Excited-state dynamics of a molecular dyad with two orthogonally-oriented fluorophores

Romain Letrun a, Bernhard Lang a, Oleksandr Yushchenko a, Roland Wilcken b, Denis Svechkarev c, Dmytro Kolodieznyi§ d, Eberhard Riedle b and Eric Vauthey *a
aDepartment of Physical Chemistry, University of Geneva, 30 Quai Ernest-Ansermet, CH-1211 Geneva 4, Switzerland. E-mail: eric.vauthey@unige.ch
bLehrstuhl für BioMolekulare Optik, Ludwig-Maximilians-Universität München, Oettingenstr. 67, 80538 München, Germany
cDepartment of Pharmaceutical Sciences, University of Nebraska Medical Center, Omaha, Nebraska 68198-6858, USA
dV.N. Karazin Kharkiv National University, 4 Svobody Square, Kharkiv 61022, Ukraine

Received 22nd August 2018 , Accepted 20th November 2018

First published on 22nd November 2018

The excited-state dynamics of a T-shaped bichromophoric molecule, consisting of two strong fluorophores, diphenyloxazole and diphenylpyrazoline, directly linked in an orthogonal geometry, was investigated. Despite the weak coupling ensured by this geometry and confirmed by the electronic absorption spectra, this dyad exhibits only weak fluorescence in both apolar and polar solvents, with fluorescence lifetimes ranging from 200 ps in CHX to 10 ps in ACN. Ultrafast spectroscopic measurements reveal that the fluorescence quenching in polar solvents is due to the population of a charge-separated state. In non-polar solvents, this process is energetically not feasible, and a quenching due to an efficient intersystem crossing (ISC) to the triplet manifold is proposed, based on quantum-chemical calculations. This process occurs via the spin–orbit charge-transfer (SOCT) ISC mechanism, which is enabled by the charge-transfer character acquired by the S1 state of the dyad upon structural relaxation and by the orthogonal arrangement of the molecular orbitals involved in the transition. The same mechanism is proposed to explain why the recombination of the charge-separated state is faster in medium than in highly polar solvents, as well as to account for the fast decay of the lowest triplet state to the ground state.

1 Introduction

Over the past years, substantial efforts have been invested in the design and synthesis of organic multichromophoric molecular systems as well as in the investigation of their excited-state properties.1–14 Potential applications include sensing,15–19 molecular electronics,20–25 white-light emitters,26–29 as well as light harvesting for photovoltaics and artificial photosynthesis.30–34 The properties of these polyads strongly depend on the coupling between the chromophoric subunits: in the weak coupling limit, each chromophore retains its specific properties, whereas in the strong coupling limit, the polyad should be considered a new chromophore with totally different properties. The most evident approach to control coupling is by tuning the interchromophoric distance. In this case, weak coupling is achieved by inserting a long-enough bridge between the subunits. However, rigid bridges are usually conjugated molecular rods that are not totally inert and, thus, enhance the interactions between the chromophores.35–37 An alternative approach to minimise coupling, while keeping short distance, is to resort to an orthogonal arrangement of the chromophores that disables conjugation.38–42

The 1,3,5-triphenyl-2-pyrazoline (PyP, Chart 1) dye is a promising platform to prepare such a T-shaped dyad. Indeed, the strong fluorescence of PyP arises from the 1,3-diphenyl-2-pyrazoline (Py) unit.43–45 Therefore, the phenyl group in C-5 position, which is almost orthogonal to the Py fragment, can be replaced by a larger conjugated system to realise a bichromophoric molecule.

image file: c8cp05356a-c1.tif
Chart 1 Structure of the PyOx dyad and of the pyrazoline (Py, PyP) and oxazole (Ox) subunits.

We report here on our investigation of the photophysics and excited-state dynamics of a pyrazoline-based T-shaped dyad (PyOx) with the phenyl ring in C-5 position shared with the 2,5-diphenyloxazole chromophore (Ox) (Chart 1). The latter is characterised by strong fluorescence and is a well-known dye used for scintillation. Surprisingly, previous measurements revealed that, although composed of two strong fluorophores, PyOx exhibits only weak fluorescence independently of the solvent polarity.46 The properties of this dyad will be compared with those of the individual chromophores, namely PyP, which can be considered as a good proxy of Py, and Ox. Previous investigations on pyrazoline derivatives showed that addition of an electron accepting substituent such as CN or NO2 on the phenyl group in the C-5 position suffices to strongly quench the fluorescence.47–49 This was tentatively explained in terms of intramolecular charge transfer (CT), although no experimental evidence of such process was reported.

The main aim of the present study is to establish the origin of the reduced fluorescence of PyOx. We will show that, in polar solvents, the quenching is due to the population of a non-emissive state with a strong CT character, most probably a charge-separated state. In non-polar solvents, this process seems not operative. However, based on quantum-chemical calculations, the fluorescence quenching of PyOx in apolar solvents is explained by invoking an equilibrium structure of the S1 state that leads to a delocalisation of the excitation on both subunits and confers a CT character to this state. The orthogonal arrangement of the chromophores enhances spin–orbit coupling and favours intersystem-crossing (ISC) from the S1 state to the triplet manifold and back to the ground state. This investigation reveals that, although an orthogonal arrangement of two chromophores leads to a weak coupling as expected, it introduces strong spin–orbit coupling to CT, making ISC a very efficient non-radiative deactivation pathway of the excited state.

2 Experimental

2.1 Samples

2,5-Diphenyloxazole (Ox) was purchased from Sigma-Aldrich and used as received, whereas 1,3,5-triphenyl-2-pyrazoline (PyP) and the dyad PyOx were synthesised as described in the ESI. The solvents, acetone (ACE), acetonitrile (ACN), cyclohexane (CHX), 1,2-dichloroethane (DCE), decahydronaphthalene (decaline, DEC), dimethylformamide (DMF), dimethyl sulfoxide (DMSO), and tetrahydrofuran (THF), were all of spectroscopic grade and used as received.

2.2 Steady-state spectroscopy

Stationary absorption and emission spectra were recorded on a Cary 50 spectrophotometer and a Cary Eclipse fluorometer, respectively. All emission and excitation spectra were corrected for the wavelength-dependent sensitivity of the instrument. The fluorescence quantum yields were determined relative to 1,4-bis(5-phenyloxazole-2-yl)benzene (POPOP) in cyclohexane (Φf = 0.97).50 Stimulated emission spectra were calculated by multiplying the spontaneous fluorescence spectra by λ4.51

2.3 Time-resolved fluorescence

Sub-nanosecond time-resolved fluorescence dynamics were measured using the time-correlated single photon counting (TCSPC) technique with the setup described in ref. 52. The excitation was carried out with a laser diode at 375 nm (LDH-P-C-375, PicoQuant). The pulse duration was 60 ps and the full width at half maximum (FWHM) of the instrument response function (IRF) was about 200 ps. A Glan–Taylor polarizer placed in front of the sample ensured linearly polarised excitation light. The fluorescence was collected at magic angle relative to the polarisation of the excitation pulse and passed through an interference filter with 8 nm bandwidth centred at 450 nm.

Fluorescence dynamics on a shorter time scale were measured with the fluorescence up-conversion (FU) instrument described in detail in ref. 53 and 54. Excitation was performed at 380 nm at a repetition rate of 80 MHz using the frequency doubled output of a Ti:sapphire oscillator (Mai Tai HP, Spectra Physics). The irradiance on the sample was 7–8 μJ cm−2 and the IRF was about 250 fs. The fluorescence was monitored at 450 nm. The polarisation of the pump pulses was at magic angle relative to that of the gate pulses. The samples were placed in a 400 μm thick rotating cell with quartz windows. Their absorbance was below 0.2 at the excitation wavelength. Sample degradation was less than 3% from scan-to-scan.

2.4 Transient absorption (TA)

Two different setups, whose description can be found in ref. 55–58 were used for the electronic TA measurements. With the first instrument, excitation was performed at 385 nm with the frequency-doubled output of a 1 kHz Ti:sapphire amplifier system (Spitfire, Spectra-Physics). The irradiance on the sample was about 0.5 mJ cm−2. A supercontinuum generated in a CaF2 plate was used for probing. The polarisation of the probe pulses was at magic angle relative to that of the pump pulses. The IRF was ∼200 fs. The samples were located in a quartz cuvette of 1 mm path length and were continuously stirred by N2 bubbling. Their absorbance was between 0.1 and 0.2 at the excitation wavelength. No significant degradation of the samples was observed throughout the measurements.

The second TA setup is based on a Ti:sapphire amplifier (CPA 2001, Clark MXR) with 1 kHz repetition rate. Frequency conversion in a noncollinear optical parametric amplifier (NOPA) and for excitation in the UV subsequent frequency doubling is used to produce the needed excitation wavelength.57,58 The excitation pulses were centered around 321–325 nm and 368–370 nm with a pulse length of 35 fs and 45 fs, respectively. The pulse length was measured by direct autocorrelation in the UV.59 The IRF depends on the probe wavelength due to group velocity mismatch. Below 100 fs the signal might be masked by the coherent artifact, especially in the UV region. The pulse energy was 100 nJ at a focal spot of 100 μm FWHM (irradiance 3 mJ cm−2). For the probe beam a supercontinuum ranging from 285 to 720 nm was generated in a CaF2 crystal. It was focused to a 35 μm FWHM spot and used to probe the temporal evolution of the sample. The sample was measured in a 100 μm path length flow cell with 200 μm fused silica windows.60 The sample absorption was kept below 0.3 for the whole probe range. Pump and probe pulse polarization were at magic angle.

2.5 Quantum-chemical calculations

Ground-state gas-phase geometry optimisation of PyP, Ox and PyOx was performed at the density functional level of theory (DFT) using the CAM-B3LYP functional,61 and the 6-31G(d,p) basis set. This functional has been shown to provide a good qualitative description of electronic transitions with a large charge-transfer character.62–64 Electronic transitions as well as geometry optimisation of the lowest singlet excited state were computed with time-dependent DFT (TD-DFT) using the same functional and basis set. All calculations were performed using Gaussian 16.65

3 Results

3.1 Steady-state spectroscopy

The electronic absorption spectrum of PyOx is dominated by overlapping bands covering the whole 270 to 400 nm region (Fig. 1C). Comparison with the absorption spectra of PyP and Ox (Fig. 1A and B) suggests that the long wavelength region of this broad band originates from an electronic transition localised on the Py subunit, whereas absorption at wavelengths shorter than ∼350 nm is due to a transition localised on Ox. Indeed, the absorption of PyOx can be well reproduced by a linear combination of the spectra of PyP and of Ox, with the latter red shifted by 700 cm−1 (Fig. S1, ESI). This composite spectrum is a clear indication of the weak coupling between the chromophoric sub-units despite their connection through a single bond. This can be attributed to the orthogonal T-shape arrangement of the chromophores in the dyad, in agreement with the structure of PyP.66 No significant solvatochromism can be observed when going from CHX to ACN. In the following, the lowest energy transition localised on Py will be called S1 ← S0, whereas that at higher energy localised on Ox will be named S2 ← S0.
image file: c8cp05356a-f1.tif
Fig. 1 Absorption and emission spectra of PyP (A), Ox (B) and PyOx (C) in CHX and ACN. The spikes at 405 and 417 nm in (C) are Raman bands of the solvent (excitation was at 371 nm).

The fluorescence quantum yield of the dyad in CHX is one order magnitude weaker (Φf = 0.04) than that of PyP (Φf = 0.6) and Ox (Φf = 0.74), in agreement with a previous observation.46 The emission spectrum is similar but not identical to that of PyP (Fig. 1). The band maximum is slightly blue shifted (423 vs. 429 nm, 330 cm−1), the band is narrower (FWHM of 3470 vs. 3720 cm−1), especially on the long-wavelength side, and exhibits a shoulder around 445 nm. The fluorescence is further reduced by a factor of ∼2 in ACN, and its spectrum, although similar to that of PyP in ACN, is significantly broader on the long-wavelength side. In both solvents, the fluorescence spectrum of PyOx is independent of the excitation wavelength, and the fluorescence excitation spectrum coincides with the absorption spectrum (Fig. S2, ESI). These results reveal that, although the absorption spectrum of PyOx points to excited states localised on either of the two chromophoric subunits, emission occurs from a single state that seems to be mostly localised on the Py subunit. This suggests very efficient S2 → S1 conversion.

3.2 Time-resolved fluorescence

The fluorescence decay of PyP and Ox occurs on the nanosecond timescale (Table 1). For PyP, the decay at 450 nm could be well reproduced using a single-exponential function with a lifetime ranging between 3.5 and 6 ns depending on the solvent. For Ox in CHX, a fluorescence lifetime of 1.4 ns was reported in literature.67 By contrast, the fluorescence dynamics of the dyad take place on a timescale ranging from tens to hundreds of picoseconds, as could be anticipated from the low fluorescence quantum yield. Fluorescence time profiles recorded at 450 nm by FU in various solvents are shown in Fig. 2. These decays were analysed using the convolution of a Gaussian function accounting for the IRF and one or two exponential functions. The resulting time constants are listed in Table 1.
Table 1 Solvent properties (dielectric constant, εs, viscosity, η)a, solvation times, τs,b time constants, τi, and relative amplitudes (in parentheses) obtained from the analysis of the fluorescence dynamics of PyP and PyOx. The fluorescence lifetime of Ox in cyclohexane amounts to 1.4 ns.67 Limits of error on τi: (±5%)
Solventc ε s η/cP τ s/ps PyP PyOx PyOx
τ/ns τ 1/ps τ 2 /ps
a From ref. 70. b 1/e time, from ref. 68. c CHX: cyclohexane, DEC: decaline; THF: tetrahydrofuran; ACN: acetonitrile; DMF: dimethylformamide; DMSO: dimethyl sulfoxide. d Value assigned to the excited-state lifetime.
CHX 2.0 0.89 3.5 224
DEC 2.2 2.76 3.7 240
THF 7.5 0.46 0.70 3.8 9.7 (0.30) 44 (0.70)
ACN 36.6 0.37 0.15 4.3 1.1 (0.32) 7.7 (0.68)
DMF 38.3 0.79 0.67 5.8 3.6 (0.34) 16 (0.66)
DMSO 47.2 1.99 0.90 6.0 2.9 (0.36) 18 (0.64)

image file: c8cp05356a-f2.tif
Fig. 2 Fluorescence time profiles measured at 450 nm upon 380 nm excitation of PyOx in various solvents.

In non-polar solvents (CHX and DEC), the fluorescence dynamics could be analysed with a single exponential function. On the other hand, two exponential functions were required to correctly reproduce the fluorescence decays measured in polar media. The shorter time constant, τ1 (carrying about 1/3 of the amplitude) scales roughly with the solvation times reported in the literature,68 but is systematically longer (Table 1). A contribution from vibrational relaxation to this early fluorescence dynamics cannot be excluded,69 but can be expected to be minor as excitation is performed close to the onset of the absorption band. The fairly large decay amplitude of the fast component and the fact that emission is detected at a wavelength where the fluorescence spectra in polar and non-polar solvents have similar relative height suggests that another process leading to a decrease of oscillator strength could be operative. The longer time constant, τ2, is attributed to the lifetime of the emitting state. The values match the ones from the TA very well (vide infra). This excited-state lifetime of PyOx exhibits a strong dependence on the solvent polarity, decreasing by one order of magnitude when going from non-polar to polar solvents. This suggests that the small fluorescence quantum yield of the dyad could be associated with a charge-transfer process.

3.3 Transient absorption (TA)

TA measurements were first performed with the individual chromophores, PyP and Ox, in order to identify the spectral signatures of their excited states. Representative transient spectra are depicted in Fig. S3 and S4 (ESI). These data can be well reproduced by global target analysis assuming a series of two or three successive exponential steps with the evolution-associated difference absorption spectra (EADS)71 and the corresponding time constants shown in Fig. 3 and 4.
image file: c8cp05356a-f3.tif
Fig. 3 Evolution-associated difference absorption spectra obtained from a global analysis of the TA data recorded with PyP in CHX (A) and ACN (B) upon 370 nm excitation assuming two successive exponential steps (A → B →). The negative stationary absorption and stimulated emission (SE) spectra are also shown for comparison.

image file: c8cp05356a-f4.tif
Fig. 4 Evolution-associated difference absorption spectra (EADS) obtained from a global analysis of the TA data recorded with Ox in CHX (A) and ACN (B) upon 325 nm excitation assuming two (A → B →) or three successive exponential steps (A → B → C →). The negative stationary absorption and stimulated emission (SE) spectra are also shown for comparison.

The TA spectra measured with PyP in CHX and ACN upon 370 nm excitation exhibit an intense positive band with a maximum at 475 and 438 nm, respectively, that decays on the nanosecond timescale and can be thus assigned to an Sn ← S1 excited-state absorption (ESA) (Fig. 3). A weaker and broader Sn ← S1 ESA band with a maximum above 650 nm is also present. Finally, the negative band below 400 nm can be ascribed to the bleach of the S1 ← S0 absorption. This band is significantly narrower than the stationary absorption spectrum indicating the presence of an ESA overlapping significantly with the blue and red sides of the bleach. The S1 → S0 stimulated emission, expected to have the same intensity as the GSB, is hardly visible as it overlaps with the intense ESA band. It can however be distinguished in CHX as a shallow dent around 420 nm. In ACN, it appears as a weak negative feature around 510 nm. Apart from the decay of all spectral features on a nanosecond timescale, very little spectral dynamics can be observed. These TA data could be well reproduced by global analysis assuming two successive steps (Fig. 3). The second step is associated with a >1 ns lifetime and can be assigned to the decay of the equilibrated S1 state of PyP. In CHX, the first step leads mostly to small changes in the stimulated emission region around 420 nm with a 6.3 ps time constant, that could originate from vibrational/structural relaxation of PyP in the S1 state. In ACN, this step is associated with a blue shift of the main ESA band and a concurrent red shift of the stimulated emission occurring in about 600 fs. This process can be assigned to the equilibration of the S1 state upon solvent relaxation.

The TA spectra measured with Ox in CHX and ACN upon 325 nm excitation are very similar and dominated by a positive band culminating at 537 and 520 nm (Fig. 4 and Fig. S4, ESI). This band decays on the nanosecond time scale and can thus be ascribed to Sn ← S1 ESA. By comparison with the stationary absorption and stimulated emission spectra of Ox, the negative bands below 410 nm can be attributed to the bleach of the S1 ← S0 absorption and to the S1 → S0 stimulated emission. In CHX, very little spectral dynamics can be observed, and the TA data could be reproduced assuming two consecutive steps, with the first step corresponding to a ∼10% rise and narrowing of the ESA band on a 7 ps timescale, that can be attributed to vibrational and/or structural relaxation connected with a slight increase in oscillator strength. The same process is also observed in ACN. However, an additional step with a ∼200 fs time constant had to be added in the global analysis to account for the initial blue shift, rise and narrowing of the ESA band as well as the ∼600 cm−1 red shift of the stimulated emission. Like for PyP in ACN, this first step is assigned to solvent relaxation.

The excited-state dynamics of PyOx were investigated at a few selected excitation wavelengths, preferentially populating either the S1 (368 or 385 nm excitation) or the S2 state (321 nm excitation). Ox does not absorb at 368 nm and longer wavelength, whereas both Py and Ox absorb at 321 nm. However, the absorbance of Ox at 321 nm is 3–4 times higher depending on the solvent (Fig. S1, ESI). The TA results in CHX will be discussed first as the dynamics are simpler.

S2 ← S0 excitation in CHX. The earliest TA spectra recorded upon 321 nm excitation are similar to those recorded with Ox (Fig. 4) and are dominated by a positive ESA band at 550 nm that can be assigned to the S2 state of PyOx (Fig. 5). This band decays in less than 300 fs and is replaced by another band at 475 nm that resembles that measured with PyP (Fig. 3) and can thus be attributed to the S1 state of PyOx. These ultrafast dynamics can be interpreted as S2 → S1 conversion. Whether it should be viewed as an excitation energy transfer from Ox to Py or rather as an internal conversion will be discussed in the next section.
image file: c8cp05356a-f5.tif
Fig. 5 Transient absorption spectra recorded at various time delays after 321 nm excitation of PyOx in CHX (A and B), evolution-associated difference absorption spectra (EADS) obtained from a global analysis of the TA data assuming four successive steps (EADS A to D) with the negative absorption and stimulated emission (SE) spectra of PyOx (C) and transient-absorption profiles at several wavelength (D).

Given that the TA data below 400 nm are masked by the coherent artefact before ∼80 fs and the wavelength dependence of the IRF, the EADS and time constants obtained from the global analysis of the first few hundreds of femtoseconds should only be considered on a semi-quantitative basis. According to this analysis, the S2 → S1 conversion appears in the <100 fs and 150 fs steps (Fig. 5A). To obtain a better estimate of the fastest stage of this process, the data at ≤1 ps in the 545 to 555 nm region, where the duration of the IRF is approximately constant, were analysed with the convolution of a Gaussian function and the sum of two exponential functions. The shortest time constant, corresponding to the S2 → S1 conversion, amounts to 90 fs. A rise with a similar time constant is found in the 470–480 nm region. At later times, the TA dynamics can be well reproduced assuming two consecutive steps with the EADS (C and D) shown in Fig. 5 and with 8.4 and 200 ps time constants.

A significant difference compared to the TA data with PyP and Ox is the 200 ps excited-state lifetime of PyOxvs. >1 ns. Another difference can be observed below 400 nm, where the negative band exhibits a pronounced structure with minima at 308, 323 and 342 nm. These wavelengths coincide with the weak vibronic features observed in the S2 ← S0 absorption band of PyOx (Fig. 1). This pronounced structure is due to the presence of an ESA band overlapping with the GSB. This ESA band originates from the S1 state and its shape is similar to that of the stationary absorption spectrum of PyOx but without vibronic structure (Fig. S5, ESI). The early change in the structure at 310 nm (Fig. 5C) coincides with the S2 → S1 conversion.

S1 ← S0 excitation in CHX. The TA spectra recorded with the dyad upon 368 nm excitation in CHX (Fig. S6, ESI) are very similar to those measured upon S2 ← S0 excitation after ∼1 ps, once the S2 → S1 conversion is complete. The data could be well reproduced globally assuming two successive steps with essentially the same EADS and time constant as those found with the >1 ps dynamics upon 321 nm excitation (Fig. 5).
S2 ← S0 excitation in ACN. The TA dynamics observed before 1 ps upon 321 nm excitation of PyOx are similar to those measured in CHX (Fig. 6). The earliest TA spectra are dominated by the Ox-like Sn ← S2 band at 542 nm, which decays with a 100 fs time constant without any discernible shift. This is accompanied by a parallel rise of the Py-like Sn ← S1 band at 475 nm, which shifts to 468 nm on a ∼200 fs timescale. These dynamics can be assigned to the S2 → S1 conversion and to solvation of the S1 state. In addition, the vibronic structure increases on the blue side due to vibrational relaxation. Compared to the S1 state dynamics in CHX, those in ACN require an additional step to be properly reproduced (Fig. 6). The relaxed S1 state spectrum transforms with a 6 ps time constant to a spectrum with a broad band centred at 458 nm and a weaker band with a maximum above 630 nm. This spectrum is also characterised by a negative band that matches relatively well the stationary absorption spectrum of PyOx. The transformation of the positive band and the changes in the bleach region points to a decay of the S1 state. The resulting spectrum decays with a 100 ps time constant to a very small residual spectrum. From the shape of the bleach, the 100 ps transient corresponds to a state that involves both Py and Ox. As it appears only in the polar ACN and is not fluorescent, it can be reasonably assigned to a charge-separated (CS) state, as discussed in more detail in the next section.
image file: c8cp05356a-f6.tif
Fig. 6 Transient absorption spectra recorded at various time delays after 321 nm excitation of PyOX in ACN (A and B), evolution-associated difference absorption spectra (EADS) obtained from a global analysis of the TA data assuming five successive steps (EADS A to E) with the negative absorption and stimulated emission (SE) spectra of PyOx (C) and TA dynamics at several wavelength (D).
S1 ← S0 excitation in ACN. The TA dynamics in ACN upon 368 nm excitation are again similar to those measured upon 321 nm excitation after completion of the S2 → S1 conversion (Fig. S7, ESI). The TA spectra measured at longer wavelengths point to an absorption maximum of the CS state around 636 nm. Two significant differences compared to 321 nm excitation can be noticed: (i) the structured bleach feature is more pronounced and resembles that in CHX and (ii) the Sn ← S1 band is slightly narrower. These two differences could be explained by assuming an ultrafast population of the CS state from the S2 state or the unrelaxed S1 state upon 321 nm excitation. As the decay of the S2 state is not faster in the polar ACN than in CHX (i.e. 100 fs vs. 90 fs), charge separation from the vibrationally hot S1 state is more plausible.

To better understand the role of solvent polarity on the excited-state dynamics of PyOx, further TA measurements were performed in THF upon S2 ← S0 excitation and in other solvents upon S1 ← S0 excitation. The TA spectra and EADS resulting from global target analysis with a two- or four-step sequential model are shown in Fig. S8–S13 (ESI). The time constants are listed in Table 2, where their interpretation is also given. The population of the CS state is observed in THF and in more polar solvents. On the other hand, the excited-state dynamics in the other non-polar solvent, decaline, are very similar to those in CHX.

Table 2 Time constants obtained from the analysis of the transient absorption data. Limits of error on τi: (±5%)
Compound Solventa λ exc /nm τ IC /ps τ rel /ps τ S1 /ps τ CR /ps
a CHX: cyclohexane, DEC: decaline; THF: tetrahydrofuran; DCE: dichloroethane; ACE: acetone; ACN: acetonitrile; DMF: dimethylformamide; DMSO: dimethyl sulfoxide. b For PyOX the state that is selectively populated is indicated in parentheses. c τ IC: S2 → S1 internal conversion; τrel: solvent/vibrational relaxation; τS1: S1 state decay; τCR: charge recombination.
PyP CHX 370 6.3 >1000
ACN 370 0.6 >1000
Ox CHX 325 7.0 >1000
ACN 325 0.2/7.4 >1000
PyOx CHX 321 (S2) 0.09 8.4 200
368 (S1) 8.5 200
DEC 385 (S1) 8.4 250
THF 321 (S2) 0.10 2.0 40 24
DCE 385 (S1) 1.7 34 66
ACE 385 (S1) 0.6 9.8 76
ACN 321 (S2) 0.10 0.2 6.0 100
368 (S1) 0.3 7.4 123
DMF 385 (S1) 0.8 14 165
DMSO 385 (S1) 1.1 19 345

The lifetimes of the S1 state obtained from the TA data are overall in good agreement with the slow component of the fluorescence decays and exhibit the same solvent dependence (Tables 1 and 2 and Fig. S14, ESI).

4 Discussion

The above results reveal that the reduced fluorescence of PyOx in polar solvents compared to the individual Py and Ox subunits can be explained by the population of a non-emissive state, most probably a CS state. However, this state seems not to be responsible for the low fluorescence quantum yield of PyOx in apolar solvents. The feasibility of a charge separation in the dyad can be first evaluated from the Weller equation.72 According to the redox potentials of the subunits (see ESI for detail), the lowest CS state is that with the hole on Py and the electron on Ox. In polar solvents, this state is predicted to be about 0.1 eV below the S1 state. The other CS state with the hole on Ox and the electron on Py is more than 0.6 eV above the S1 state. The spectrum of the CS state obtained from the analysis of the TA data is consistent with that of the radical cation of a dimer of PyP with a broad absorption from 450 up to ∼650 nm.73 The absorption spectrum of the radical anion of Ox has been reported in a 2-methyltetrahydrofuran glass at 77 K.74 It displays a maximum at 500 nm and decreases towards the blue. It is very similar to the S1 ESA of Ox as expected from the electronic structure. Therefore, charge separation does not lead to a significant change in this spectral region.

Intramolecular electron transfer from Py to Ox is also supported by the quantum chemical calculations. Fig. 7 shows that the frontier molecular orbitals (MOs), calculated at the ground-state geometry of PyOx in the gas phase, are strongly localised, in agreement with the orthogonal arrangement of the sub-units. According to TD-DFT calculations summarised in Table S1 (ESI), the S1 ← S0 and S2 ← S0 transitions are localised on Py (HOMO → LUMO+1) and Ox (HOMO−1 → LUMO), respectively. This result agrees with the absorption spectrum of the dyad and confirms the local character of the first two electronic transitions. The next singlet excited state is associated with a HOMO → LUMO transition and is a CS state with the hole on Py and the electron on Ox, in agreement with the Weller equation. This CS state is predicted to be 0.35 eV above the S1 state. In polar solvents, this state can be expected to be strongly stabilised by solvation and to become the lowest singlet excited state.

image file: c8cp05356a-f7.tif
Fig. 7 Frontier molecular orbitals of PyOx calculated at the ground-state equilibrium geometry. The arrows indicate the orbitals associated with the S1 ← S0 (red) and S2 ← S0 transitions (blue).

Both the energetic consideration and the quantum-chemical calculations support our assignment to charge separation as the origin of the fluorescence quenching of PyOx in polar solvents. In apolar solvents, the CS state is predicted to be higher than the S1 state and, thus this process should not be operative. As polarity increases, charge separation becomes exoergic and shortens the excited-state lifetime of PyOx, in agreement with Marcus theory for electron transfer in the normal region (Fig. S14, ESI).75 The 40–50 ps S1 state lifetime in THF is substantially shorter than in non-polar solvents although no CS state is spectroscopically observed. A possible explanation could be that, in this solvent, the recombination of the CS state is faster than the charge separation itself. A negligible transient population would result that does not lead to an appreciable TA signal. To test this idea, the TA data in THF were analysed globally assuming an A → B → C → target model with the third step (charge recombination) faster than the second one (charge separation). This analysis results in a spectrum associated with C that is compatible with the CS state (Fig. S9D, ESI) and yields time constants for charge separation and recombination of 40 and 24 ps, respectively (Table 2). The first time constant is in good agreement with the 44 ps fluorescence lifetime of PyOx in THF (Table 1).

The lifetime of the CS state varies from 60 ps (or 24 ps if THF is taken into account) to 350 ps depending on the solvent. According to the Weller equation, the driving force for charge recombination (CR) to the ground state is larger than 2.9 eV, and thus, this process should occur in the inverted region. However, the solvent dependence of CR does not follow the behaviour expected according to Marcus theory: CR to the ground state should become faster upon increasing solvent polarity, i.e. upon decreasing the energy gap between the CS and the ground states. This is the opposite of the observation: CR is the fastest in THF and DCE and the slowest in DMSO (Table 2).

The quantum-chemical calculations of PyOx in the gas phase at the ground-state equilibrium geometry predict five triplet states below the S1 state (Table S1, ESI), all of them localised either on Py (T1, T4, T5) or Ox (T2, T3). Due to spin conservation, CR to any of these energetically accessible triplet states is, in principle, forbidden.76 However upon CR, an electron undergoes a change from an Ox (LUMO) to a Py orbital (HOMO). Given the orthogonal arrangement of Py and Ox in the dyad, this also involves a change in angular orbital momentum that should induce sufficiently large spin–orbit coupling to allow CR to a triplet state.77 This process, known as SOCT ISC (for spin–orbit, charge-transfer ISC) was observed for intramolecular CR in twisted donor–acceptor dyads.78–82 Spin–orbit coupling in diradicals has been theoretically shown to increase with increasing orthogonality of the associated MOs and with decreasing distance.77,83,84 Given the T-shape geometry of PyOx and the direct connection of the donor–acceptor sub-units, this SOCT ISC can be expected to be particularly efficient. CR to the triplet state in less than 45 ps has been reported for pyrene linked to dimethylaniline.78 Additionally, efficient ISC takes place in curved polycyclic aromatic systems, such as fullerenes,85–87 carbon nanotubes,88,89 and core-twisted aromatics.90 Although in these cases no CT is involved, the curvature of the MOs suffices to result in a large spin–orbit coupling.91–93

Triplet CR is less exergonic than singlet CR (i.e. CR to the ground state) by at least 2 eV. Consequently, it can be expected to occur in the normal region and to slow down with increasing solvent polarity, as observed here. Moreover, in medium polarity solvents like THF and DCE, where the CS state is higher in energy than in the most polar solvents, a larger number of triplet CR pathways might be operative. In principle, this process should lead to a large population of the triplet states, which, at the geometry of the ground state, should be localised on either Py or Ox, and thus be long lived. However, apart from a very weak residual TA signal in polar solvents, no signal that could be attributed to a triplet state was observed. Moreover, the negative bleach below 400 nm decreases in parallel with the band attributed to the CS state, pointing to a close to full recovery of the ground-state population. In summary, for this triplet CR scenario to be compatible with the TA data, the triplet state has to be short-lived, i.e. to have a lifetime shorter than that of the CS state. Such a fast ISC is totally unexpected for a triplet state localised on either Py or Ox.

The assumption of localised excited states made so far is based on the absorption spectrum of PyOx and on the quantum-chemical calculations at the ground-state geometry. However, the situation might be different for the relaxed excited states. To explore this possibility, the geometry of the lowest singlet excited state of PyOx was optimised at the TD-DFT level of theory. Given the relatively close energy spacing of the excited states and the inherent difficulty of DFT to describe excited states with significant charge-transfer character, no attempt to minimise the geometry of other excited states was made. The results presented here for the lowest singlet excited state should only be considered on a qualitative basis.

According to these calculations, the largest geometry changes concern the mutual orientation of the Py and Ox sub-units (Fig. 8): the angle between the short axis of Py and the long axis of Ox decreases by 20 deg. (Fig. 8 top), whereas the angle between the long axis of Py and the short axis of Ox decreases from close to orthogonal (88 deg.) to 68 deg. (Fig. 8 bottom). These changes are associated with substantial differences in the electronic distribution as testified by the frontier molecular orbitals calculated at the excited-state geometry (Fig. 9). Whereas the HOMO and HOMO−1 remain localised on Py and Ox as in the ground-state geometry (Fig. 7), both LUMO and LUMO+1 are distributed over the whole dyad. According to these calculations, the relaxed S1 state is associated with a HOMO–LUMO configuration. These results imply that, in the Franck–Condon S1 state, excitation is localised on Py and that, upon structural relaxation, the excited electron density is redistributed over the whole dyad, whereas the hole remains localised on Py. In other words, the S1 state acquires significant CT character upon relaxation. Once again, these results should not be considered on a quantitative basis, but they clearly reveal how relatively modest structural changes can have profound effects on the nature of the excited state of PyOx.

image file: c8cp05356a-f8.tif
Fig. 8 Optimised gas-phase geometry of PyOx in the ground and S1 states. The bottom views were obtained by rotating the top views around the vertical axis by 90 deg.

image file: c8cp05356a-f9.tif
Fig. 9 Frontier molecular orbitals of PyOX calculated at the S1 state equilibrium geometry.

A possible experimental indication for the rapid structural relaxation of the S1 state of PyOx is given by the biexponential fluorescence decay in polar solvents (Table 1). The fast component, responsible for the decay of about 30% of the intensity, could originate, at least partially, from the decrease of the oscillator strength associated with the increase of the CT character of the S1 state upon structural relaxation.

Experimental evidence of a delocalised excited state can be also found in the TA spectra recorded in apolar solvents upon S1 ← S0 excitation (Fig. S6, ESI). Indeed, the structured bleach feature below 400 nm originating from Ox is clearly visible upon S1 ← S0 excitation. If the S1 excitation were strictly localised on Py, the bleach should not contain any significant contribution from Ox. Additionally, differences in the fluorescence spectrum of PyOx and PyP (Fig. 1) as well as in the position of the ESA bands are consistent with an excited state that is not entirely localised on Py.

Moreover, the timescale of the S2 → S1 conversion, <100 fs, seems more consistent with an internal conversion than an excitation energy transfer (EET) from Ox to Py and also suggests a delocalisation of the excitation. Given the spectral overlap of Ox emission and Py absorption, an EET time constant of <100 fs would require an electronic coupling about five times as large as that calculated for a Förster-type EET process (see ESI).

The quantum-chemical calculations also predict that the S1 state is stabilised by about 0.45 eV upon structural relaxation and moves below the T3 state (Table S1, ESI). The T2 state is mostly associated with a HOMO−1–LUMO configuration, and the T1 state with a HOMO–LUMO configuration. As HOMO−1 and HOMO are localised on Ox and Py, respectively (Fig. 9), the S1 → T2 ISC involves a hole moving from Py (HOMO) to Ox (HOMO−1) (or an electron from Ox to Py) (Fig. S16, ESI). Given the orthogonal orientation of these subunits, this ISC should be strongly accelerated by the SOCT mechanism. Afterwards, rapid T2 → T1 internal conversion should take place.

Such fast ISC could explain the short fluorescence lifetime of the dyad in non-polar solvents, where charge separation is not operative. However, in these solvents also, no triplet state population was observed in the TA data, and the recovery of the ground-state population takes place within a few hundreds of ps. Like for the above-discussed hypothesis of triplet CR, a short-lived triplet state has to be invoked here as well.

If we assume that the equilibrium geometry of the T1 state is similar to that of S1 state, then T1 → S0 ISC involves an electron going from the delocalised LUMO to the Py-centered HOMO (Fig. 9 and Fig. S16, ESI). Here again, SOCT should make this process very efficient. To account for the experimental data in polar solvents, this ISC should be faster than triplet CR, i.e. take place in less than 100 ps. At the S1 state geometry, T1 is only 1.3 eV above the ground state (Table S1, ESI). Therefore, such a fast ISC is not unrealistic.

Conical intersections have been shown to often open efficient non-radiative decay pathways towards the ground state.94–101 The existence of a S1/S0 conical intersection cannot be ruled out here, but it is not supported by the quantum-chemical calculations that predict only a modest narrowing of the S1–S0 gap of PyOx upon equilibration of the excited state.

From what precedes, the deactivation pathways depicted in Fig. 10 are proposed to account for the decay of the emissive state of PyOx in non-polar and polar solvents.

image file: c8cp05356a-f10.tif
Fig. 10 Energy-level scheme illustrating the photocycle of PyOx in non-polar (CHX) and polar solvents (ACN). Only the relevant excited states are included. The excited-state energies at the ground-state geometry were estimated from the experimental absorption spectra (they are about 0.5 eV lower than the calculated values). The excited-state energies at the S1 state geometry are taken from the quantum-chemical calculations (Table S1, ESI) and were downshifted by 0.5 eV.

The most important features of the photophysics of PyOx are the very efficient ISC processes taking place via the SOCT mechanism. They are operative here because of (i) the CT character of the transitions and (ii) the close to orthogonal orientation of the associated MOs. Additional quantum-chemical calculations with PyP indicate that the HOMO and LUMO remain localised on the Py plane in both the ground and excited states, although the relative orientation of C5-phenyl ring also changes in the relaxed S1 state (Fig. S17, ESI). However, if a cyano group is added on the phenyl ring, the situation becomes similar to that found with PyOx, namely the LUMO is localised on the Py plane in the ground state geometry, but is delocalised over the whole molecule in the excited-state geometry (Fig. S18, ESI). Most probably, the mechanism proposed here to account for the fluorescence quenching of PyOx in non-polar solvents is also responsible for the absence of fluorescence of other pyrazoline derivatives with an electron accepting group in C-5 position. In principle, the fast fluorescence quenching in apolar solvents could be suppressed by preventing structural relaxation of the S1 state. Preliminary measurements in polymer films point to a fluorescence lifetime of PyOx on the nanosecond time scale (Fig. S15, ESI).

5 Conclusions

The above results reveal that the T-shaped PyOx dyad, originally conceived for dual emission, exhibits rich photophysics and efficient but relatively complicated decay pathways from the optically populated excited states down to the ground state. This leads to a very small fluorescence quantum yield. Weak fluorescence in polar solvents can be easily accounted for by the occurrence of a Py to Ox intramolecular electron transfer. However, the decay of the resulting charge-separated state exhibits totally unexpected solvent dependence, being significantly faster in medium-polar than in strongly polar solvents. We explain this by a charge recombination to the triplet manifold via the SOCT ISC mechanism that is favoured by the orthogonal arrangement of the donor and acceptor subunits. Very efficient decay of the lowest triplet is also accounted for by this mechanism. Charge separation is not operative in non-polar solvents and the 200 ps lifetime of the S1 state is explained by SOCT ISC alone. Quantum-chemical calculations reveal that this mechanism can be operative because, upon equilibration, the S1 state undergoes structural changes that result in a partial charge-transfer character of the excited state.

Our investigation shows that an orthogonal arrangement of two chromophoric units indeed allows for a weak coupling despite a short distance. Here however, this geometry is not fully maintained in the relaxed excited state, and consequently, excitation is no longer localised. Because of the different redox properties of the chromophoric sub-units, the excited state acquires some charge-transfer character. Together with the T-shape geometry, this seems to open an efficient non-radiative decay channel of the excited state via the triplet manifold that suppresses fluorescence. Therefore, high fluorescence could be maintained by inhibiting structural relaxation in the excited state. This could be done by replacing the single bond by a rigid spacer or by working in a rigid environment. Further work will lead to a comprehensive understanding of the photophysics of such T-shaped dyads, which are promising candidates for mechanosensing.

Conflicts of interest

There are no conflicts to declare.


We thank Prof. Andrey Doroshenko (Kharkiv National University) for fruitful discussions on the photophysics of these pyrazoline dyes. This work was supported by the Deutsche Forschungsgemeinschaft (SFB749), the cluster of excellence “Munich Centre for Advanced Photonics (MAP)”, the Fonds National Suisse de la Recherche Scientifique (Project No. 200020-165890) and the University of Geneva.


  1. M. Berberan-Santos, J. Canceill, E. Gratton, L. Jullien, J.-M. Lehn, P. So, J. Sutin and B. Valeur, J. Phys. Chem., 1996, 100, 15–20 CrossRef CAS.
  2. E. Yeow, K. Ghiggino, J. Reek, M. Crossley, A. Bosman, A. Schenning and E. Meijer, J. Phys. Chem. B, 2000, 104, 2596–2606 CrossRef CAS.
  3. N. Aratani, A. Osuka, H. Cho and D. Kim, J. Photochem. Photobiol., C, 2002, 3, 25–52 CrossRef CAS.
  4. A. Morandeira, E. Vauthey, A. Schuwey and A. Gossauer, J. Phys. Chem. A, 2004, 108, 5741–5751 CrossRef CAS.
  5. U. Megerle, F. Selmaier, C. Lambert, E. Riedle and S. Lochbrunner, Phys. Chem. Chem. Phys., 2008, 10, 6245–6251 RSC.
  6. N. Aratani, D. Kim and A. Osuka, Acc. Chem. Res., 2009, 42, 1922–1934 CrossRef CAS PubMed.
  7. C. Kirmaier, H.-E. Song, E. Yang, J. K. Schwartz, E. Hindin, J. R. Diers, R. S. Loewe, K.-y. Tomizaki, F. Chevalier, L. Ramos, R. R. Birge, J. S. Lindsey, D. F. Bocian and D. Holten, J. Phys. Chem. B, 2010, 114, 14249–14264 CrossRef CAS PubMed.
  8. C. Sissa, A. Painelli, M. Blanchard-Desce and F. Terenziani, J. Phys. Chem. B, 2011, 115, 7009–7020 CrossRef CAS PubMed.
  9. F. Schlosser, J. Sung, P. Kim, D. Kim and F. Wurthner, Chem. Sci., 2012, 3, 2778–2785 RSC.
  10. R. Ziessel, G. Ulrich, A. Haefele and A. Harriman, J. Am. Chem. Soc., 2013, 135, 11330–11344 CrossRef CAS PubMed.
  11. O. Mongin, M. Sankar, M. Charlot, Y. Mir and M. Blanchard-Desce, Tetrahedron Lett., 2013, 54, 6474–6478 CrossRef CAS.
  12. J. Q. Gong, L. Favereau, H. L. Anderson and L. M. Herz, J. Phys. Chem. Lett., 2016, 7, 332–338 CrossRef CAS PubMed.
  13. E. Martinou, K. Seintis, N. Karakostas, A. Bletsou, N. S. Thomaidis, M. Fakis and G. Pistolis, J. Phys. Chem. C, 2017, 121, 5341–5355 CrossRef CAS.
  14. R. Maragani, M. B. Thomas, R. Misra and F. D'Souza, J. Phys. Chem. A, 2018, 122, 4829–4837 CrossRef CAS PubMed.
  15. C. Le Droumaguet, O. Mongin, M. H. V. Werts and M. Blanchard-Desce, Chem. Commun., 2005, 2802–2804 RSC.
  16. D. Astruc, E. Boisselier and C. Ornelas, Chem. Rev., 2010, 110, 1857–1959 CrossRef CAS PubMed.
  17. H.-Q. Peng, L.-Y. Niu, Y.-Z. Chen, L.-Z. Wu, C.-H. Tung and Q.-Z. Yang, Chem. Rev., 2015, 115, 7502–7542 CrossRef CAS PubMed.
  18. G. Sathiyan and P. Sakthivel, RSC Adv., 2016, 6, 106705–106715 RSC.
  19. M. Vishe, T. Lathion, S. Pascal, O. Yushchenko, A. Homberg, E. Brun, E. Vauthey, C. Piguet and J. Lacour, Helv. Chim. Acta, 2018, 101, e1700265 CrossRef.
  20. J. M. Tour, Acc. Chem. Res., 2000, 33, 791–804 CrossRef CAS PubMed.
  21. D. Holten, D. F. Bocian and J. S. Lindsey, Acc. Chem. Res., 2002, 35, 57–69 CrossRef CAS PubMed.
  22. P. A. de Silva and N. D. McClenaghan, Chem. – Eur. J., 2004, 10, 574–586 CrossRef PubMed.
  23. A. Painelli and F. Terenziani, Synth. Met., 2004, 147, 111–115 CrossRef CAS.
  24. A. Jain and S. J. George, Mater. Today, 2015, 18, 206–214 CrossRef CAS.
  25. M.-N. Yu, H. Soleimaninejad, J.-Y. Lin, Z.-Y. Zuo, B. Liu, Y.-F. Bo, L.-B. Bai, Y.-M. Han, T. A. Smith, M. Xu, X.-P. Wu, D. E. Dunstan, R.-D. Xia, L.-H. Xie, D. D. C. Bradley and W. Huang, J. Phys. Chem. Lett., 2018, 9, 364–372 CrossRef CAS PubMed.
  26. G. L. Tu, C. Y. Mei, Q. G. Zhou, Y. X. Cheng, Y. H. Geng, L. X. Wang, D. G. Ma, X. B. Jing and F. S. Wang, Adv. Funct. Mater., 2005, 16, 101–106 CrossRef.
  27. Y. Liu, M. Nishiura, Y. Wang and Z. Hou, J. Am. Chem. Soc., 2006, 128, 5592–5593 CrossRef CAS PubMed.
  28. Q.-Y. Yang and J.-M. Lehn, Angew. Chem., Int. Ed., 2014, 53, 4572–4577 CrossRef CAS PubMed.
  29. B. Xu, Y. Mu, Z. Mao, Z. Xie, H. Wu, Y. Zhang, C. Jin, Z. Chi, S. Liu, J. Xu, Y.-C. Wu, P.-Y. Lu, A. Lien and M. R. Bryce, Chem. Sci., 2016, 7, 2201–2206 RSC.
  30. M. J. Ahrens, L. E. Sinks, B. Rybtchinski, W. Liu, B. A. Jones, J. M. Giaimo, A. V. Gusev, A. J. Goshe, D. M. Tiede and M. R. Wasielewski, J. Am. Chem. Soc., 2004, 126, 8284–8294 CrossRef CAS PubMed.
  31. G. Kodis, Y. Terazono, P. A. Liddell, J. Andreasson, V. Garg, M. Hambourger, T. A. Moore, A. L. Moore and D. Gust, J. Am. Chem. Soc., 2006, 128, 1818–1827 CrossRef CAS PubMed.
  32. I. Oesterling and K. Muellen, J. Am. Chem. Soc., 2007, 129, 4595–4605 CrossRef CAS PubMed.
  33. N. Sakai, M. Lista, O. Kel, S.-i. Sakurai, D. Emery, J. Mareda, E. Vauthey and S. Matile, J. Am. Chem. Soc., 2011, 133, 15224–15227 CrossRef CAS PubMed.
  34. O. Yushchenko, D. Villamaina, N. Sakai, S. Matile and E. Vauthey, J. Phys. Chem. C, 2015, 119, 14999–15008 CrossRef CAS.
  35. K. Petterson, A. Kyrychenko, E. Rönnow, T. Ljungdahl, J. Martensson and B. Albinsson, J. Phys. Chem. A, 2006, 110, 310–318 CrossRef PubMed.
  36. B. Albinsson and J. Martensson, J. Photochem. Photobiol., C, 2008, 9, 138–155 CrossRef CAS.
  37. P. Kölle, I. Pugliesi, H. Langhals, R. Wilcken, A. J. Esterbauer, R. de Vivie-Riedle and E. Riedle, Phys. Chem. Chem. Phys., 2015, 17, 25061–25072 RSC.
  38. H. Langhals, A. J. Esterbauer, A. Walter, E. Riedle and I. Pugliesi, J. Am. Chem. Soc., 2010, 132, 16777–16782 CrossRef CAS PubMed.
  39. G. J. Hedley, A. Ruseckas, A. C. Benniston, A. Harriman and I. D. W. Samuel, J. Phys. Chem. A, 2015, 119, 12665–12671 CrossRef CAS PubMed.
  40. C. Menelaou, J. t. Schiphorst, A. M. Kendhale, P. Parkinson, M. G. Debije, A. P. H. J. Schenning and L. M. Herz, J. Phys. Chem. Lett., 2015, 6, 1170–1176 CrossRef CAS PubMed.
  41. D. N. H. Meineke, M. L. Bossi, H. Ta, V. N. Belov and S. W. Hell, Chem. – Eur. J., 2016, 23, 2469–2475 CrossRef PubMed.
  42. V. Gray, B. Küçükoz, F. Edhborg, M. Abrahamsson, K. Moth-Poulsen and B. Albinsson, Phys. Chem. Chem. Phys., 2018, 20, 7549–7558 RSC.
  43. H. Strähle, W. Seitz and H. Güsten, Ber. Bunsenges. Phys. Chem., 1976, 80, 288–294 CrossRef.
  44. H. Strähle, W. Seitz and H. Güsten, Z. Naturforsch. B, 1976, 31, 1248–1255 Search PubMed.
  45. H. Güsten, G. Heinrich and H. Früihbeis, Ber. Bunsenges. Phys. Chem., 1977, 81, 810–815 CrossRef.
  46. A. O. Doroshenko, V. T. Skripkina, V. M. Schershukov and O. A. Ponomaryov, J. Fluoresc., 1997, 7, 131–138 CrossRef CAS.
  47. Y. Zhenglin and W. Shikang, J. Lumin., 1993, 54, 303–308 CrossRef CAS.
  48. M. R. V. Sahyun and G. P. Crooks, Proc. SPIE, 1991, 1436, 125–133 CrossRef CAS.
  49. J. T. Blair, M. R. V. Sahyun and D. K. Sharma, J. Photochem. Photobiol., A, 1994, 77, 133–139 CrossRef CAS.
  50. M. Mardelli and J. Olmsted, J. Photochem., 1977, 7, 277–285 CrossRef CAS.
  51. A. V. Deshpande, A. Beidoun, A. Penzkofer and G. Wagenblast, Chem. Phys., 1990, 142, 123–131 CrossRef CAS.
  52. P.-A. Muller, C. Högemann, X. Allonas, P. Jacques and E. Vauthey, Chem. Phys. Lett., 2000, 326, 321–327 CrossRef CAS.
  53. A. Morandeira, L. Engeli and E. Vauthey, J. Phys. Chem. A, 2002, 106, 4833–4837 CrossRef CAS.
  54. G. Duvanel, J. Grilj, H. Chaumeil, P. Jacques and E. Vauthey, Photochem. Photobiol. Sci., 2010, 9, 908–915 RSC.
  55. G. Duvanel, N. Banerji and E. Vauthey, J. Phys. Chem. A, 2007, 111, 5361–5369 CrossRef CAS PubMed.
  56. N. Banerji, G. Duvanel, A. Perez-Velasco, S. Maity, N. Sakai, S. Matile and E. Vauthey, J. Phys. Chem. A, 2009, 113, 8202–8212 CrossRef CAS PubMed.
  57. E. Riedle, M. Beutter, S. Lochbrunner, J. Piel, S. Schenkl, S. Spörlein and W. Zinth, Appl. Phys. B: Lasers Opt., 2000, 71, 457–465 CrossRef CAS.
  58. E. Riedle, M. Bradler, M. Wenninger, C. F. Sailer and I. Pugliesi, Faraday Discuss., 2013, 163, 139–158 RSC.
  59. C. Homann, N. Krebs and E. Riedle, Appl. Phys. B: Lasers Opt., 2011, 104, 783–791 CrossRef CAS.
  60. U. Megerle, I. Pugliesi, C. Schriever, C. F. Sailer and E. Riedle, Appl. Phys. B: Lasers Opt., 2009, 96, 215–231 CrossRef CAS.
  61. T. Yanai, D. P. Tew and N. C. Handy, Chem. Phys. Lett., 2004, 393, 51–57 CrossRef CAS.
  62. M. J. G. Peach, P. Benfield, T. Helgaker and D. J. Tozer, J. Chem. Phys., 2008, 128, 044118 CrossRef PubMed.
  63. P. Krawczyk, J. Mol. Model., 2010, 16, 659–668 CrossRef CAS PubMed.
  64. D. Jacquemin, A. Planchat, C. Adamo and B. Mennucci, J. Chem. Theory Comput., 2012, 8, 2359–2372 CrossRef CAS PubMed.
  65. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz, A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F. Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson, D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W. Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K. Throssell, J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. J. Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov, T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Millam, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin, K. Morokuma, O. Farkas, J. B. Foresman and D. J. Fox, Gaussian 16 Rev. B.01, 2016 Search PubMed.
  66. C. Foces-Foces, N. Jagerovic and J. Elguero, Z. Kristallogr., 2001, 216, 240–244 CAS.
  67. N. Boens, W. Qin, N. Basaric, J. Hofkens, M. Ameloot, J. Pouget, J.-P. Lefevre, B. Valeur, E. Gratton, M. vandeVen, N. D. Silva, Y. Engelborghs, K. Willaert, A. Sillen, G. Rumbles, D. Phillips, A. J. W. G. Visser, A. van Hoek, J. R. Lakowicz, H. Malak, I. Gryczynski, A. G. Szabo, D. T. Krajcarski, N. Tamai and A. Miura, Anal. Chem., 2007, 79, 2137–2149 CrossRef CAS PubMed.
  68. M. L. Horng, J. A. Gardecki, A. Papazyan and M. Maroncelli, J. Phys. Chem., 1995, 99, 17311–17337 CrossRef CAS.
  69. A. Pigliucci, G. Duvanel, L. M. L. Daku and E. Vauthey, J. Phys. Chem. A, 2007, 111, 6135–6145 CrossRef CAS PubMed.
  70. D. R. Lide, Handbook of Chemistry and Physics, CRC Press, Boca Raton, 91st edn, 2010 Search PubMed.
  71. I. H. M. van Stokkum, D. S. Larsen and R. van Grondelle, Biochim. Biophys. Acta, Bioenerg., 2004, 1657, 82–104 CrossRef CAS PubMed.
  72. A. Weller, Z. Phys. Chem., 1982, 133, 93–98 CrossRef CAS.
  73. F. Pragst, J. Prakt. Chem., 1973, 315, 549–564 CrossRef CAS.
  74. T. Shida, Electronic Absorption Spectra of Radical Ions, Elsevier, Amsterdam, 1988, vol. 34 Search PubMed.
  75. R. A. Marcus and N. Sutin, Biochim. Biophys. Acta, 1985, 811, 265–322 CrossRef CAS.
  76. E. Haselbach, E. Vauthey and P. Suppan, Tetrahedron, 1988, 44, 7335–7344 CrossRef CAS.
  77. I. V. Khudyakov, Y. A. Serebrennikov and N. J. Turro, Chem. Rev., 1993, 93, 537–570 CrossRef CAS.
  78. T. Okada, M. Migita, N. Mataga, Y. Sakata and S. Misumi, J. Am. Chem. Soc., 1981, 103, 4715–4720 CrossRef CAS.
  79. M. R. Wasielewski, D. G. Johnson, W. A. Svec, K. M. Kersey and D. W. Minsek, J. Am. Chem. Soc., 1988, 110, 7219–7221 CrossRef CAS.
  80. Z. E. X. Dance, S. M. Mickley, T. M. Wilson, A. B. Ricks, A. M. Scott, M. A. Ratner and M. R. Wasielewski, J. Phys. Chem. A, 2008, 112, 4194–4201 CrossRef CAS PubMed.
  81. A. R. Mallia, P. S. Salini and M. Hariharan, J. Am. Chem. Soc., 2015, 137, 15604–15607 CrossRef CAS PubMed.
  82. S. Das, W. G. Thornbury, A. N. Bartynski, M. E. Thompson and S. E. Bradforth, J. Phys. Chem. Lett., 2018, 9, 3264–3270 CrossRef CAS PubMed.
  83. L. Salem and C. Rowland, Angew. Chem., Int. Ed. Engl., 1972, 11, 92–111 CrossRef CAS.
  84. L. Carlacci, C. Doubleday, T. R. Furlani, H. F. King and J. W. McIver, J. Am. Chem. Soc., 1987, 109, 5323–5329 CrossRef CAS.
  85. V. A. Nadtochenko, I. V. Vasil'ev, N. N. Denisov, I. V. Rubtsov, A. S. Lobach, A. P. Moravskii and A. F. Shestakov, J. Photochem. Photobiol., A, 1993, 70, 153–156 CrossRef CAS.
  86. A. Watanabe, O. Ito, M. Watanabe, H. Saito and M. Koishi, J. Phys. Chem., 1996, 100, 10518–10522 CrossRef CAS.
  87. A. Sassara, G. Zerza, M. Chergui, V. Ciulin, J. D. Ganiere and B. Deveaud, J. Chem. Phys., 1999, 111, 689–697 CrossRef CAS.
  88. H. Ago, M. S. P. Shaffer, D. S. Ginger, A. H. Windle and R. H. Friend, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 2286–2290 CrossRef CAS.
  89. J. Park, P. Deria and M. J. Therien, J. Am. Chem. Soc., 2011, 133, 17156–17159 CrossRef CAS PubMed.
  90. K. Nagarajan, A. R. Mallia, K. Muraleedharan and M. Hariharan, Chem. Sci., 2017, 8, 1776–1782 RSC.
  91. B. F. Habenicht and O. V. Prezhdo, J. Am. Chem. Soc., 2012, 134, 15648–15651 CrossRef CAS PubMed.
  92. Y. Liu, M. Lin and Y. Zhao, J. Phys. Chem. A, 2017, 121, 1145–1152 CrossRef CAS PubMed.
  93. F. Franco de Carvalho, C. A. Pignedoli and I. Tavernelli, J. Phys. Chem. C, 2017, 121, 10140–10152 CrossRef CAS.
  94. M. J. Bearpark, F. Bernardi, S. Clifford, M. Olivucci, M. A. Robb, B. R. Smith and T. Vreven, J. Am. Chem. Soc., 1996, 118, 169–175 CrossRef CAS.
  95. A. Sanchez-Galvez, P. Hunt, M. A. Robb, M. Olivucci, T. Vreven and H. B. Schlegel, J. Am. Chem. Soc., 2000, 122, 2911–2924 CrossRef CAS.
  96. B. G. Levine and T. J. Martinez, Annu. Rev. Phys. Chem., 2007, 58, 613–634 CrossRef CAS PubMed.
  97. T. Gustavsson, A. Banyasz, E. Lazzarotto, D. Markovitsi, G. Scalmani, M. J. Frisch, V. Barone and R. Improta, J. Am. Chem. Soc., 2006, 128, 607–619 CrossRef CAS PubMed.
  98. A. Nakayama and T. Taketsugu, J. Phys. Chem. A, 2011, 115, 8808–8815 CrossRef CAS PubMed.
  99. J. Grilj, E. N. Laricheva, M. Olivucci and E. Vauthey, Angew. Chem., Int. Ed., 2011, 50, 4496–4498 CrossRef CAS PubMed.
  100. M. Quick, A. L. Dobryakov, M. Gerecke, C. Richter, F. Berndt, I. N. Ioffe, A. A. Granovsky, R. Mahrwald, N. P. Ernsting and S. A. Kovalenko, J. Phys. Chem. B, 2014, 118, 8756–8771 CrossRef CAS PubMed.
  101. T. Kumpulainen, B. Lang, A. Rosspeintner and E. Vauthey, Chem. Rev., 2017, 117, 10826–10939 CrossRef CAS PubMed.


Electronic supplementary information (ESI) available: Synthesis, details on the global analysis, stationary electronic spectra, transient absorption data, fluorescence in polymer films, application of the Weller equation and of the Förster model, quantum-chemical calculations. See DOI: 10.1039/c8cp05356a
Present address: European XFEL GmbH, Holzkoppel 4, 22869 Schenefeld, Germany.
§ Present address: Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, USA.

This journal is © the Owner Societies 2018