEtOH 540 35 000 – – – trum of the ethynyl precursor, 5 dioxane 574 60 000 – – – this represents a hypsochromic EtOH 575 57 000 – – – shift of 21 nm and a decrease of 6 dioxane 558 77 000 – – – EtOH 554 74 000 – – – the extinction coefficient by 599 45 500 – – – 7 dioxane about 50 %. Weak solvatochrom+ 579 39 300 – – – dioxane + H ism is observed in solvents with 602 44 200 – – – EtOH higher dipole moments. Upon 572 36 500 – – – EtOH + H + protonation of 7 with anhydrous 6 [a] Quantum yield determined for dilute solutions (1 10 m) using rhodamine 6G (FF = 0.78 in H2O, lexc = HCl gas, a hypsochromic shift of 488 nm), or cresyl violet as reference (FF = 0.51 in EtOH, lexc = 578 nm). Protonation was performed with HCl(g).[45] about 20 and 30 nm is observed in dioxane and ethanol, respectively (Table 1). In all cases, the fluorescence of the chromophore is completely quenched, probably due to an oxidative photoinduced electron transfer from the BODIPY* to the nitrile group.[46–48] A strong driving force of DG0 720 meV has been calculated using the Rehm–Weller equation, disregarding the electrostatic effects.[49, 50] The optical transition E00 2.08 eV has been estimated using the absorption profile and the steady-state emission of the 5-styryl-2-ethynylanisole precursor.[51] Table 1. Optical properties of the dyes.
2.2. Nonlinear Optical Properties
Figure 2. Absorption (105 m) and emission (106 m) spectra of compounds 1 and 2 in CH2Cl2 at room temperature. Excitation wavelengths were 490 and 480 nm for compounds 1 and 2, respectively.
Figure 3. Absorption spectra of 3 and 4 in dioxane (105 m, room temperature).
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
The hyperpolarizability of the push–pull chromophores was investigated using the electric-field-induced second-harmonic generation (EFISHG) technique and the results are shown in Figure 6. The experiments were performed on chloroform solutions of the BODIPY dyes using an Nd:YAG laser source that was actively Q-switched at 10 Hz and emitted 7 ns pulses at
Figure 4. Absorption spectra of 5 and 6 in dioxane (105 m, room temperature).
ChemPhysChem 2014, 15, 2693 – 2700
2696
CHEMPHYSCHEM ARTICLES
Figure 5. Absorption spectra of 7 and 7 + H + in dioxane (105 m, room temperature).
the wavelength of 1.064 mm. These pulses were Raman-shifted to 1.907 mm through a high-pressure hydrogen cell so that their second harmonic at 953 nm was clearly separated from the absorption bands of the chromophores. Absolute values have been defined against a quartz wedge reference by taking its quadratic susceptibility to be d11 = 1.2 109 esu at 1.064 mm and by extrapolating it to 1.1 109 esu at 1.907 mm. In EFISHG, a fundamental-mode-locked laser beam is focused into a solution of the chromophore being analyzed, to which a strong static electric field is applied. The interaction of the field with the permanent dipole moments of the molecules causes a bias in the average orientation of the molecules. The partial removal of the isotropy allows second-harmonic genera-
www.chemphyschem.org tion to occur.[1, 5] The intensity of the detected second-harmonic light is proportional to mb, the scalar product of the permanent dipole moment with the vectorial part of the hyperpolarizability tensor.[3, 52, 53] The vectorial part can be extracted by separately measuring the permanent dipole moment,[1, 54] however, such measurements must be delicately performed and we have not made any attempts to do so with these BODIPY compounds.[1, 19] In EFISHG, a minimal dye concentration of 103 m is typically needed for a meaningful measurement. The poor solubility of compound 1 in CHCl3 therefore made these measurements difficult. Nevertheless, 2 was found to be highly soluble although the mb value is somewhat small.[2] Surprisingly, 3, with the strongly electron-withdrawing group TCBD, showed almost no NLO response. Dye 4 has a mb value comparable to that of 2. Interestingly, being conjugated with a strong electron-donating group, the mono- and di-TCBD derivatives 5 and 6 show high mb NLO responses of 1500 1048 and 2050 1048 esu, respectively. The p-dimethylaminostyryl 7 has the highest mb value of all the molecules in this study, 5900 1048 esu, which is almost four times that of 5. Another interesting property of 7 is that it shows a strong NLO contrast upon protonation–deprotonation. EFISHG measurements have been recently used to quantify this type of effect in diphenyl-(4-{2-[4-(2-pyridin-4-yl-vinyl)-phenyl]-vinyl}phenyl)-amine (DPVPA).[55] The NLO response of 7 can be reduced from 5900 1048 to 680 1048 esu by exposing the solution to HCl vapor. This causes protonation of the dimethylamino group, which results in an 80 % NLO contrast, significantly higher than the one reported in DPVPA molecules.[55] This process is completely reversible, as the high mb value can be
Figure 6. The mb values for the push–pull chromophores in CHCl3 solution, TEA = triethylamine.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPhysChem 2014, 15, 2693 – 2700
2697
CHEMPHYSCHEM ARTICLES restored by adding a drop of triethylamine to the protonated solution, which restores the original dimethylamino donor group. 2.3. Calculations Geometry optimizations of BODIPY compounds were performed with the program MOPAC2009[56] using the recent PM6[57] semi-empirical Hamiltonian and the COSMO solvation model[58] for chloroform (see the Supporting Information for the effect of solvation model). MOPAC2009 has been used in a recent study of analogous chromophores.[8] The permanent ground-state dipole moment vector components and the zerofrequency hyperpolarizability tensor components were obtained using the same Hamiltonian for at least two stable geometries of each compound. In all cases, it was possible to converge to distinct geometries after starting from initial structures that essentially differed by a 1808 rotation of one of the substituting groups around the axis that was defined by a bond linking it to the BODIPY core (see Figures S3–S17 for proximal projections of geometries used). To uncover a rationale behind the trends in the experimental quadratic responses, the calculations were made only for the dyes that were characterized successfully by EFISHG (2–7). Unfortunately, the purely semi-empirical approach yielded erratic results. In particular, for the TCBD dyes, even the sign of the response depends on which one of two orientations was adopted by the electron-withdrawing group. Indeed, the calculated mb values for dye 4 change from 140 1048 to 220 1048 esu and for dye 7 from 235 1048 to 365 1048 esu as the TCBD unit is rotated by 1808. For dye 5, the corresponding values are 300 1048 and 80 1048 esu. In the case of dye 6, which has two TCBD moieties, the three distinct geometries they can adopt with respect to the BODIPY core give rise to the three values of 200 1048, 800 1048, and 1360 1048 esu. It can be concluded that the changes in sign merely reflect the fact that the values given by these semi-empirical calculations are so low as to be meaningless, with the possible exception for dye 6. With the development of the CAM–B3LYP functional,[59] into which the idea of long-range correction to the coulomb correlation into the popular hybrid functional B3LYP is implemented, the performance of density functional theory (DFT) has been greatly enhanced with respect to the description of charge-transfer systems. Because this functional has recently been shown to satisfactorily reproduce EFISHG results in push– pull phenyl-polyenes,[53] we have used a suitably patched version of the DALTON program[60] to evaluate the properties of dyes 3 and 7. Because geometry optimizations could not be carried out with this implementation, we could only evaluate CAM–B3LYP properties for PM6 geometries. For dye 3, the quadratic responses are still negative, but almost an order of magnitude lower than at the semi-empirical level, and are consistent with the fact that the experimental value is low and therefore difficult to measure. For the two geometries of dye 7, we obtained two positive values of 360 1048 and 900 1048 esu at the DFT level. Although an improvement 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org with respect to the PM6 value, the best result is still too small by a factor of six with respect to the experimental target value of 5900 1048 esu. We must conclude that the interpretation of the NLO properties of the BODIPY dyes cannot be based on an effective two-level charge-transfer model. This also implies that the extrapolation of the experimental (finite frequency) results to static responses, although useful as a first approximation when comparing dyes with different charge-transfer absorption bands, is questionable here also. We are thus led to attribute the large quadratic response to three-level contributions to b where only the three transition dipole moments come into play. Moreover, in the formal expression for b, one of the terms corresponding to such contributions becomes resonant when the transition between the two excited states occurs at twice the frequency of the excitation field. Unfortunately, when one of the transitions becomes resonant, it is necessary to include not only the vibrational structure but also the environment-induced broadening of the vibronic levels in the general expression for the first hyperpolarizability.[61] The situation is more complex in the case of TCBD-containing dyes. Because of the large amplitude motions these groups undergo in solution at room temperature, a representation in terms of vibronic states falls short in the description of dynamics that give rise to non-adiabatic couplings between electronic excited states. Under these circumstances, it is incorrect to expect realistic predictions from response calculations based on the electronic structure of a representative equilibrium structure. However, we might consider the large amplitude motions to be slow with respect to the three optical transitions that are involved in the generation of the second harmonic from the excitation field. Indeed, this process does not require the intermediate states to be populated, but only for transient coherences between these states to be created, as happens, for example, in stimulated Raman processes. For our purposes, we can thus view the absorption bands of the BODIPY dyes as being inhomogeneously broadened, and we can explain the surprisingly high quadratic responses of some of them as stemming from resonantly enhanced three-level contributions to b. To test this idea, excited states were computed for BODIPY dyes at the PM6 geometries in the INDO/S Hamiltonian of the ZINDO program.[62] Between 460–624 singly excited configurations were used in these computations, depending on the size of the molecule and the nature of the occupied and virtual orbitals. Although the excited state properties derived from such calculations are far from definitive, they provide a generally useful qualitative first assessment. For all the compounds, and all their geometries, the transition to the first excited state is the one that carries the largest oscillator strength, but always leads to a reduction in permanent dipole moment (Supporting Information). For a conventional push–pull chromophore for which the charge-transfer band rules the quadratic property, this would imply a negative first hyperpolarizability. Here, however, the change in permanent dipole moment is approximately orthogonal to the transition dipole moment, resulting in a vanishing contribution. For the BODIPY dyes studied here, this clearly implies that their properties derive from higher exChemPhysChem 2014, 15, 2693 – 2700
2698
CHEMPHYSCHEM ARTICLES
www.chemphyschem.org
cited states. Dye 6 stands out as being the compound in which some of the Sn S0 transitions show substantial alignment between their transition dipole moments and the corresponding difference (increase) in permanent dipole moment. Thus 6 is likely to derive at least part of its strong response from additive two-level contributions from higher excited states. Dye 7 on the other hand is remarkable because both the ground and first excited state have rather strong transition dipole moments of 6–8 Debye to the S5 state with an S5 S2 energy gap of 10 000 cm1, close to twice the photon energy of the 1.907 mm excitation source used in our EFISHG experiment (Figure 7). Dye 7 is therefore likely to derive its high response from resonance enhancement of a three-level contribution to b.
ior of its quadratic susceptibility as its first absorption band is scanned. Dye 7, on the other hand, has the potential for improved quadratic responses at specific wavelength ranges because of resonance effects between excited states, yet remains perfectly transparent. There is evidence for interesting NLO contrast upon protonation–deprotonation cycles. Future studies will investigate the use of fluorescent hyperpolarizable compounds in polymers and thin films.
!
!
Acknowledgements This work was financially supported by the CNRS providing research facilities and financial support. The authors thank Dr. Alexander Tarnovsky and his lab for assisting with the transient absorption measurements. Keywords: BODIPY · dyes/pigments · fluorescence · nonlinear optics · push–pull chromophores
!
Figure 7. ZINDO S1 S0 transition dipole moment for dye 7 (see also the Supporting Information).
3. Conclusions Several BODIPY bridge compounds have been synthesized and studied. Compounds with strong push–pull character were obtained by reaction of TCNE with an electron-rich ethyne bond. A nonlinear donor–BODIPY–acceptor system was obtained with a p-dimethylaminostyryl donor function at the 5-position of the BODIPY and formation of TCBD-group acceptor at the 2-position. As expected, push–pull compounds exhibit solvatochromic behavior and a reduced emission efficiency with increasing solvent dipole moment. The TCBD moiety strongly stabilizes the radical anion and the dianion, and moderately affects the reduction and oxidation of the BODIPY. The oxidation of the BODIPY core is reversible and much easier than the reversible oxidation of the dimethylamino function in 7 (Supporting Information). Evaluation of the quadratic responses (mb) of several of the compounds using EFISHG measurements has shown the TCBD-containing compounds 6 and 7 to have particularly interesting NLO responses, although probably for different microscopic reasons. Dye 6 is unconventional in the sense that its first absorption band probably does not contribute to its quadratic response, resulting in a monotonic behav 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[1] a) A. Loudet, K. Burgess, Chem. Rev. 2007, 107, 4891 – 4932; b) G. Ulrich, R. Ziessel, A. Harriman, Angew. Chem. Int. Ed. 2008, 47, 1184 – 1207; Angew. Chem. 2008, 120, 1202 – 1219; c) R. Ziessel, G. Ulrich, A. Harriman, New J. Chem. 2007, 31, 496 – 501; d) N. Boens, V. Leen, W. Dehaen, Chem. Soc. Rev. 2012, 41, 1130 – 1172. [2] a) S. Erten-Ela, M. D. Yilmaz, B. Icli, Y. Dede, S. Icli, E. U. Akkaya, Org. Lett. 2008, 10, 3299 – 3302; b) S. Kolemen, Y. Cakmak, S. Erten-Ela, Y. Altay, J. Brendel, M. Thelakkat, E. U. Akkaya, Org. Lett. 2010, 12, 3812 – 3815. [3] P. A. Bouit, K. Kamada, P. Feneyrou, G. Berginc, L. Toupet, O. Maury, C. Andraud, Adv. Mater. 2009, 21, 1151 – 1154. [4] P. Didier, G. Ulrich, Y. Mely, R. Ziessel, Org. Biomol. Chem. 2009, 7, 3639 – 3642. [5] Q. Zheng, G. Xu, P. Prasad, Chem. Eur. J. 2008, 14, 5812 – 5819. [6] D. Zhang, Y. Wang, Y. Xiao, S. Qian, X. Qian, Tetrahedron 2009, 65, 8099 – 8103. [7] Q. Zheng, G. S. He, P. N. Prasad, Chem. Phys. Lett. 2009, 475, 250 – 255. [8] W.-J. Shi, P.-C. Lo, A. Singh, I. Ledoux-Rak, D. K. P. Ng, Tetrahedron 2012, 68, 8712 – 8718. [9] a) D. Collado, J. Casado, S. R. Gonzlez, J. T. L. Navarrete, R. Suau, E. Perez-Inestrosa, T. M. Pappenfus, M. M. M. Raposo, Chem. Eur. J. 2011, 17, 498 – 507; b) R. Ziessel, P. Retailleau, K. J. Elliott, A. Harriman, Chem. Eur. J. 2009, 15, 10369 – 10374. [10] M. Blanchard-Desce, V. Alain, P. V. Bedworth, S. R. Marder, A. Fort, C. Runser, M. Barzoukas, S. Lebus, R. Wortmann, Chem. Eur. J. 1997, 3, 1091 – 1104. [11] D. R. Kanis, M. A. Ratner, T. J. Marks, Chem. Rev. 1994, 94, 195 – 242. [12] a) B. F. Levine, Chem. Phys. Lett. 1976, 37, 516 – 520; b) S. J. Lalama, A. F. Garito, Phys. Rev. A 1979, 20, 1179 – 1194. [13] a) P. T. C. So, C. Y. Dong, B. R. Masters, K. M. Berland, Annu. Rev. Biomed. Eng. 2000, 2, 399 – 429; b) F. M. Geiger, Ann. Rev. Phys. Chem. 2009, 60, 61 – 83. [14] L. R. Dalton, P. A. Sullivan, D. H. Bale, Chem. Rev. 2010, 110, 25 – 55. [15] a) T. Verbiest, S. Houbrechts, M. Kauranen, K. Clays, A. Persoons, J. Mater. Chem. 1997, 7, 2175 – 2189; b) F. Wrthner, R. Wortmann, K. Meerholz, ChemPhysChem 2002, 3, 17 – 31. [16] S. R. Marder, B. Kippelen, A. K. Y. Jen, N. Peyghambarian, Nature 1997, 388, 845 – 851. [17] A. J. Moore, A. Chesney, M. R. Bryce, A. S. Batsanov, J. F. Kelly, J. A. K. Howard, I. F. Perepichka, D. F. Perepichka, G. Meshulam, G. Berkovic, Z. Kotler, R. Mazor, V. Khodorkovsky, Eur. J. Org. Chem. 2001, 2001, 2671 – 2687. [18] N. Bloembergen, Selected Topics in Quantum Electronics, IEEE J. Sel. Top. Quantum Eletron. 2000, 6, 876 – 880. [19] a) J. F. Ward, Rev. Mod. Phys. 1965, 37, 1 – 18; b) J. L. Oudar, J. Chem. Phys. 1977, 67, 446 – 457. [20] J. L. Oudar, H. Le Person, Opt. Commun. 1975, 15, 258 – 262.
ChemPhysChem 2014, 15, 2693 – 2700
2699
CHEMPHYSCHEM ARTICLES [21] S. M. LeCours, H.-W. Guan, S. G. DiMagno, C. H. Wang, M. J. Therien, J. Am. Chem. Soc. 1996, 118, 1497 – 1503. [22] J. Zyss, J. Non-Cryst. Solids 1982, 47, 211 – 225. [23] M. C. Zerner, W. M. F. Fabian, R. Dworczak, D. W. Kieslinger, G. Kroner, H. Junek, M. E. Lippitsch, Int. J. Quantum Chem. 2000, 79, 253 – 266. [24] T. Yoshimura, Phys. Rev. B 1989, 40, 6292 – 6298. [25] a) T. Yoshimura, Mol. Cryst. Liq. Cryst. Inc. Nonlinear Opt. 1990, 182, 43 – 50; b) T. Yoshimura, Appl. Phys. Lett. 1989, 55, 534 – 536. [26] J. Zyss, I. Ledoux, Chem. Rev. 1994, 94, 77 – 105. [27] S. R. Marder, C. B. Gorman, B. G. Tiemann, J. W. Perry, G. Bourhill, K. Mansour, Science 1993, 261, 186 – 189. [28] F. Meyers, S. R. Marder, B. M. Pierce, J. L. Bredas, J. Am. Chem. Soc. 1994, 116, 10703 – 10714. [29] I. D. L. Albert, T. J. Marks, M. A. Ratner, J. Am. Chem. Soc. 1997, 119, 6575 – 6582. [30] I. D. L. Albert, T. J. Marks, M. A. Ratner, Chem. Mater. 1998, 10, 753 – 762. [31] R. A. Huijts, G. L. J. Hesselink, Chem. Phys. Lett. 1989, 156, 209 – 212. [32] P. R. Varanasi, A. K.-Y. Jen, J. Chandrasekhar, I. N. N. Namboothiri, A. Rathna, J. Am. Chem. Soc. 1996, 118, 12443 – 12448. [33] I. F. Perepichka, D. F. Perepichka, M. R. Bryce, A. Chesney, A. F. Popov, V. Khodorkovsky, G. Meshulam, Z. Kotler, Synth. Met. 1999, 102, 1558 – 1559. [34] W. Frank, M. Klaus, Chem. Eur. J. 2010, 16, 9366 – 9373. [35] A. V. Kulinich, A. A. Isschenko, Russ. Chem. Rev. 2009, 78, 141 – 164. [36] H. Burckstummer, N. M. Kronenberg, K. Meerholz, F. Wrthner, Org. Lett. 2010, 12, 3666 – 3669. [37] K. D. Belfield, D. J. Hagan, E. W. Van Stryland, K. J. Schafer, R. A. Negres, Org. Lett. 1999, 1, 1575 – 1578. [38] C. Barsu, R. Cheaib, S. Chambert, Y. Queneau, O. Maury, D. Cottet, H. Wege, J. Douady, Y. Bretonniere, C. Andraud, Org. Biomol. Chem. 2010, 8, 142 – 150. [39] K. D. Belfield, K. J. Schafer, W. Mourad, B. A. Reinhardt, J. Org. Chem. 2000, 65, 4475 – 4481. [40] S. Niu, G. Ulrich, P. Retailleau, R. Ziessel, Tetrahedron Lett. 2011, 52, 4848 – 4853. [41] S. Niu, G. Ulrich, P. Retailleau, R. Ziessel, Org. Lett. 2011, 13, 4996 – 4999. [42] W. Qin, M. Baruah, M. Van der Auweraer, F. C. De Schryver, N. Boens, J. Phys. Chem. A 2005, 109, 7371 – 7384.
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemphyschem.org [43] W. Qin, T. Rohand, M. Baruah, A. Stefan, M. van der Auweraer, W. Dehaen, N. Boens, Chem. Phys. Lett. 2006, 420, 562 – 568. [44] J. Karolin, L. B. A. Johansson, L. Strandberg, T. Ny, J. Am. Chem. Soc. 1994, 116, 7801 – 7806. [45] J. Olmsted, J. Phys. Chem. 1979, 83, 2581 – 2584. [46] A. P. de Silva, H. Q. N. Gunaratne, C. P. McCoy, J. Am. Chem. Soc. 1997, 119, 7891 – 7892. [47] A. Coskun, E. Deniz, E. U. Akkaya, Org. Lett. 2005, 7, 5187 – 5189. [48] M. A. Fox, Photochem. Photobiol. 1990, 52, 617 – 627. [49] D. Rehm, A. Weller, Ber. Bunsen-Ges. 1969, 73, 834 – 839. [50] D. Rehm, A. Weller, Isr. J. Chem. 1970, 8, 259 – 271. [51] R. Ziessel, L. Bonardi, P. Retailleau, G. Ulrich, J. Org. Chem. 2006, 71, 3093 – 3102. [52] I. Maltey, J. A. Delaire, K. Nakatani, P. Wang, X. Shi, S. Wu, Adv. Mater. Opt. Elec. 1996, 6, 233 – 238. [53] L. Ferrighi, L. Frediani, C. Cappelli, P. Salek, H. Agren, T. Helgaker, K. Ruud, Chem. Phys. Lett. 2006, 425, 267 – 272. [54] F. Wrthner, F. Effenberger, R. Wortmann, P. Krmer, Chem. Phys. 1993, 173, 305 – 314. [55] E. Cariati, D. Dragonetti, E. Lucenti, F. Nisic, S. Righetto, R. E. Tordin, Chem. Commun. 2014, 50, 1608 – 1610. [56] J. P. J. Stewart, Computational Chemistry, Colorado Springs, CO, 2008, http://OpenMOPAC.net. [57] J. P. J. Stewart, J. Mol. Model. 2007, 13, 1173 – 1213. [58] A. Klamt, G. Schrmann, J. Chem. Soc. Perkin Trans. 2 1993, 799 – 805. [59] H. Iikura, T. Tsuneda, T. Yanai, K. Hirao, J. Chem. Phys. 2001, 115, 3540 – 3544. [60] DALTON, Release 2.0, a molecular electronic structure program (with CAM – B3LYP functional enabled) 2005, http://www.kjemi.uio.no/software/dalton/dalton.html. [61] C. H. Wang, J. Chem. Phys. 2000, 112, 1917 – 1924. [62] ZINDO version of the Cerius2 suite of programs by MSI under license from Accelrys.
Received: March 12, 2014 Revised: May 12, 2014 Published online on June 20, 2014
ChemPhysChem 2014, 15, 2693 – 2700
2700
