Tuning the Direction of Intramolecular Charge Transfer and the Nature

Article pubs.acs.org/JPCA

Tuning the Direction of Intramolecular Charge Transfer and the Nature of the Fluorescent State in a T‑Shaped Molecular Dyad Abdellah Felouat,† Anthony D’Aléo,*,† Azzam Charaf-Eddin,†,‡ Denis Jacquemin,‡,§ Boris Le Guennic,*,∥ Eunsun Kim,⊥ Kwang Jin Lee,⊥ Jae Heun Woo,⊥,# Jean-Charles Ribierre,⊥ Jeong Weon Wu,*,⊥ and Frédéric Fages*,† †

Aix Marseille Université, CNRS, CINaM UMR 7325, Campus de Luminy, Case 913, 13288 Marseille, France Laboratoire CEISAM, UMR CNRS 6230, Université de Nantes, 2 Rue de la Houssinière, BP 92208, 44322 Nantes Cedex 3, France § Institut Universitaire de France, 103 Boulevard Saint-Michel, 75005 Paris Cedex 05, France ∥ Institut des Sciences Chimiques de Rennes, UMR 6226 CNRS-Université de Rennes 1, 263 Avenue du Général Leclerc, 35042 Rennes Cedex, France ⊥ Department of Physics, CNRS-Ewha International Research Center, Ewha Womans University, Seoul, South Korea ‡

S Supporting Information *

ABSTRACT: Controlling photoinduced intramolecular charge transfer at the molecular scale is key to the development of molecular devices for nanooptoelectronics. Here, we describe the design, synthesis, electronic characterization, and photophysical properties of two electron donor− acceptor molecular systems that consist of tolane and BF2containing curcuminoid chromophoric subunits connected in a T-shaped arrangement. The two π-conjugated segments intersect at the electron acceptor dioxaborine core. From steady-state electronic absorption and fluorescence emission, we find that the photophysics of the dialkylamino-substituted analogue is governed by the occurrence of two closely lying excited states. From DFT calculations, we show that excitation in either of these two states results in a distinct shift of the electron density, whether it occurs along the curcuminoid or tolane moiety. Femtosecond transient absorption spectroscopy confirmed these findings. As a consequence, the nature of the emitting state and the photophysical properties are strongly dependent on solvent polarity. Moreover, these characteristics can also be switched by protonation or complexation at the nitrogen atom of the amino group. These features set new approaches toward the construction of a three-terminal molecular system in which the lateral branch would transduce a change of electronic state and ultimately control charge transport in a molecular-scale device.

■

INTRODUCTION

T-shaped molecules are 2D systems formally related to cruciforms. The construction of such systems is governed by the choice of the three-terminal junction (hereafter called Tjunction) unit that assembles the branches because it determines both the overall molecular geometry and the extent of electronic communication between the branches. The benzene ring has been considered, because it allows access to a variety of structures with pseudo-3-fold symmetry.32−34 For example, electron donor (D) and acceptor (A) arrays were proposed as ultrafast molecular logic gates.33 Efforts were also directed to devise those connectors that are sensitive to external stimuli. A benzimidazole junction has led to T-shaped πconjugated molecules with acid-responsive photophysical properties.35 A spiropyran moiety served as a photoactive Tjunction for directional electron transfer.36 It was demonstrated that 2,2′-3,3″-terthiophene derivatives undergo photochemi-

Controlling delocalization of π electrons along a linear (macro)molecular backbone is key to the development of conjugated organic materials for applications in electronics, photovoltaics, and photonics.1−7 Increasing dimensionality with the use of branched scaffolds enables the generation of functional structures that combine multiple conjugation pathways. Among recent examples, cruciforms, conjugated dendrimers, and two-dimensional (2D) polymers have emerged as very promising platforms for the generation of materials for sensing or semiconducting applications.8−24 Even more challenging is the investigation of those structurally related but more complex molecular systems in which the direction of electron delocalization can be switched upon application of a physical, chemical, or photo-/electrochemical external stimulus, because they have unique optoelectronic properties and potential applications in molecular electronics.8−15,25−29 For example, redox,30,31 coordination,10 and surface self-assembly9 processes have been employed to control the electronic structures of cruciforms. © XXXX American Chemical Society

Received: April 17, 2015 Revised: May 27, 2015

A

DOI: 10.1021/acs.jpca.5b03699 J. Phys. Chem. A XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry A

distribution of excited rotamers leading to differently relaxed CT species with identical orientations of their excited-state dipole moment. As a result, a direct filiation of relaxation between CT states was evidenced. The possibility of controlling the nature of the emitting species in such systems can offer exciting prospects of applications in molecular electronics or sensing.35,43,44,52−58 We report herein the first case of a T-shaped molecule that displays a dual emission arising from two distinct ICT states that are (i) produced by independent light excitation and (ii) involve the perpendicular D1−A and D2−A molecular branches. We also show that the change in nature of the solvent provides a clear-cut means to fine-tune the competition between the two ICT processes. Moreover, we demonstrate that the basic amino group in T-2 can act as a transducer for efficient molecular side switching of optical properties. To this end, we used steady-state and time-resolved absorption and fluorescence emission spectroscopy. Quantum chemical (QC) calculations based on density functional theory (DFT) and its time-dependent counterpart (TD-DFT) were performed in order to reach better insights into the electronic structure of the ground and first excited states of these molecules and consequently into their absorption and emission properties. As described in a previous publication,59 B-curcuminoids strongly absorb visible light and emit fluorescence with quantum yields as high as 60%. The introduction of a side aryl ring at the central carbon atom (called meso) of the acetylacetonate subunit has been shown not to quench the Bcurcuminoid excited state. This feature is exploited here to further extend the π conjugation from the meso position along a direction quasi-orthogonal to that of the B-curcuminoid backbone. We opted for the introduction of a diphenylacetylene (DPA or tolane) segment in T-1 and T-2. T-1 will serve as a reference compound in the photophysical investigations. With the di-n-octylamino (shortened as amino) substituent, the tolane subunit in T-2 is related to that of p-N,Ndimethylamino-p′-cyanodiphenylacetylene (DACN-DPA).60−62

cally reversible cyclization and cycloreversion reactions and achieve rerouting of the π-conjugation system.37 The two compounds investigated in this study, T-1 and T-2 (Chart 1), are difluoroboron complexes of curcuminoid Chart 1. Structure of T-1 and T-2

derivatives (called thereafter B-curcuminoid) in which the central (1,3-diketonato)boron difluoride (dioxaborine) ring serves as a T-junction. The latter is a strong electron acceptor moiety38,39 and supports two types of donor groups in T-2, D1 (anisole) and D2 (dialkylaniline). We thus envisioned that T-2 could enable investigating the interplay between two photoinduced intramolecular charge transfer (ICT) processes arising from the quasi-orthogonal D1−A and D2−A fragments. Photoinduced ICT is a key phenomenon in advanced optoelectronic materials.11 For example, D−A π-conjugated rods are of considerable importance for nanoscience and nanotechnology because they represent elementary functional building blocks showing interesting optical and electronic properties.42 In that connection, molecular cruciforms8−15,43,44 and T-shaped systems35,45 having ICT transitions were investigated as fluorescence polarity indicators or metal ion sensors. Moreover, dual fluorescence was observed in several D−A systems owing to the interplay between locally excited (LE) and charge transfer (CT) states.46,47 In some cases, a more complex photophysics resulted from the occurrence of two CT states.48−51 This effect was due to the existence of a

■

RESULTS AND DISCUSSION Synthesis. The synthetic route toward compounds T-1 and T-2 is outlined in Scheme 1. We used here the meso-p-

Scheme 1. Synthesis of T-1 and T-2a

LEgend: (a) B2O3, ethyl acetate, 60 °C; (b) anisaldehyde, tri-n-butylborane, ethyl acetate, 60 °C; (c) n-butylamine, ethyl acetate, 80 °C, then HCl, 60 °C; (d) 3 or 4, CuI, Pd(PPh3)2Cl2, THF, Et3N, 50 °C; (e) BF3·Et2O, DCM, reflux. a

B


Article


Figure 1. Normalized UV−visible absorption spectra (concentration ca. 10−5 M): (a) T-1 (black , ■) and T-2 (red , ●) in DCM; (b) T-1 (bottom) and T-2 (top) in CH (black , ■), DBE (blue , ▲) and EA (red , ●).

ing T-2 from T-1 in DCM are retained in each solvent investigated: i.e., a decrease in molar absorption coefficient and an increase in fwhm (Table S1). We also noticed the presence of a more or less apparent absorbing tail, depending on solvent nature, at the red edge of the low-energy band of T-2. Quantum chemical calculations provided valuable insights to rationalize the electronic absorption properties of T-1 and T-2 (see Computational Details in the Experimental Section). In order to estimate electron delocalization in the molecular orbitals (MOs) of compounds T-1 and T-2, we performed DFT optimization of the ground state geometries using the M06-2X functional. The corresponding MOs are sketched in Figure 2. Theoretical calculations on the parent 1,3,2-dioxaborine compounds acac(BF2) and dbm(BF2) have been reported in the literature.39 They show that the trimethine fragment of the βdiketone unit displays the characteristic frontier MO patterns of an allylic unit. Of note, the LUMO coefficient at the meso carbon atom is 0. The latter feature translates onto the LUMOs of compounds T-1 and T-2, indicating that electronic communication between the tolane and the B-curcuminoid units is interrupted through this position in the LUMO. The HOMO and LUMO of T-1 are delocalized with the main contributions along the B-curcuminoid. For T-2, the situation is different, with the LUMO delocalized on the B-curcuminoid branch while the HOMO resides on the tolane branch. Therefore, one notes the spatial separation of the frontier MOs in T-2. We underline that calculations do not point to any significant orbital overlap at the dioxaborine T-junction in the frontier MOs of T-2, which is likely to stem from the twisted geometry of the dioxaborine−phenyl meso junction. The vanishing electron density at the meso carbon in the LUMO may further limit electronic communication between tolane and dioxaborine units. The first calculated vertical excitation energies with their assignment in terms of MOs contributions are given in Table 1. These data show similar features for T-1 and T-2. The transition band at low energy measured at around 2.44 eV in DCM is calculated at ca. 2.90 eV for both compounds. The MO assignment clearly shows that these transitions are centered on the B-curcuminoid part of the molecules and are described mainly as HOMO → LUMO and HOMO-1 → LUMO excitations for T-1 and T-2, respectively. They may be viewed

iodophenyl acetylacetone precursor 6 obtained from acetylacetone (acacH).63 Compound 5 is conveniently obtained from 6 and anisaldehyde,59 which is subsequently subjected to standard Sonogashira cross coupling with the appropriate acetylene derivatives to introduce the orthogonal tolane unit, yielding the free ligands 1 and 2. The boron difluoride complexes T-1 and T-2 were prepared simply by reacting the ligands with boron trifluoride etherate in dichloromethane (DCM). All new compounds were fully characterized by NMR spectroscopy (see the Supporting Information) and highresolution mass spectrometry. Compounds T-1 and T-2 could be obtained with the high purity grade requested for spectroscopic measurements and were found to be soluble in organic solvents. No aggregation was detected at the concentrations used in this study. All compounds were found chemically and photochemically stable in solution, and particularly, no detectable BF2 decoordination was observed in all solvents used here. Electronic Absorption Properties. The electronic absorption spectra of compounds T-1 and T-2 in dichloromethane (DCM) are given in Figure 1a. Both compounds feature two main well-separated transition bands. From literature data,60−62 the high-energy band, located in the UV part of the spectrum, is assigned to the tolane chromophore. This transition shifts bathochromically when going from T-1 (λmax 290 nm) to T-2 (λmax 360 nm) as a result of the auxochromic effect of the amino group, becoming similar to that of DACN-DPA.20 For T-1, the low-energy band observed in the visible region, around 500 nm, arises from the Bcurcuminoid chromophore, as it features a shape and a molar absorption coefficient (around 60000 M−1cm−1) that both match those recorded for related compounds bearing a mesoaryl group.59 In contrast, the presence of the strong electron donor amino group in T-2 alters both the intensity and the shape of that low-energy band. The molar absorption coefficient becomes significantly smaller (ca. 52000 M−1cm−1) and the full width at half-maximum (fwhm) is larger (3263 cm−1 vs 2895 cm−1 for T-1) in DCM (Table S1 in the Supporting Information). When solvents of increasing polarity are used (Figure 1b and Figures S1 and S2 in the Supporting Information), the low-energy transition band broadens and shifts to the red for both compounds, as observed in previous studies,59 but the spectral characteristics that allow distinguishC


Article


2. The solvatochromism is well reproduced by the calculations with a slightly underestimated blue shift (ca. 10 nm instead of 20 nm) of the absorption band when going from DCM to cyclohexane (CH) (Table S2 in the Supporting Information). The HOMO-2 → LUMO excitation gives rise to a weak Bcurcuminoid electronic transition calculated at 3.7 eV for both T-1 and T-2. The absorption band centered on the tolane moiety at high energy is theoretically described as a mix of excitations with similar contributions. For compound T-1, the experimental wavelength for this transition (290 nm) is well reproduced at 284 nm. A poorer agreement between the calculated and experimental values is obtained for T-2 (312 nm vs 360 nm), but the experimentally observed red shift remains qualitatively reproduced. In comparison to T-1, the calculated data obtained for T-2 point to the occurrence of an extra transition at low energy (3.23 eV, 384 nm) close to that centered on the B-curcuminoid chromophore. That transition dominated by a HOMO to LUMO contribution involves MOs that are spatially separated. Such types of transitions have usually been found in cruciforms or other T-shaped molecules.8−15 For T-1, the related transition is traceable to the HOMO-1 → LUMO excitation and its calculated energy is much higher (3.78 eV) than for T-2. These observations led us to scrutinize the low-energy transition profile of T-1 and T-2, and we performed a spectral deconvolution of the spectra using a Gaussian-shaped function (Figure 3). Four bands represent the spectrum of T-1. For the sake of clarity, Figure 3a shows only the three bands of lower energy that are assigned to the main vibronic structure, the fourth band at high energy being attributed to the HOMO-2 → LUMO excitation. In the case of T-2, the curve-fitting method leads to a total of five bands. Omitting here again the representation of the HOMO-2 → LUMO transition, the spectrum in Figure 3b contains the three vibronic contributions, identical with those found in T-1, and a new band assigned to the HOMO → LUMO transition. The latter is broad and is located underneath the curcuminoid band, displaying a maximum at 490 nm in DCM. Deconvolution analyses performed on spectra recorded in other solvents, apolar and polar, lead qualitatively to the same spectral decomposition and show a dependence of the maximum

Figure 2. Selected MOs of compounds T-1 (top) and T-2 (bottom).

as cyanine-type transitions. It is well-known that vertical TDDFT calculations do not quantitatively reproduce the experimental transition energy of the cyanine class,64−66 and this explains the rather large deviation (ca. 0.5 eV) observed between experimental and calculated data. Moreover, calculations confirm the nonparticipation of the amino moiety in this transition, resulting in similar transition energies for T-1 and T-

Table 1. Experimental and Calculated λ Values (nm) and Transition Energies (eV) for Compounds T-1 and T-2 in DCMa T-1

T-2 theor

exptl λ/nm (eV)

λ/nm (eV)

f

508 (2.44)

425 (2.92)

2.10

290 (4.27)

theor assignt H → L (89%)

333 (3.72)

0.02

H-2 → L (86%)

328 (3.78) 284 (4.36)

0.08 1.27

H-1 → L (85%) H-1 → L+1 (42%) H → L+1 (41%)

274 (4.52)

0.03

H-3 → L (75%) H-8 → L (9%)

exptl λ/nm (eV)

λ/nm (eV)

f

assignt

506 (2.45)

428 (2.89)

1.90

490 (2.53)

384 (3.23)

0.25

336 (3.69)

0.20

H-1 → L (77%) H → L (17%) H → L (75%) H-1 → L (16%) H-2 → L (73%) H → L+1 (8%) H → L+2 (8%)

312 (3.97)

1.50

287 (4.32)

0.05

360 (3.44)

H → L+2 (40%) H → L+1 (35%) H-2 → L (18%) H-3 → L (67%) H-5 → L (13%)

a

The theoretical values are evaluated at the PCM(DCM)-TD-M06-2X/6-31G(d)//PCM(DCM)-M06-2X/6-31G(d) level of approximation. Vertical transition assignments are based on MO contributions (only MO pairs with percentages larger than 8% are given). D


Article


Figure 3. Example of Gaussian deconvolution of the low-energy part of the electronic absorption spectra of (a) T-1 and (b) T-2 in ethyl acetate (EA).

Table 2. Spectroscopic Data and Photophysical Properties of Compounds T-1 and T-2 in Solvents of Different Polarity at Room Temperaturea T-1

T-2

solvent

λabs

λem

ΔνST

Φf

τf

pentane MCH CH Tol DBE DEE Chlo THF EA DCM BN acetone AN

483 487 488 505 494 494 509 504 501 508 498 504 504

516 520 522 537 525 524 558 545 545 564 558 556 566

1324 1303 1335 1180 1195 1159 1725 1493 1611 1955 2160 1856 2173

0.14 0.18 0.19 0.28 0.25 0.29 0.39 0.35 0.35 0.42 0.35 0.378 0.38

2.5 ns) component at 510 and 650 nm, respectively (Figure S6 in the Supporting Information). When the emission was monitored at the red edge of the emission band (>690 nm), only the long component was measured. No rise time could be observed with our experimental setup operating at the nanosecond time resolution. In the Lippert−Mataga plot (Figure 5),69,70 we used the Δν values obtained in polar solvents and those corresponding to the maximum wavelength of the high-energy, short-lived emission measured in TOL and DBE. We obtained a slope of ca. 6500 cm−1, which, within experimental error, is close to the value obtained for T-1. Then, a plot of the Stokes shifts measured for T-2 in apolar alkane solvents together with those of the long-lived, low-energy emission in TOL and DBE gave a straight line with a much larger value of the slope (ca. 21000 cm−1). These correlations show that fluorescence emission occurs from the B-curcuminoid-centered state (LUMO → HOMO-1 transition), a strong CT state, or both, depending on solvent polarity. The CT state is characterized by a long lifetime and a low kf value relative to those of B-curcuminoids and DACN-DPA,59−62 which correlates with the weak oscillator strength calculated for the radiative transition between the

Plotting the Stokes shift (Δν) versus the solvent polarity parameter (Δf ′) according to the Lippert−Mataga formalism69,70 (Figure 5) yields a linear relationship with a slope of ca.

Figure 5. Lippert−Mataga plots of 1 (blue ▲), high-energy emission of 2 vs the curcuminoid absorption maxima (black ■), and low-energy emission of 2 vs the curcuminoid absorption maxima (red ●) with their linear regression.

6000 cm−1 for the whole range of solvents under investigation. This value is characteristic of the B-curcuminoid chromophore.59 The fluorescence emission quantum yields (Φf) of T-1 are found between 0.2 and 0.4, and the fluorescence decays fitted with a monoexponential function give short excited-state lifetimes (2 ns). Scheme 2 sketches the excited-state dynamics in T-2. In DCM, both compounds T-1 and T-2 behave as T-1 in CH. They exhibit positive signals at 807 nm showing excitedstate absorption with ultrafast decay (

Tuning the Direction of Intramolecular Charge Transfer and the Nature

Recommend Documents