Two-Photon Absorption Properties of Proquinoidal D-A-D and A-D-A Quadrupolar Chromophores

Abstract

We report the synthesis, one- and two-photon absorption spectroscopy, fluorescence, and electrochemical properties of a series of quadrupolar molecules that feature proquinoidal π-aromatic acceptors. These quadrupolar molecules possess either donor-acceptor-donor (D–A–D) or acceptor-donor-acceptor (A–D–A) electronic motifs, and feature 4-N,N-dihexylaminophenyl, 4-dodecyloxyphenyl, 4-(N,N-dihexylamino)benzo[c][1,2,5]thiadiazolyl or 2,5-dioctyloxyphenyl electron donor moieties and benzo[c][1,2,5]thiadiazole (BTD) or 6,7-bis(3’,7’-dimethyloctyl)[1,2,5]thiadiazolo[3,4-g]quinoxaline (TDQ) electron acceptor units. These conjugated structures are highly emissive in nonpolar solvents and exhibit large spectral red-shifts of their respective lowest energy absorption bands relative to analogous reference compounds that incorporate phenylene components in place of BTD and TDQ moieties. BTD-based D-A-D and A-D-A chromophores exhibit increasing fluorescence emission red-shifts, and a concomitant decrease of the fluorescence quantum yield (Φ_f) with increasing solvent polarity; these data indicate that electronic excitation augments benzothiadiazole electron density via an internal charge transfer mechanism. The BTD- and TDQ-containing structures exhibit blue-shifted two-photon absorption (TPA) spectra relative to their corresponding one-photon absorption (OPA) spectra, and display high TPA cross-sections (>100 GM) within these spectral windows. D-A-D and A-D-A structures that feature more extensive conjugation within this series of compounds exhibit larger TPA cross-sections consistent with computational simulation. Factors governing TPA properties of these quadrupolar chromophores are discussed within the context of a three-state model.

Introduction

Two-photon absorption (TPA) is the basis for an increasing number of optical and electrooptical applications that include data storage,^1–3 3-dimensional microfabrication,^4–6 biological imaging,^7–12 optical limiting,^13–16 and photodynamic therapy.^17–21 The potential of these technologies, in turn, continues to drive development of materials with improved two-photon absorption (TPA) properties. For example, TPA-enhanced materials boost absorption efficiency at low excitation intensity, and therefore help minimize damage due to laser irradiation. Therefore, an improved understanding of chromophoric structural factors that may be manipulated to improve TPA will undoubtedly impact a variety of technologies that depend upon this nonlinear optical (NLO) property.

To date, a number of design strategies for materials with large TPA cross-sections (8) have been proposed and demonstrated. d-π-d, D-A-D, and A-D-A structural motifs, where D, π, and A refer, respectively, to electron donor, π-conjugated bridge, and electron acceptor units, have been shown to exhibit exceptionally large TPA cross-sections for many D, π, and A moieties.^22–34 Quadrupolar structures, in particular, have experimentally²⁸ and theoretically^35,36 demonstrated enhanced TPA cross sections that are approximately an order of magnitude greater than those of dipolar analogues. In addition, quadrupolar compounds have higher fluorescence quantum yields relative to corresponding dipolar compounds, most likely due to cancellation of D-A dipoles which otherwise would tend to induce charge transfer quenching of fluorescence; such augmented fluorescence quantum yields are useful, for example, in imaging technologies that rely on TPA.^22,24,28 Corresponding octopolar chromophores,^37–42 as well multi-branched conjugated systems that include dendrimers,^43–48 can also provide materials with high TPA cross-sections.

Incorporation of proquinoid structures, such as benzothiadiazole and thiadiazoloquinoxaline, into the main chain of π-conjugated oligomers and polymers is an established means to diminish optical and potentiometric band gaps.^49–58 π-Conjugated materials that feature benzothiadiazole or thiadiazoloquinoxaline units have been exploited as both emissive and charge transport functional elements for organic light-emitting diodes,^{54,55,59–63} and as key components of photovoltaic cells^60,64–73 and field effect transistors.^59,74–76 Since proquinoidal units function as electron accepting moieties in structures having extended conjugation, D–A–D and A–D–A type chromophores that take advantage of benzothiadiazole or thiadiazoloquinoxaline building blocks might be expected to exhibit enhanced TPA properties.

In this paper, we report the synthesis, one- and two-photon absorption spectroscopy, fluorescence, and electrochemical properties of a series of quadrupolar molecules that feature benzothiadiazole or thiadiazoloquinoxaline π-aromatic acceptors, and compute the one- and two-photon absorption spectra for these species, to provide further insight into the origins of the OPA and TPA properties that derive from the presence of the proquinoidal units.

Experimental Section

Materials

Manipulations were carried out under argon previously passed through an O₂ scrubbing tower (Schweitzerhall R3–11 catalyst) and a drying tower (Linde 3-Å molecular sieves) unless otherwise stated. Air sensitive solids were handled in a Braun 150-M glove box. Standard Schlenk techniques were employed to manipulate air-sensitive solutions. Tetrahydrofuran (THF) was distilled from K/4-benzoylbiphenyl under N₂. Triethylamine, MeOH, and CH₂Cl₂ were distilled from CaH₂ under N₂. Pyridine was also dried over CaH₂ and distilled under reduced pressure. The catalysts Pd(PPh₃)₄, tris(dibenzylideneacetone)dipalladium(0) (Pd₂dba₃), Pd(PhCN)₂Cl₂, triphenylarsine (AsPh₃) and P(t-Bu)₃ (10 wt% solution in hexanes) were purchased from Strem Chemicals and used as received. 4-Bromobenzo[c][1,2,5]thiadiazole,⁷⁷ 4,7-dibromobenzo[c][1,2,5]thiadiazole,⁷⁷ 5,6-diamino-4,7-dibromobenzo[c][1,2,5]thiadiazole,⁷⁸ 4,7-bis[(trimethylsilyl)ethynyl]benzo[c][1,2,5]thiadiazole,⁵⁶ and 4,7-diiodobenzo[c][1,2,5]thiadiazole⁷⁹ were prepared by literature methods. All NMR solvents were used as received. The supporting electrolyte used in the electrochemical experiments, tetra-n-butylammonium hexafluorophosphate, was recrystallized twice from ethanol and dried under vacuum at 70 °C overnight prior to use. All the other chemicals were used as received.

Chemical shifts for ¹H NMR spectra are relative to the tetramethylsilane (TMS) signal in deuterated solvent (TMS, δ = 0.00 ppm). All J values are reported in Hertz. Flash and size exclusion column chromatographic separations were performed on the bench top, using respectively silica gel (EM Science, 230–400 mesh) and Bio-Rad Bio-Beads SX-1 as media. CI mass spectra were acquired at the Mass Spectrometry Center at the University of Pennsylvania. MALDI-TOF mass spectroscopic data were obtained with a PerSeptive Biosystems Voyager-DE RP; samples for these experiments were prepared as micromolar solutions in THF; dithranol (Aldrich) was utilized as the matrix.

Instrumentation

¹H NMR spectra were recorded on 360 MHz DMX-360, 300 MHz DMX-300, or 250 MHz AC-250 Brüker spectrometers. Electronic absorption spectra were recorded on a Shimadzu PharmaSpec UV1700 spectrophotometer. Fluorescence spectra were obtained with a Spex Fluorolog-3 spectrophotometer (Jobin Yvon Inc, Edison, NJ) that utilized a PMT detector. These fluorescence spectra were corrected using the spectral output of a calibrated light source supplied by the National Bureau of Standards. Fluorescence quantum yields were measured in argon-purged solutions at room temperature. Quinine sulfate in 1.0 N H₂SO₄ (Ф_f = 0.55)⁸⁰ and rhodamine 101 inner salt in EtOH (Ф_f = 1.00)⁸¹ were used as standards. Time-correlated single-photon counting (TCSPC) experiments that measured fluorescence lifetimes were carried out using an Edinburgh Analytical Instruments FL/FS 900 spectrometer, using either a nanosecond flash lamp operating under an atmosphere of H₂ gas (0.50–0.55 bar, 1.2 ns full width at half maximum (fwhm), 40 kHz repetition rate) or a blue (450 ± 15 nm) light-emitting diode (Picoquant PLS 450/PDL 800-B) triggered at a frequency of 100 kHz by a Berkeley Nucleonics 555 pulse generator, as excitation sources.^82,83 This diode produced light pulses of 450 ps fwhm with a power of 15 μW. TCSPC data were analyzed by iterative convolution of the fluorescence decay profile with the instrument response function using software provided by Edinburgh Instruments.

Cyclic voltammetric measurements were carried out on an EG&G Princeton Applied Research model 273A Potentiostat/Galvanostat. The electrochemical cell for these experiments utilized a platinum disk working electrode, a platinum wire counter electrode, and a saturated calomel reference electrode (SCE). The reference electrode was separated from the bulk solution by a junction bridge filled with the corresponding solvent/supporting electrolyte solution. The ferrocene/ferrocenium redox couple was utilized as an internal potentiometric standard.

Two-photon Absorption Spectra

The background theory and a detailed experimental description of our TPA spectra measurements were described previously.^84,85 The samples were dissolved in toluene at a concentration of 3.8–10.5 × 10⁻⁵ M. The sample cells were degassed by freeze-pump-thaw methods (3 cycles) and filled with argon gas. Throughout our experiments, we used a 100 µM solution of fluorescein in H₂O (pH ~ 13) as a standard calibration sample. We assume herein that the one-photon and two-photon fluorescence quantum efficiencies of fluorescein are equal, and that the fluorescence quantum efficiency is constant over the spectral range of the experiment.

Computational Studies

We used time-dependent density-functional theory (TDDFT) within the collective electronic oscillator (CEO) framework recently described by Tretiak et al.⁸⁶ to calculate one- and two-photon absorption spectra for this series of quadrupolar proquinoid compounds. In our implementation of the CEO framework, frequency-dependent polarizabilities are described by functions that can be obtained from linear-response theory, including corresponding functional derivatives up to third order in the driving field.⁸⁷ Briefly, the procedure for computing frequency-dependent polarizabilities using TDDFT involved: (1) computing the ground-state density matrix, (2) obtaining a set of transition densities and frequencies (for coupled harmonic oscillators), (3) calculating anharmonicity tensors, and (4) summing over states for the desired response. This method was found to be efficient computationally and accurate for calculating nonlinear polarizabilities of large D-A substituted conjugated molecules.^86,88,89 The method typically performs better than semi-empirical methods (such as ZINDO and PM3 methods) and other common ab initio approaches (such as CI singles and time-dependent Hartree-Fock methods) for describing molecular excited states and electronic excitation energies.^90,91

All electronic structure and non-linear response calculations were carried out using the GAUSSIAN 98⁹² and CEO programs,⁹³ while geometry optimization was performed using the Hartree-Fock (HF) method with a 6–31G basis set. In order to maximize computational efficiency, the solubilizing substituents of these proquinoidal chromophores (N,N-dihexylamino, dodecyloxy, octyloxy and 3,7-dimethyloctyl groups) were replaced by structurally simpler analogues (N,N-dimethylamino, methoxy, and methyl groups, respectively)). All molecular geometries were optimized beginning with planar structures with symmetry constraints. The excitation energies, dipole matrices, and Coulomb exchange-interaction matrices were calculated using the TDDFT method in GAUSSIAN 98 with a 6–31G basis set and the B3LYP functional. The energies and corresponding transition densities of the first 10 singlet excited states were calculated. We did not include more excited states in the calculation because it has been reported that increasing the number of excited states from 6 to 30 changes the resulting cross-section by only about 10%.^86,89 The CEO program was used to compute first-, second-, and third-order responses based on electronic structure analysis from GAUSSIAN 98.⁹³ These simulations neglected solvation effects. An empirical damping factor Γ = 0.1 eV (obtained from experimental studies of phenylene-vinylene derivatives) was used.^22,24

Synthesis and Characterization

A full account of the synthesis and characterization of all new compounds, complete with tabulated spectral data, is provided in the Supporting Information (SI).

Results and Discussion

Synthesis

Structures of D-A-D and A-D-A chromophores examined in this study are shown in Figure 1; these molecules were synthesized by Pd-mediated cross-coupling reactions involving substituted aromatic donor and proquinoidal acceptor building blocks (Scheme 1; see Supporting Information for details). 4,7-Diethynylbenzo[c][1,2,5]thiadiazole (BTD) and 4,9-diethynyl-6,7-bis(3’,7’-dimethyloctyl)[1,2,5]thiadiazolo[3,4-g]quinoxaline (TDQ) were selected as proquinodal spacer units. For the thiadiazoloquinoxaline-containing structures, note that 3,7-dimethyloctyl groups serve as 6- and 7-TDQ-ring substituents, which facilitate solubility of these large planar molecules.

Figure 1.

Structures of D-A-D and A-D-A chromophores, along with their respective abbreviations; note that t and d signify connectivities utilizing respectively C-C triple and double bonds.

Scheme 1.

Synthetic routes to D-A-D and A-D-A chromophores and their precursors.

Proquinoidal BTD and TDQ units have been used to drive enhanced conjugation in semiconducting oligomers and polymers, and as electron acceptor moieites;^49,51,56,94 these units play a similar role in the Figure 1 chromophores. For these D-A-D and A-D-A structures, aromatic units are connected by either ethyne or ethene linkers, through respective Sonogashira or Heck coupling reactions. These simple, efficient coupling protocols enabled facile synthesis of the Figure 1 compounds (Supporting Information).

One-Photon Absorption Spectra

The linear (one-photon) electronic absorption spectra of the proquinoidal D-A-D and A-D-A molecules, measured in solvents of varying dielectric constant, are displayed in Figure 2. Fluorescence and comparative one-photon absorptive extinction coefficient data acquired in toluene solvent are compiled in Table 1. As highlighted in the Figure 2 spectral data, these chromophores feature excellent solubility in a wide range of solvents, and underscore the chromophoric impact of the proquinoidal components. Notice, for example, that the lowest absorption maxima of (DHAt)₂BTD and (DHAd)₂BTD are respectively 111 nm (5,681 cm⁻¹) and 131 nm (5,929 cm⁻¹) red-shifted relative to their corresponding reference compounds featuring phenylene spacers (1,4-bis[(4’-N,N-dimethylaminophenyl)ethyn-1’-yl]benzene⁹⁵ and 1,4-bis[(4’-N,N-di-n-butylaminophenyl)ethen-1’-yl]benzene).²⁴ (ROPht)₂TDQ and (ROPhd)₂TDQ express similar magnitude spectroscopic red-shifts (10,392 cm⁻¹ (toluene) and ~9,500 cm⁻¹ (CH₂Cl₂)) of their respective lowest energy absorption maxima, relative to analogous 1,4-bis[(4′-methoxyphenyl)ethyn-1′-yl]benzene⁹⁵ and 1,4-bis[(4′-n-butoxyphenyl)ethen-1′-ylbenzene⁹⁶ spectroscopic benchmarks.

Figure 2.

Electronic absorption spectra of: (A) (DHAt)₂BTD, (B) (DHAd)₂BTD, (C) (ROPht)₂TDQ, (D) (ROPhd)₂TDQ, (E) (DHA-BTDt)₂BTD (F) (BTDt)₂ROPh, (G) (DHAtPht)₂ROPh, (H) (DHAtBTDt)₂ROPh, and (I) (DHAtROPht)₂BTD, recorded in cyclohexane, toluene, tetrahydrofuran (THF), CH₂Cl₂, acetone, and dimethylformamide (DMF) solvents.

Table 1.

The lowest energy linear absorption band maxima, extinction coefficients, and the fluorescence maxima of proquinoidal (DHAt)₂BTD, (DHAd)₂BTD, (ROPht)₂TDQ, (ROPhd)₂TDQ, (DHA-BTDt)₂BTD, (BTDt)₂ROPh, (DHAtPht)₂ROPh, (DHAtBTDt)₂ROPh, and (DHAtROPht)₂BTD chromophores determined in toluene solvent.

	Linear Absorption			Fluorescence Emission
	λabs(nm)	ν(cm−1)	ε(×104 M−1cm−1)	λem(nm)	ν(cm−1)
(DHAt)2BTD	501	19,960	3.53	598	16,722
(DHAd)2BTD	540	18,519	3.70	651	15,361
(ROPht)2TDQ	562	17,794	2.04	655	15,267
(ROPhd)2TDQ	626	15,974	2.01	727	13,755
(DHA-BTDt)2BTD	549	18,215	4.61	612	16,340
(BTDt)2ROPh	424	23,585	2.77	501	19,960
(DHAtPht)2ROPh	401	24,938	9.41	445	22,472
(DHAtBTDt)2ROPh	515	19,417	6.96	590	16,949
(DHAtROPht)2BTD	494	20,243	5.64	587	17,036

For systems with identical donor and acceptor units, compounds featuring ethene-based connectivity manifest more substantial low energy absorption band spectroscopic red-shifts relative to corresponding compounds in which ethynyl moieties link aromatic units [λ_max((DHAt)₂BTD, toluene) = 501 nm < λ_max((DHAd)₂BTD, toluene) = 540 nm; λ_max((ROPht)₂TDQ, toluene) = 562 nm < λ_max((ROPhd)₂TDQ, toluene) = 626 nm)]. While all of these D-A-D and A-D-A chromophores evince only modest solvatochromism of their respective lowest energy absorption band maximum (Figure 2), (DHAt)₂BTD, (DHAd)₂BTD, (DHA-BTDt)₂BTD, (DHAtPht)₂ROPh, (DHAtBTDt)₂ROPh, and (DHAtROPht)₂BTD exhibit systematic red-shifts in their absorption edges with increasing solvent polarity. Note that (ROPht)₂TDQ, (ROPhd)₂TDQ, and (BTDt)₂ROPh do not display similar red-shifts of their long wavelength absorption band edges with increasing solvent polarity, consistent with the diminished electron-releasing character of alkoxyphenyl (ROPh) relative to the dialkylaminophenyl group.

Fluorescence Spectra

In contrast to the linear absorption spectroscopic data, the fluorescence spectra of these proquinoidal D-A-D and A-D-A compounds show a clear dependence on solvent dielectric constant (Figure 3). As the solvent polarity increases, the fluorescence maxima red-shift and the corresponding fluorescence quantum yield decreases (Table 2). Identical spectroscopic trends have been observed in other D-A-D and A-D-A systems, and have been attributed to an S₀➔S₁ electronic transition having internal charge transfer (ICT) character.^39,44,97 For the chromophores featuring dialkylamino electron releasing groups, substantial solvent-dependent Stokes shifts are observed [(DHAt)₂BTD (2,176 (cyclohexane) − 4,918 (DMF) cm⁻¹ ); (DHAd)₂BTD (2,410 (cyclohexane) − 4,260 (DMF) cm⁻¹ ); (DHA-B Dt)₂BD (1,279 (cyclohexane) − 2,985 (DMF) cm⁻¹); (DHAtPht)₂ROPh (2,158 (cyclohexane) − 7,209 (DMF) cm⁻¹); (DHAtBTDt)₂ROPh (1,123 (cyclohexane) − 5,243 (acetone) cm⁻¹ ); (DHAtROPht)₂BTD (2,316 (cyclohexane) − 5,688 (CH₂Cl₂) cm⁻¹)]; at the other extreme lie the (ROPht)₂TDQ and (ROPhd)₂TDQ chromophores, which evince only modest solvent-dependent Stokes shifts (ROPht)₂TDQ (2,046 (cyclohexane) − 3,068 (acetone) cm⁻¹); (ROPhd)₂TDQ (1,969 (cyclohexane) − 2,468 (CH₂Cl₂) cm⁻¹)], suggesting a diminished degree of ICT character to their respective S₀➔S₁ and S₁➔S₀ electronic transitions.

Figure 3.

Fluorescence spectra of: (A) (DHAt)₂BTD, (B) (DHAd)₂BTD, (C) (ROPht)₂TDQ, (D) (ROPhd)₂TDQ, (E) (DHA-BTDt)₂BTD, (F) (BTDt)₂ROPh, (G) (DHAtPht)₂ROPh, (H) (DHAtBTDt)₂ROPh, and (I) (DHAtROPht)₂BTD, recorded in cyclohexane, toluene, THF, CH₂Cl₂, acetone, and DMF solvents. Sharp peaks evident at ~500 and 800 nm derive from scattered light from the excitation source.

Table 2.

Fluorescence Quantum Yields of Proquinoidal D-A-D and A-D-A Chromophores Determined in Cyclohexane, Toluene, THF, CH₂Cl₂, Acetone, and DMF Solvents.

	Cyclohexane	Toluene	THF	CH2Cl2	Acetone	DMF
(DHAt)2BTD	0.87	0.84	0.31	0.25	0.011	0.0031
(DHAd)2BTD	0.75	0.42	0.13	0.037	0.031	0.0083
(ROPht)2TDQ	0.48	0.42	0.27	0.27	0.25
(ROPhd)2TDQ	0.054	0.034	0.017	0.013
(DHA-BTDt)2BTD	0.13	0.088	0.11	0.095	0.014	0.0034
(BTDt)2ROPh	0.59	0.66	0.48	0.37	0.15
(DHAtPht)2ROPh	0.86	0.89	0.88	0.78	0.35	0.13
(DHAtBTDt)2ROPh	0.65	0.80	0.14	0.080	0.0012	<0.001
(DHAtROPht)2BTD	0.59	0.56	0.0097	0.0013	<0.001	<0.001

Solvent-polarity-dependent fluorescence spectra can be interpreted in terms of the Lippert equation (eq. (1)), which expresses the magnitude of the Stokes shift in terms of changes in the molecular dipole moment that occur concomitantly with electronic excitation.

$${\tilde{\nu }}_a-{\tilde{\nu }}_f=\left[\frac{\epsilon -1}{2\epsilon +1}-\frac{n^2-1}{2n^2+1}\right]\frac{2{({\mu }_E-{\mu }_G)}^2}{hc{a}^3}+\text{constant}$$ (1)

Here ṽ_a and ṽ_f are the respective peak absorption and emission energies expressed in wavenumbers, ε is the dielectric constant, n is the refractive index of the solvent, ε_E and ε_G correspond respectively to the magnitudes of the excited- and ground-state dipole moments, h is Planck’s constant, c is the speed of light, and a is the radius of the solute spherical cavity.

The bracketed term in eq. 1 is often referred to as the orientational polarizability (Δf; eq. 2).

$$\mathrm{\Delta }f=\frac{\epsilon -1}{2\epsilon +1}-\frac{n^2-1}{2n^2+1}$$ (2)

Plots of the magnitude of the Stokes shift (ṽ_a – ṽ_f) versus the orientational polarizability are displayed in Figure 4. This analysis highlights the fact that the slopes of the Lippert plots for the TDQ-containing chromophores are smaller than those of the D-A-D and A-D-A structures that are based on the BTD unit. This result likely derives from (i) the reduced electron-releasing character of alkoxyphenyl relative to the dialkylaminophenyl group, which leads to smaller magnitude ε_E –ε_G values, and (ii) the fact that TDQ possesses a larger solute spherical cavity radius relative to BTD. Likewise, it is interesting that the observed Lippert plot slopes for the longer (DHAtBTDt)₂ROPh and (DHAtROPht)₂BTD molecules are similar in magnitude to the slope for the shorter (BTDt)₂ROPh chromophore; this effect is consistent with the expectation that while (DHAtBTDt)₂ROPh and (DHAtROPht)₂BTD should possess larger (μ_E–μ_G) values relative to that for (BTDt)₂ROPh, the solvent spherical cavity sizes for these species are also correspondingly larger.

Figure 4.

Lippert plots determined for the (DHAt)₂BTD, (DHAd)₂BTD, (ROPht)₂TDQ, (ROPhd)₂TDQ, (DHA-BTDt)₂BTD, (BTDt)₂ROPh, (DHAtPht)₂ROPh, (DHAtBTDt)₂ROPh, and (DHAtROPht)₂BTD chromophores.

Electrochemical properties

Figure 5 shows the potentiometrically determined E_1/2^0/+ and E_1/2^−/0 valuesfor these proquinoidal D-A-D and A-D-A chromophores; these electrochemical data are tabulated in Table 3. The Figure 5 potentiometric data highlight a number of key electronic structural features of these compounds. For example, replacing ethyne by ethene in the spacer structure [(DHAt)₂BTD, (DHAd)₂BTD; (ROPht)₂TDQ, (ROPhd)₂TDQ] lowers the first oxidation and reduction potentials by ~350 and ~150 mV, respectively, in these BTD- and TDQ-based chromophores and therefore leads to a correspondingly smaller potentiometrically determined HOMO–LUMO gap (E_p), mirroring the trend observed for the optical band gaps (E_op) of these species (Figure 2).

Figure 5.

Potentiometrically determined E_1/2^0/+ and E_1/2^−/0 values measured in CH₂Cl₂ solvent for: (A) (DHAt)₂BTD, (B) (DHAd)₂BTD, (C) (ROPht)₂TDQ, (D) (ROPhd)₂TDQ, (E) (DHA-BTDt)₂BTD, (F) (BTDt)₂ROPh, (G) (DHAtPht)₂ROPh, (H) (DHAtBTDt)₂ROPh, and (I) (DHAtROPht)₂BTD. Experimental conditions are described in Table 3. Redox potentials shown are relative to the ferrocene/ferrocenium (Fc/Fc⁺) redox couple, which was used as an internal standard in these experiments.

Table 3.

Comparative Potentiometric Data Acquired for Proquinoidal D-A-D and A-D-A Chromophores.^a

	E1/20/+ (Epa)b	E1/2−/0
(DHAt)2BTD	----- (0.40)	−1.74
(DHAd)2BTD	0.04 (0.12)	−1.92
(ROPht)2TDQ	0.76c (0.80)	−1.28
(ROPhd)2TDQ	----- (0.38)	−1.39
(DHA-BTDt)2BTD	----- (0.36)	−1.70
(BTDt)2ROPh	0.91 (0.98)	−1.84
(DHAtPht)2ROPh	----- (0.43) 0.46d (0.57)	----- −2.54d
(DHAtBTDt)2ROPh	----- (0.46) 0.36e (0.44)	−1.72 −1.66e
(DHAtROPht)2BTD	0.32 (0.41)	−1.66

^aCyclic voltammetric experimental conditions: [chromophore] = ~1 mM; solvent = CH₂Cl₂; [TBAPF6] = 0.10 M; scan rate = 200 mV/s; reference electrode = SCE; working electrode = Pt disk or glassy carbon. All potentials are reported relative to the ferrocene/ferrocenium redox couple, which was used as an internal standard in these experiments.

^bAnodic peak potential

^cScan rate = 400 or 500 mV/s

^din THF

^ein PhCN

The structures of (DHA-BTDt)₂BTD and (DHAt)₂BTD are closely related; (DHA-BTDt)₂BTD features terminal dihexylamino groups directly attached to benzothiadiazole rather than phenyl groups. Figure 5 data suggest that replacement of the terminal phenyl by BTD destabilizes the HOMO level by 0.04 V, while stabilizing the LUMO to a similar degree; these data thus indicate that (DHA-BTDt)₂BTD and (DHAt)₂BTD possess comparable extents of electronic delocalization and ICT character in their respective ground and low energy excited states.

Note that the reduction potential of (BTDt)₂ROPh (−1.84 V vs. Fc/Fc⁺) is shifted cathodically relative to that for 4,7-bis[(trimethylsilyl)ethynyl]benzo[c][1,2,5]thiadiazole (BTD(E-TMS)₂) (−1.66 V). Since the ethyne group typically acts as a weak acceptor, the higher reduction potential of BTD(E-TMS)₂ compared to (BTDt)₂ROPh suggests that the two terminal benzothiadiazole groups in (BTDt)₂ROPh are not strongly conjugated to the central dialkoxyphenyl unit in the LUMO level.

Finally, Figure 5 shows that the reduction potential of (DHAtBTDt)₂ROPh (−1.72 V) is ~ 100 mV stabilized relative to that determined for (BTDt)₂ROPh, as expected due its expanded conjugation. In contrast, note that the oxidation potential of (DHAtBTDt)ROPh is similar to that of (DHAtPht)ROPh, suggesting that the HOMO is localized extensively on the dihexylaminophenyl groups, or only weakly delocalized with the remainder of linear π-conjugated system. Likewise, these potentiometric data indicate that the reduction potential of (DHAtROPht)₂BTD (−1.66 V) resembles that of BTD(E-TMS)₂, suggesting a LUMO localized primarily on the diethynylbenzothiadiazole unit; this trend is evident in the Figure 5 data, which shows clearly that the BTD and TDQ units play a dominant role in determining the potentiometrically determined LUMO levels.

Two-photon absorption spectra

Two-photon absorption cross-sections were determined by measurements of two-photon fluorescence excitation spectra^98,99 with femtosecond laser pulses in the spectral range between 750 and 960 nm. The two-photon induced fluorescence excitation spectra are shown in Figure 6. TPA cross-sections (δ values; expressed in Göppert-Mayer or GM units), where 1 GM = 1×10⁻⁵⁰ cm⁴ s molecule⁻¹ photon⁻¹), and maxima, are given in Table 4.

Figure 6.

Two-photon fluorescence excitation spectra determined in toluene solvent for: (A) (DHAt)₂BTD, (B) (DHAd)₂BTD, (C) (ROPht)₂TDQ, (D) (ROPhd)₂TDQ, (E) (DHA-BTDt)₂BTD, (F) (BTDt)₂ROPh, (G) (DHAtPht)₂ROPh, (H) (DHAtBTDt)₂ROPh, and (I) (DHAtROPht)₂BTD. Systematic error for δ is ±30% for each point.

Table 4.

TPA Data Acquired for Proquinoidal D-A-D and A-D-A Chromophores Along with those for Reference Compounds in Toluene Solvent.

	Egea (eV)	εmaxb (M−1cm−1)	Mgec (D)	TPA maxima (nm)	ħωd (eV)	(Ege–ħω) (eV)	δe(GM)	Mee’f (D)
(DHAt)2BTD	2.47	35,300	8.00	850	1.46	1.01	230	6.9
(DHAd)2BTD	2.30	37,000	8.53	890	1.39	0.91	270	6.5
(ROPht)2TDQ	2.21	20,400	6.39	g	g	g	g	g
(ROPhd)2TDQ	1.98	20,100	6.62	g	g	g	g	g
(DHA-BTDt)2BTD	2.26	46,100	9.19	910	1.36	0.90	100	3.7
(BTDt)2ROPh	2.92	27,700	6.96	780	1.59	1.33	120	6.9
(DHAtPht)2ROPh	3.09	94,100	12.66	760	1.63	1.46	630	9.2
(DHAtBTDt)2ROPh	2.41	69,600	11.41	830	1.49	0.92	330	5.1
(DHAtROPht)2BTD	2.51	56,400	10.30	890	1.39	1.12	390	8.0
Reference Ah	3.19		11.4	712	1.74	1.45	980	12.4
Reference Bi	3.03		10.7	730	1.70	1.33	995	12.1

^aS₀➔S₁ electronic transition.

^bExtinction coefficient.

^cS₀➔S₁ transition dipole moment.

^dExcitation energy.

^eTwo photon absorption cross section.

^fTransition dipole moment between the S₁ state and two-photon allowed higher-lying excited states.

^gThese parameters were not evaluated for TDQ derivatives due to possible mixing of OP A and TPA.

^hThe values were extracted from ref. 102.

ⁱThe values were extracted from ref. 24.

Based on their spectral shapes (Figure 6), the actual TPA peaks of (DHA-BTDt)BTD, (BTDt)ROPh, (DHAtPht)ROPh, and (DHAtROPht)BTD may lie outside of the spectral window analyzed in the present study; otherwise, the peak TPA cross-sections for all of the present compounds are comparable to those of similar benzothiadiazole derivatives reported recently.^100,101 To facilitate comparison with the corresponding OPA spectra, TPA spectra are plotted as a function of λ_ex/2 (Figure 7). All of the compounds showed TPA maxima higher in energy than the corresponding OPA maxima.

Figure 7.

Normalized one-photon absorption and two-photon excitation spectra of: (A) (DHAt)₂BTD, (B) (DHAd)₂BTD, (C) (ROPht)₂TDQ, (D) (ROPhd)₂TDQ, (E) (DHA-BTDt)₂BTD, (F) (BTDt)₂ROPh, (G) (DHAtPht)₂ROPh, (H) (DHAtBTDt)₂ROPh, and (I) (DHAtROPht)₂BTD. The solid line represents the one-photon absorption spectrum. The two-photon fluorescence excitation spectral data are displayed as solid circles (the dotted line serves only as a guide to the eye). Except for (ROPhd)₂TDQ (D), the two-photon fluorescence excitation spectral data are plotted as a function of the total transition energy (twice the photon energy). Note that the (ROPhd)₂TDQ fluorescence excitation data is plotted as a function of the photon energy to highlight the extent to which the TPA magnitude rises near the low energy absorption edge.

Linearity tests of power-squared fluorescence dependence were performed at varying laser excitation wavelengths for selected samples and are shown in Figure 8; the corresponding magnitudes of these slopes are given in Table 5. While the large TPA cross-sections below 800 nm for (ROPht)₂TDQ may arise from residual OPA, it is important to note that the slope (1.82) of the two-photon excited fluorescence intensity as a function of the logarithm of the laser power at 780 nm indicates significant TPA. Note that the available spectral window for (ROPhd)2TDQ (Figure 7D) was mostly dominated by OPA; all other compounds showed pure TPA within the spectral window interrogated.

Figure 8.

The dependence of the two-photon excited fluorescence intensity, plotted logarithmically as a function of the logarithm of the laser power P, for chromophores (A) (DHAt)₂BTD, (B) (DHAd)₂BTD, (C) (ROPht)₂TDQ, and (D) (DHAtPht)₂ROPh.

Table 5.

Linearity Tests for Power-Squared Fluorescence Dependence of Proquinoid D-A-D and A-D-A Chromophores in Toluene Solvent.

Compounds	λ(nm)	Slope a,b
(DHAt)2BTD	850	2.06
(DHAd)2BTD	900	1.89
(ROPht)2TDQ	780	1.82
(BTDt)2ROPh	810	1.65
(DHAtPht)2ROPh	830	1.86
(DHAtBTDt)2ROPh	870	1.77
(DHAtROPht)2BTD	900	2.07

^aValues displayed are the slopes of the linear regression fits to logarithmic plots of fluorescence intensity versus incident excitation laser intensity.

^bEstimated uncertainty of the slope is ±0.1.

The TPA cross-section δ is related to the imaginary part of the second-order hyperpolarizability γ through eq. (3):

$$\delta =\frac{4{\pi }^2\hslash {\omega }^2}{n^2{c}^2}{L}^4\text{Im}\gamma \langle -\omega ;\omega ,-\omega ,\omega \rangle$$ (3)

where ħ is Planck’s constant divided by 2π, ω is the excitation frequency, n is the refractive index of the medium, c is the speed of light, L is the local field factor (L = (n²+2)/3), and Im indicates the imaginary part of the complex quantity <γ>, the orientational average of γ.

$$\text{Im}\gamma =\frac{4}{5}\times \text{Im}\left[\begin{array}{c}\frac{M_{ge}^2\times \mathrm{\Delta }{\mu }_{ge}^2}{{(E_{ge}-\hslash \omega -i{\mathrm{\Gamma }}_{ge})}^2(E_{ge}-2\hslash \omega -i{\mathrm{\Gamma }}_{ge})}\hfill \\ +{\displaystyle \sum _{e\text{'}}\frac{M_{ge}^2\times M_{ee\text{'}}^2}{{(E_{ge}-\hslash \omega -i{\mathrm{\Gamma }}_{ge})}^2(E_{ge}-2\hslash \omega -i{\mathrm{\Gamma }}_{ge\text{'}})}}\hfill \\ -\frac{M_{ge}^4}{{(E_{ge}-\hslash \omega -i{\mathrm{\Gamma }}_{ge})}^2(E_{ge}-\hslash \omega -i{\mathrm{\Gamma }}_{ge})}\hfill \end{array}\right]$$ (4)

Here M_ge is the transition dipole moment between the ground state (S₀) and the first excited state (S₁), Δμ_ge is the difference in dipole moments between S₀ and S₁, E_ge is the transition energy between S₀ and S₁, ħω is the excitation energy, Г is a damping factor, M_ee’ is the transition dipole moment between S₁ and two-photon allowed higher excited states.

The first term in eq. (4) is the dipolar term, which is absent for quadrupolar compounds because Δμ_ge vanishes. The third term is related to one-photon resonance and has no contribution to the TPA. Therefore, eq. (4) is dominated by the second term for quadrupolar compounds. Despite cancellation of the dipolar term, TPA cross sections of quadrupolar structures have been demonstrated to be higher than those of dipolar analogues; note that the M_ee, term similarly vanishes for corresponding dipolar chromophores.²⁸ The three-state TPA cross-section for quadrupolar compounds with electronic states g,e, and e′ is

$$\delta \approx \frac{16{\pi }^2{(\hslash \omega )}^2{L}^4}{5n^2{c}^2\hslash }\frac{M_{ge}^2{M}_{ee\text{'}}^2}{{(E_{ge}-\hslash \omega )}^2{\mathrm{\Gamma }}_{ge\text{'}}}.$$ (5)

The transition moments M_ge are obtained from the area under the lowest energy absorption band. The energy difference (E_ge – ħω) is a detuning factor. Γ_ge’ was assumed to be 0.1 eV in all cases.^22,24,28 M_ee’ was derived using eq. (5), along with the other experimental parameters. Each parameter of eq. (5) is tabulated in Table 4.

As expected from eq. (5), the TPA cross-section of quadrupolar compounds can be enhanced (within a three-state ansatz) by (i) increasing the transition dipole moment between the ground state and the first excited state, (ii) increasing the transition dipole moment between the first and higher excited states, and (iii) decreasing the detuning factor. Since δ is proportional to the square of these quantities, the magnitude of the TPA response can be very sensitive to small changes in electronic structure.

The M_ge values of (DHAt)₂BTD (8.00 D) and (DHAd)₂BTD (8.53 D) are slightly smaller than those of the corresponding reference compounds (11.4 and 10.7 D, respectively, for 1,4-bis[(4′-N,N-bis(4″-methoxyphenyl)aminophenyl)ethyn-1′-yl]-2,3-dimethylbenzene (Reference A in Table 4)¹⁰² and 1,4-bis[(4′-N,N-dibutylaminophenyl)ethen-1′ -yl] benzene (Reference B in Table 4),²⁴ in which the BTD moiety is replaced by phenyl. As the radiative rate constant (k_r) is proportional to the square of M_ge,¹⁰³ these proquinoidal D-A-D and A-D-A molecules manifest diminished k_r values with respect to these phenylene-based reference compounds (Table 6).^102,104 Radiative rate constants were determined from the fluorescence quantum yields and fluorescence lifetimes using:

$${Ф}_f=\frac{k_r}{k_r+k_{nr}}$$ (6)

$${\tau }_f=\frac{1}{k_r+k_{nr}}$$ (7)

In these equations, Φ_f is the fluorescence quantum yield, k_r and k_nr are radiative and nonradiative rate constants, respectively, and τ_f is the fluorescence lifetime. Radiative rate constants for these chromophores were also estimated from Strickler-Berg relation and their corresponding absorption and fluorescence spectra.¹⁰³ The experimental and calculated radiative rate constants matched well for these BTD- and TDQ-based chromophores, and are smaller than those determined previously for their corresponding phenylene-based reference compounds (see Table 6 references). Note as well that the detuning factors of these proquinoidal chromophores are smaller than those determined previously for their respective corresponding reference compounds (Table 4), in which the BTD and TDQ units are replaced by simple aryl structures. Since smaller detuning factors increase TPA cross-sections, measured decreases of TPA cross-sections in these proquinoidal chromophores relative to their respective phenylene analogues would therefore be expected to derive from smaller M_ee’ values for these species, consistent with the data in Table 4. Note that (DHAtBTDt)₂ROPh and (DHAtROPht)₂BTD, which have extended π-conjugated structures relative to (DHAt)₂BTD, evince higher TPA cross-sections, suggesting that increased conjugation length is an important factor to enhance TPA, consistent with other reports.^24,32,105

Table 6.

Fluorescence Lifetime Data for Proquinoidal D-A-D and A-D-A Chromophores Determined in Toluene Solvent.

	τf (ns)a	kr (× 107 s−1)	knr (× 107 s−1)	kr (calc) (× 107 s−1)
(DHAt)2BTD	4.40	19.1	3.63	19.1
(DHAd)2BTD	4.71	8.92	12.3	17.6
(ROPht)2TDQ	6.63	6.33	8.75	9.68
(ROPhd)2TDQ	3.55	0.958	27.2	7.94
(DHA-BTDt)2BTD	0.78	11.3	117	23.3
(BTDt)2ROPh	3.13	21.1	10.8	23.5
(DHAtPht)2ROPh	0.73	122	15.0	109
(DHAtBTDt)2ROPh	3.10	25.8	6.46	40.0
(DHAtROPht)2BTD	2.80	20.0	15.7	33.0
Reference A’b	0.74	104	31.1
Reference B’c	1.1	84.5	6.4

^aThe uncertainty of the lifetime measurements was ±0.05 ns.

^bThe values were obtained from ref. 102.

^cThe values were obtained from ref. 104.

Theoretical Analysis

In order to provide further insight into the origins of the OPA and TPA properties of these proquinoidal D-A-D and A-D-A quadrupolar chromophores, the one- and two-photon absorption spectra for these species were computed. Each structure was simplified and optimized using HF/6–31G methods for structures with π-systems constrained to remain planar; time-dependent density-functional theory (TDDFT) with the collective electronic oscillator (CEO) framework⁸⁶ was used to compute the one- and two-photon absorption spectra for these structures (see Experimental Section). Data obtained from these calculations are tabulated in Table 7.

Table 7.

Calculated OPA and TPA Band Maxima, and TPA Cross Sections (δ), of Proquinoidal D-A-D and A-D-A Chromophores, Along with Corresponding Experimental Data Obtained in Toluene Solvent.^a

	OPA(calc)	TPA(calc)	δ(calc)	OPA(expt)	TPA(expt)	δ(expt)
	λ(nm)	λ(nm)	GM	λ(nm)	λ(nm)	GM
(DHAt)2BTD	599	969	504	501	850	230
(DHAd)2BTD	633	990	464	540	890	270
(ROPht)2TDQ	679	966	487	562	b	b
(ROPhd)2TDQ	772	1,033	598	626	b	b
(DHA-BTDt)2BTD	608	981	652	549	910	100
(BTDt)2ROPh	523	992	114	424	780	120
(DHAtPht)2ROPh	430	783	1,308	401	760	630
(DHAtBTDt)2ROPh	633	1,093	1,725	515	830	330
(DHAtROPht)2BTD	659	1,181	1,200	494	890	390

^aComputational values were determined for vacuum.

^bThese parameters were not evaluated for TDQ derivatives due to possible mixing of OPA and TPA.

For OPA, these TDDFT calculations provided qualitative agreement with experimental excitation frequencies and transition intensities. Note that the calculated frequencies of the lowest energy excitations are red-shifted by ~100 – 200 nm (about 0.2– 0.6 eV) relative to that determined by experiment for all of the molecules studied. Such transition energy discrepancies between theory and experiment have been noted previously,⁸⁶ and has been ascribed to several sources; these include: 1) solute-solvent interactions, not accounted for computationally, that lead to geometrical distortions and shortening of the effective conjugation length or localization of charge, and 2) the fact that the TDDFT approach overestimates the extent of exciton delocalization.^106,107

For TPA spectra, the simulations revealed that compounds with increased conjugation length (e.g., (DHAtPht)₂ROPh, (DHAtBTDt)₂ROPh, and (DHAtROPht)₂BTD) relative to (DHAt)₂BTD exhibit stronger TPA, consistent with experiment. Note that these calculated cross-sections are about 2 – 5 times larger than the experimental values. Such overestimation of calculated TPA cross-sections relative to those determined experimentally is a common artifact of this computational approach and can be ascribed to several factors. Foremost, the conventional exchange-correlation functionals employed in these simulations often fail to effectively describe charge transfer.^89,108,109 Further, the molecular structures were optimized using symmetry constraints described above; this simplification neglects the rotational degrees of freedom between the donor and acceptor units of these chromophores and access to other lower symmetry species in the ensemble). To test this hypothesis, we calculated TPA spectra for nonplanar (DHAt)₂BTD structures. In this set of simulations, the dihedral angles between the phenylene and BTD ring systems were varied, and TPA spectra were calculated for multiple combinations of twist angles (data not shown). These studies found that the calculated TPA values were strongly influenced by the planarity of the structure, and that the absorption cross-section decreased dramatically with the increasing twist-angle magnitude; the calculated minimum TPA cross-section was found to be less than one fifth of the maximum cross-sectional value determined for the planar structure.

The agreement between calculation and experiment was closest for (DHAt)₂BTD, (DHAd)₂BTD, and (DHAtPht)₂ROPh, which evinced a factor of two difference between the measured and computed TPA cross-sections, and computed red-shifts of the lowest absorption maxima within 100 nm of the respective experimental values). The discrepancy between measured and computed values of δ and the TPA band maximum was larger for the more conjugated ((DHAtPht)₂ROPh, (DHAtBTDt)₂ROPh, and (DHAtROPht)₂BTD) chromophores; this result is consistent with the expectation that these species likely access a range of structures having varying degrees of conjugation in solution. The calculated δ values for (DHAtPht)₂ROPh, (DHAtBTDt)₂ROPh, and (DHAtROPht)₂BTD) exceed those computed for (DHAt)₂BTD, consistent both with experiment and the expectation that increasing conjugation length in these proquinoidal quadrupolar chromophores is an important contributor to enhance TPA cross-section.

Conclusion

π-Conjugated proquinoidal quadrupolar chromophores that feature either donor-acceptor-donor (D–A–D) or acceptor-donor-acceptor (A–D–A) electronic motifs and utilize benzo[c][1,2,5]thiadiazole (BTD) or 6,7-bis(3′,7′-dimethyloctyl)[1,2,5]thiadiazolo[3,4-g]quinoxaline (TDQ) electron acceptor units were designed and synthesized. The lowest energy absorption bands of these proquinoidal compounds are red-shifted ~100 nm relative to the transitions of corresponding reference chromophores that incorporate phenylene components in place of BTD and TDQ moieties. These chromophores exhibit large fluorescence spectral red-shifts and concomitant decreases of fluorescence quantum yield with increasing solvent polarity, indicating that these species possess relaxed singlet excited states having significant internal charge transfer (ICT) character. These proquinoidal species manifest blue-shifted two-photon absorption (TPA) spectra relative to their corresponding one-photon absorption (OPA) spectrum, and display high TPA cross-sections (>100 GM) within these spectral windows. D-A-D and A-D-A structures within this series that feature more extensive conjugation exhibit larger TPA cross-sections; theoretical analysis of these spectral properties are consistent with this trend. Two proquinoidal quadrupolar chromophores, (DHAtBTDt)₂ROPh and (DHAtROPht)₂BTD, showed TPA cross-sections over 150 GM in the spectral region between 750 and 960 nm. This optical excitation window is easily achieved with a mode-locked femtosecond titanium-sapphire laser, which is most frequently used for two-photon-excited fluorescence imaging. Proquinoidal D-A-D and A-D-A chromophores that feature BTD components manifest high fluorescence quantum yields (0.6 < Φ_f <0.9) in environments with low solvent polarity. This property makes these species attractive dopants in nanoscale matrices, such as polymersomes,^110–112 as the combination of large TPA cross-section and substantial fluorescence emission provides added imaging utility; as the nonlinear properties of dipolar and quadrupolar compounds are governed by different parameters, such D-A-D and A-D-A chromophores compliment the orthogonal imaging utility of D-A structures, that include applications such as second harmonic generation imaging.^113–115