Unexpected multiple activated steps in the excited state decay of some bis(phenylethynyl)-fluorenes and -anthracenes

Abstract

The temperature effect on the photophysical parameters of four acetylene-derivatives [bis(phenylethynyl)-anthracenes and -fluorenes with substituents of different electron acceptor efficiencies] has been investigated by absorption and emission spectroscopy, using stationary and pulsed (ns/fs resolution) techniques. The nature of the central nucleus (anthracene or fluorene) and the peripheral electron-withdrawing group (nitro or formyl) strongly affect the deactivation of the excited states of these push–pull molecules. In some cases the study evidenced an interesting role of two activated steps in the deactivation of the excited singlet state, namely an activated inter-system crossing to an upper triplet state of n,π* nature (previously hypothesized on the basis of TD-DFT calculations) and a sort of activated internal conversion, discussed also on the basis of maximum entropy method analysis of the fluorescence decay data. Nicely, an efficient ISC was found for the fluorene-derivatives where small energy gaps between S₁ (π,π*) and T_n (n,π*) states had been calculated while no activated ISC was evidenced in the case of anthryl-derivatives where higher S₁–T_n energy gaps are expected. A peculiar temperature effect for a fluorene-derivative was pointed out and also explained on the basis of quantum-mechanical calculations at the DFT level taking into account the solvation effects by means of the conductor-like polarizable continuum model CPCM. The presence of dual emission, at first evidenced by a shoulder in the emission spectrum of the fluorene-derivative featuring a peripheral formyl group in dichloromethane at low temperatures, was nicely confirmed by femtosecond up-conversion measurements at room temperature.

---

Introduction

The present contribution is part of a research project on the excited state properties of organic dyes such as bis(phenylethynyl)-anthracenes and -fluorenes^1,2 with substituents of different electron acceptor efficiencies (Chart 1). Acetylene-based materials are really interesting for applications in organic electronic devices including light emitting diodes, field-effect transistors, organic photovoltaic cells and fluorescent sensors.^3–5 Moreover, the presence of electron-acceptor units on one side of the molecules and the mild electron-donor alkoxy groups on the other side give these systems a push–pull character that can induce intramolecular charge transfer (ICT), as observed in ethene-derivatives,⁶ and solvent effects of potential interest for applications in nonlinear optics (NLO).⁷ In fact, the decay processes of the excited states of these compounds may be complicated by the involvement of ICT processes occurring during their lifetime.⁸

Chart 1 Investigated compounds.

In a previous work⁸ the spectral and photophysical behaviour of ethynyl-fluorenes and -anthracenes in solvents of different polarities were investigated by stationary, fast/ultrafast spectroscopy and theoretical DFT calculations. A substantial positive fluorosolvatochromism was observed and the photobehaviour was found to dramatically change upon increasing the medium polarity, also strongly depending on the structure of the aromatic central units.^8,9 Namely, when the central unit was the anthracene a significantly red-shifted absorption spectrum was observed making these systems promising for applications as absorbers in organic solar cells.³ On the other hand, when the central unit was the fluorene transitions with high CT character were favoured, making these molecules more appealing for NLO applications.⁷ Both the ICT character of the lowest excited states and the population of the triplet state decrease in the order NO₂ > CHO > CN, in agreement with the electron affinity of the side groups and the energy of the ³(n,π*) state. Therefore, only in the case of the fluorene-derivatives and the nitro-anthracene compound the fs transient absorption measurements evidenced the presence of two singlet excited states, namely a locally excited (LE) state, produced by absorption, and an ICT state originating from the LE state, their lifetimes being strongly affected by the solvent polarity. Moreover, a dramatic reduction of the fluorescence quantum yield (ϕ_F) and the parallel decrease of the triplet production in polar solvents indicated that an efficient internal conversion (IC) plays a fundamental role in the deactivation of the strongly stabilized ICT state, due to the small S₁–S₀ energy gap.⁸ In the other two anthracene-derivatives, having CHO and CN as substituents, no ICT states have been found in agreement with their high ϕ_F in both non-polar and polar solvents.

A peculiar behaviour of the ISC rate constant (k_ISC) was found for some of them in agreement with a previous study on analogous compounds,¹⁰ characterized by an efficient ISC process in non-polar solvents that decreases by two orders of magnitude by slightly increasing the solvent polarity, while a further increase of the latter causes a new increase of k_ISC, which when investigated in more polar solvents (acetone and dimethylformamide) reaches values similar to that in cyclohexane (CH). This behaviour was interpreted on the basis of quantum-mechanical calculations at the DFT level of theory that predicted the presence of (n,π*) states at rather low energy for the fluorene-derivatives, where the lowest ³(n,π*) and ¹(π,π*) states are located in the interval 3.1–3.9 eV in vacuum⁸ and can be differently stabilized in polar solvents (see Scheme 2 of ref. 8). In fact, it is well known that high spin–orbit coupling is expected for ISC between (n,π*) and (π,π*) states, according to the El-Sayed rules, and the increase of solvent polarity stabilises the (π,π*) states and destabilises the (n,π*) ones, thus their relative position and the energy gap between them becoming solvent dependent. Therefore, in solvents of intermediate polarity where the lowest excited state is ¹LE* of π,π* nature, a ³(n,π*) state is expected to lie close to the optical ¹(π,π*) state (see Chart 2) and then an activated singlet–triplet ISC can be operative, in agreement with the significant reduction of the kinetic constant. However, the further increase of k_ISC in highly polar solvents is due to the intrinsic nature of S₁, the CT character of which increases with solvent polarity, and is, in this case, a measure of the efficient spin–orbit coupling between the ¹CT* and low-lying ³(π,π*) states of different nature. For the anthryl-derivatives, where the ³(n,π*) state was predicted to be too high with respect to S₁ (see Chart 2), the ISC was found to be very slow and practically inefficient at room temperature.⁸

Chart 2 Sketch of the ground and lowest excited states involved in the photobehaviour of the investigated compounds as derived by quantum-mechanical calculations.⁸

This paper reports a temperature effect on the photophysical parameters in solvents of suitable polarity carried out to highlight the presence of activated ISC pathways predicted on the basis of quantum-mechanical calculations and the effect of solvent polarity. New DFT calculations, MEM analysis applied to single photon counting decay kinetics and femtosecond up-conversion measurements were required to understand the results regarding the presence of additional activated steps and dual-fluorescence.

Results and discussion

The choice of the solvent

As discussed in the Introduction section, in the case of fluorene-derivatives a peculiar trend of k_ISC in different media as a function of dielectric constant (ε) was observed (Fig. 1) and explained on the basis of the presence of an efficient ISC between ¹LE* of π,π* nature and a closely located ³(n,π*) state that becomes less likely after increasing the solvent polarity, probably because the different stabilization of the two involved states of different nature leads to an activated ISC from the ¹LE* to ³(n,π*) state. The new increase of k_ISC due to a further increase of ε has been attributed to an increased CT character of the lowest excited singlet state that allows an efficient ISC from ¹CT* to a low-lying ³(π,π*) state to become operative.

Fig. 1 Trend of k_ISC by changing the dielectric constant (ε) of the medium in the case of the fluorene-derivatives. Data retrieved from ref. 8 (DEE = diethylether, EtAc = ethylacetate, DCE = dichloroethane, Ac = acetone, DMF = dimethylformamide).

In this situation toluene (Tol) seemed to be a good solvent to explore the temperature range below 293 K, owing to its low melting point, the low dependence of its dielectric constant on the temperature and because it is at the beginning of a plateau in the plot shown in Fig. 1. In fact, k_ISC is expected not to change in this ε range (2.38–4.27) although the decrease of temperature causes some slight increase of ε. Moreover, fs transient absorption measurements ensured that only ¹LE* is present for NF and AF in Tol and no sign of the ICT process was evidenced in this solvent. For AF the measurements at low T were also performed in dichloromethane (DCM) to explore a solvent where the lowest k_ISC value had been found even if in this medium an ICT process after excitation is expected on the basis of previous fs measurements.⁸

For the anthryl-derivatives methylcyclohexane (MCH) was considered a good low polar solvent suitable to explore high temperatures to favour an eventual population of high-energy ³(n,π*) states.

Temperature effect on the spectral behaviour. The absorption and emission spectra of the investigated compounds recorded at different temperatures are reported in Fig. S1 and S2 (ESI†). The absorption spectra showed the expected increase of the vibronic structure at low temperatures. The emission spectra shifted towards the red and the fluorescence intensity generally increased (strongly for NF and AF, see Fig. S3 and S4 in the ESI†) on decreasing the temperature. The largest red-shift caused by the increase of ε at low temperatures was observed for the fluorescence of NF in Tol (from 478 nm at 293 K to 513 nm at 195 K) in agreement with its strongest push–pull character.⁸

A different behaviour was observed for AF in the more polar DCM. In fact, in this solvent the emission spectrum (red-shifted by about 100 nm with respect to that in Tol) decreases in intensity and moves towards the red by cooling. The normalized fluorescence spectra in Fig. S2 (ESI†) evidenced that below 250 K the red-shift is accompanied by the appearance of a hypsochromic shoulder at ca. 420 nm (falling in the spectral region where AF emits in Tol). Probably this shoulder is present in the entire investigated temperature range but hidden at T > 250 K under the more intense and blue-shifted emission (see Fig. S5, ESI†).

Temperature effect on the emission properties. Tables 1 and 2 report the emission properties of the investigated compounds at different temperatures. The emission parameters measured for NA in MCH do not change in the range 293–333 K. Moreover, ϕ_F ≅ 1 at all the monitored T. These findings suggest that fluorescence remains the only deactivation pathway also at high T. For AA in MCH the probability of radiative relaxation is 90% at room temperature and slightly reduces on increasing T up to 358 K.

Table 1 Fluorescence quantum yield (ϕ_F), lifetime (τ_F) and kinetic constant (k_F) of NA and AA in MCH at different temperatures

T/K	NA in MCH			AA in MCH
	ϕ F	τ F/ns	k F/10^8 s^−1	ϕ F	τ F/ns	k F/10^8 s^−1
293	0.98	2.63	3.7	0.90	2.63	3.4
303	1.02	2.64	3.9	0.89	2.62	3.4
313	1.05	2.66	4.0	0.90	2.64	3.4
323	1.06	2.68	4.0	0.81	2.67	3.0
333	1.07	2.71	4.0	0.79	2.70	2.9
343	1.06	2.73	3.9	0.78	2.64	2.9
358	1.05	2.73	3.9	0.75	2.61	2.8

---

Table 2 Fluorescence quantum yield (ϕ_F), lifetime (τ_F) and kinetic constant (k_F) of NF in Tol and AF in Tol and DCM at different temperatures

T/K	NF in Tol			AF in Tol			AF in DCM
	ϕ F	τ F/ns	k F/10^8 s^−1	ϕ F	τ F/ns	k F/10^8 s^−1	ϕ F	τ F/ns	k F/10^8 s^−1
a From fs measurements (ref. 8). b At 200 K.
293	0.013	0.045^a	2.9	0.15	0.28	5.4	0.17	0.88	1.9
280	0.026	—	—	0.19	0.34	5.6	0.14	0.74	1.9
265	0.055	—	—	0.23	0.40	5.8	0.10	0.59	1.7
250	0.12	0.41	2.9	0.28	0.47	6.0	0.08	0.47	1.7
235	0.24	0.73	3.3	0.33	0.53	6.2	0.05	0.38	1.3
220	0.42	1.22	3.4	0.37	0.61	6.1	0.05	0.33	1.5
205	0.62	1.79	3.5	0.41	0.68	6.0	0.04^b	0.33^b	1.2^b
195	0.71	2.07	3.4	0.43	0.74	5.8

---

The fluorene-derivatives studied in Tol below room T showed parallel increases of ϕ_F and τ_F by decreasing T, particularly marked for NF (where ϕ_F changes by almost one order of magnitude from 0.013 at 293 K to 0.71 at 195 K), maintaining the kinetic constant practically unchanged in the investigated T range.

An opposite trend was found for ϕ_F and τ_F of AF in DCM as a function of T. In fact, a significant reduction of their values was measured upon decreasing T. Such a trend is accompanied by a 40% reduction of k_F pointing out a progressive change of the nature of the emitting state making fluorescence a less allowed process at low T.

Temperature effect on the triplet properties. The spectral and kinetic properties and the population quantum yields for the triplet state of the investigated compounds at different T (obtained by laser flash photolysis with ns resolution) are reported in Tables 3 and 4. The T₁ → T_n absorption spectra of AA in MCH at 358 K and AF in Tol at 195 K are reported in Fig. S6 (ESI†) as an example. They show two main bands, the first at 360 nm (AA) and 525 nm (AF), blue-shifted with respect to those at room T (380 nm for AA and 535 nm for AF), while the latter at 540 nm (AA) and 760 nm (AF) were comparable with those at 530 nm (AA) and 710 nm (AF) at 293 K.

Table 3 Triplet absorption (ΔA), quantum yield (ϕ_T), lifetime (τ_T) and kinetic constant (k_ISC) of AA in MCH at different temperatures

T/K	ΔA	ϕ T	k ISC/10^6 s^−1
293	0.016	0.008	3.0
303	0.021	0.011	4.2
313	0.023	0.012	4.6
323	0.024	0.012	4.5
333	0.023	0.012	4.5
343	0.023	0.012	4.4
358	0.023	0.011	4.0

---

Table 4 Triplet absorption (ΔA), quantum yield (ϕ_T), lifetime (τ_T) and kinetic constant (k_ISC) of NF in Tol and AF in Tol and DCM at different temperatures

T/K	NF in Tol				AF in Tol				AF in DCM
	ΔA	τ T/μs	ϕ T	k ISC /10^8 s^−1	ΔA	τ T/μs	ϕ T	k ISC /10^8 s^−1	ΔA	τ T/μs	ϕ T	k ISC /10^8 s^−1
a For kISC = ϕT/τF, the τF value at 293 K is from fs measurements (ref. 8). b At 200 K.
293	0.358	2.9	0.38	84	0.219	1.1	0.37	13.2	0.033	21	0.036	0.40
280	0.349	3.2	0.37		0.228		0.39	11.5	0.031	18	0.034	0.46
265	0.342	3.8	0.37		0.238	3.3	0.40	10.0	0.026	12	0.028	0.47
250	0.349	4.3	0.37	9.3	0.275	5.3	0.46	9.79	0.026	15	0.028	0.60
235	0.364	8.3	0.40	5.5	0.271	8.2	0.46	8.68	0.018	17	0.020	0.53
220	0.340	7.6	0.36	3.0	0.268	13	0.45	7.38	0.021	10	0.023	0.70
205	0.360	10	0.38	2.1	0.268	19	0.45	6.62	0.016^b	12^b	0.017^b	0.52^b
195	0.334		0.35	1.7	0.249	29	0.42	5.68

---

The triplet population is not influenced by T for both NF and AF in Tol, but k_ISC strongly decreases on going from 293 K to 195 K confirming the presence in this solvent of an activated ISC to an upper triplet state of n,π* nature. In fact, the k_ISC value derived at room T was found to be particularly high (10⁸–10⁹ s⁻¹) in agreement with the El-Sayed rules. In contrast, the slight decrease of ϕ_T observed for AF in DCM, that parallels the decrease of τ_F, leaves k_ISC practically unchanged in the 293–200 K range investigated. The triplet lifetime increases at low T for the two fluorene-derivatives, as expected, but it is roughly halved on going from 293 K to 200 K in DCM, probably due to an increased CT character of the triplet state upon increasing the dielectric constant at low T.

Activated internal conversion and MEM analysis of decay kinetics. The IC quantum yield values (ϕ_IC) calculated for the investigated compounds at different T by taking into account all the absorbed quanta and the efficiency of the competitive relaxation processes of S₁ (ϕ_IC = 1 − ϕ_F − ϕ_ISC) and the derived kinetic constant (k_IC = ϕ_IC/τ_F) are reported in Table 5. Generally, they show a significant increase on increasing T. These results seem to suggest the presence of some kind of activated IC.¹¹ A similar behavior was also found in aza- and diaza-trans-stilbenes^12,13 and explained on the basis of calculated k_IC taking into account the role of Lim's proximity effect in molecules containing non-bonding electrons.^14–16 In fact, in these flexible molecules in fluid solutions it is expected that low frequency torsional modes, including the twisting of the aromatic fragments around single bonds, thermally populated, can act as promoting modes for vibronic coupling between excited electronic singlet states of n,π* and π,π* nature (see Chart 2), but also as accepting modes leading to an increase of the rate of the S₁ → S₀ IC process.^12–15 Moreover, the increase of T causes the presence, in fluid solutions, of different sets of twisted conformations usually responsible for the enlarged UV-Vis spectrum,^1,12,13,17 while more planar and stable forms are selected at low T. It has been reported,¹² (and nicely confirmed by TDDFT calculations for AF in Tol, see below) that a reduced energy gap between S₁ (π,π*) and the upper excited singlet state of n,π* nature is expected in the twisted geometries, which causes an increase of deactivation to the ground state through IC. The latter explanation is in agreement with the results obtained by the MEM analysis performed on the fluorescence decay data at different T obtained by single photon counting technique with ns resolution.

Table 5 Internal conversion quantum yield (ϕ_IC) and relative kinetic constant (k_IC/10⁸ s⁻¹) derived for the investigated compounds at different T

AA in MCH			NF in Tol			AF in Tol			AF in DCM
T/K	ϕ IC	k IC	T/K	ϕ IC	k IC	T/K	ϕ IC	k IC	T/K	ϕ IC	k IC
293	0.092	0.35	293	0.61	135	293	0.48	17	293	0.79	9.0
303	0.099	0.38	280	0.60		280	0.42	12	280	0.83	11
313	0.088	0.33	265	0.57		265	0.37	9.3	265	0.87	15
323	0.18	0.67	250	0.51	13	250	0.26	5.4	250	0.89	19
333	0.20	0.74	235	0.36	4.9	235	0.21	4.0	235	0.93	24
343	0.21	0.78	220	0.22	1.8	220	0.18	3.0	220	0.93	28
358	0.24	0.88	205	—	—	205	0.14	2.1	200	0.94	28
			195	—	—	195	0.15	2.0

---

Processing the fluorescence decay data by MEM analysis gave interesting results shown in Fig. 2 and Fig. S7–S10 (ESI†). In fact, at high T a broad distribution of lifetimes centered around the mean value was found in the case of compounds (AA in MCH, NF and AF in Tol) where an increase of k_IC with T has been derived. This finding supports the presence at high T of a large set of conformations with slightly different geometries that deactivate through different pathways, IC being the predominant one in the twisted conformations. Decreasing the T, narrow distributions of τ_F are found pointing to a selection of more planar and rigid conformations where fluorescence prevails on IC as a deactivation route for S₁.

Fig. 2 Distributions of lifetimes for AF in toluene at 293 K (A) and 205 K (D); NF in toluene at 250 K (B) and 220 K (E); AA in MCH at 358 K (C) and 293 K (F).

Different is the case of AF in DCM where IC is the main deactivation pathway in all the investigated T range but ϕ_IC and k_IC reach higher values at low T. This behaviour is due to the presence of a low-lying ¹ICT* state in this solvent of medium polarity where AF deactivates to the ground state mainly through IC. The increase of ε (from 9.08 at 293 K to 12.57 at 200 K) causes an increased CT character of S₁ implying an increase of k_IC.

The fluorescence quantum yield (ϕ_F) and the kinetic constants for intersystem-crossing (k_ISC) and internal conversion (k_IC) as a function of temperature were elaborated according to the following Arrhenius-type equations to derive the activation energy (E_a) and the pre-exponential factor (A) of different activated pathways:

where k⁰_ISC/IC is the residual k_ISC/IC present at low T. Eqn (2)is used when IC is not present at low T. where ϕ^lim_F is the limit value for ϕ_F at low T.

The fitting procedures usually allowed a good correlation (r² in the range 0.999–0.98) and the obtained results are collected in Table 6.

Table 6 Arrhenius parameters (activated energy barrier, E_a/kcal mol⁻¹, and pre-exponential factor, A/10¹⁰ s⁻¹, together with limit values at low T for the fluorescence quantum yield, ϕ^lim_F, and the ISC rate constant, k⁰_ISC/10⁸ s⁻¹) in MCH for AA and in Tol for NF and AF, as derived by application of eqn (1)–(3)

Compound	From kISC values (eqn (1))			From kIC values		From ϕF values (eqn (3))
	k ^0ISC	E a,ISC	A	E a,IC	A	ϕ ^limF	E a	A
a From eqn (1). b From eqn (2).
AA				3.2^a	0.49^a	0.90	3.7	1.4
NF	1.7	7.4	2.6 × 10^5	7.6^b	6.4 × 10^5 ^b	0.73	7.7	1.4 × 10^6
AF	1.3	1.1	0.7	2.8^a	21^a	0.46	3.9	210

---

The plots for the ϕ_F values as a function of T and their fittings according to eqn (3) are shown in Fig. 3 for NF and AF in Tol as an example (the other trends are reported in Fig. S11–S16, ESI†).

Fig. 3 Trends of ϕ_F as a function of T and relative fittings (red curves) according to eqn (3) for NF and AF in Tol (the fitting parameters A, B and C are A/k_F, E_a/RT and ϕ^lim_F respectively).

The different treatments agree with the presence of an activated IC (E_aca. 3.5 kcal mol⁻¹) in the case of AA in MCH evidenced by both the trends of k_IC and ϕ_F at different T; both activated ISC and IC processes with similar energy barriers (E_aca. 7.5 kcal mol⁻¹) in the case of NF in Tol, and a slightly activated ISC (E_a = 1.1 kcal mol⁻¹) and an activated IC with higher barrier (E_aca. 3 kcal mol⁻¹) in the case of AF in Tol.

Quantum mechanical calculations: a comparison between AF behaviour in Tol and DCM. As mentioned in the Introduction section, previous calculations predicted (n,π*) states at rather low energy for the investigated compounds, where the lowest ³(n,π*) and ¹(π,π*) states are located in the interval 70–90 kcal mol⁻¹.⁸ Particularly interesting is the case of NF and AF where a ³(n,π*) state is close lying to the optical ¹(π,π*) state and then can be involved in the activated singlet–triplet ISC (see Chart 2) experimentally revealed through the present study of the photophysical parameters as a function of T. The energy barriers experimentally determined and reported in Table 6 for NF (higher) and AF (lower) and the absence of an activated ISC in the case of AA and NA are in agreement with the previously calculated ¹(π,π*)–³(n,π*) gaps (>13 kcal mol⁻¹ for NA/AA, 12 kcal mol⁻¹ for NF and 2.8 kcal mol⁻¹ for AF, in gas phase-TDDFT/B3LYP calculations after HF optimization of the ground state).⁸

However, the results of the spectra and the photophysical parameters obtained for AF in Tol and DCM (in particular the opposite trend of ϕ_F with T, significantly increasing and decreasing at low T in Tol and DCM, respectively, see Table 2) prompted us to perform new quantum mechanical DFT calculations for this compound taking into account the solvation effects by means of the conductor-like polarizable continuum model CPCM.²⁰ The achievements are reported in Tables S1, S2 and Fig. S17, S18 (ESI†). The results predicted the observed red-shift of the absorption spectrum and the increase of the ¹(π,π*)–³(n,π*) energy gap on going from Tol (5.3 kcal mol⁻¹) to the more polar solvent DCM (8.3 kcal mol⁻¹), that makes the activated ISC less efficient.

Moreover, previous fs transient absorption measurements evidenced the presence of a ¹CT* state for AF in DCM (produced by the charge transfer process from ¹LE*) not observed in Tol. Therefore, the increase of ϕ_F at low T measured in Tol is explained by the decreased efficiency of the activated ISC and IC processes, while ³(n,π*) becomes inaccessible from the stabilized ¹LE* in the more polar DCM, the activated ISC being no longer competitive with the fast ICT. Moreover, the decrease of ϕ_F at low T measured in DCM can be due to the increase of dielectric constant (from 9.08 at 293 K to 12.57 at 200 K) that causes an increased CT character of S₁, the latter implying an increase of k_IC (by roughly three times, see Table 5) and a decrease of k_F (by ca. 40%, see Table 2).

A relaxed potential energy surface scan was performed by HF optimization in the case of AF in Tol, to confirm the eventual presence of sets of conformations for these flexible molecules at the investigated temperatures and the expected reduction of the energy gap between the S₁ (π,π*) state and the upper excited singlet state of n,π* nature (Δ[S₁–S_n]) in the more twisted geometries, as previously reported for analogous compounds.^12,13 The results obtained are reported in Fig. S19 and Table S3 (ESI†). In fact, the relative stability of AF twisted conformations (shown in Fig. S19, ESI†), originated by rotation around the single bond adjacent to the benzaldehyde group (see Chart S1, ESI†), and the strong reduction of Δ[S₁–S_n] observed for the twisted conformations (the two states becoming practically isoenergetic when θ = 90 deg, Table S3, ESI†), are in agreement with the presence of multiple species with slightly distorted geometries in the investigated 293–220 K range. These distorted conformations mainly contribute to the S₁ deactivation through IC according to Lim's proximity effect.

Up-conversion measurements: evidence of dual fluorescence for AF in DCM. Femtosecond fluorescence up-conversion measurements were carried out for the four compounds in the same solvents where the temperature effect on the fluorescence and triplet properties was investigated (NA and AA in CH; NF in Tol; AF in Tol and DCM; the results are shown in Fig. 4 and Fig. S20–S23, ESI† and summarized in Table 7). Experiments were performed by employing a 400 nm laser excitation, which leads to the population of S₁ with a certain excess of vibrational energy, particularly in the case of the anthracene derivatives (see Fig. S1, ESI†). When the signal/noise ratio was good enough that nice fluorescence kinetics could be acquired in a wide spectral range, time resolved emission spectra were reconstructed (panel B of the figures) and Global/Target analysis allowed us to obtain information not only about the lifetimes but also about the Species Associated Spectra (SAS) of the transients resulting from the best fitting of the data (panel C of the figures).

Fig. 4 Fluorescence up-conversion (λ_exc = 400 nm) of AA in CH (left) and AF in DCM (right): (A) contour plot of the experimental data and (B) time resolved emission spectra. Inset: Decay kinetics recorded at meaningful wavelengths. (C) SAS of the decay components obtained by Target Analysis.

Table 7 Excited state lifetimes of the transients revealed by femtosecond up-conversion measurements (τ_UC). Those derived by transient absorption spectroscopy (τ_TA) are also reported for comparison

Compound	Solvent	τ UC/ps	τ TA /ps	Assignment
a From ref. 8.
NA	CH	0.45	0.85	VC
		11	8.2	VC
		985	>2100	S1(LE)
		Rest		S1(LE)
AA	CH	0.5	0.73	VC
		14	11	VC
		1100	1360	S1 (LE)
		Rest		S1 (LE)
AF	Tol	1.2		Solv.
		18	10	VC
		270	220	S1 (LE)
			Rest	T1
AF	DCM	0.69	0.47	Solv.
		3.8	2.0	Solv.
		100	83	S1 (LE)
		995	1200	S1 (ICT)
NF	Tol	0.16	0.68	Solv.
		7.8	8.4	Solv. + VC
		48	45	S1 (LE)
			Rest	T1

---

The kinetics acquired for AA in CH (see Fig. 4 left) show a fast rise followed by a slow decay (not completed in the investigated time window) at all the investigated wavelengths. The time resolved spectra are structured and do not exhibit any significant shift in time. The best fitting of these data was obtained when three exponential components are considered (characterized by lifetimes of 0.50, 14, and 1100 ps) together with the rest that corresponds to a longer-living transient not decaying in the considered temporal range. The first two lifetimes are in fair agreement with those previously obtained by femtosecond transient absorption measurements and are probably related to vibrational cooling, VC (see ref. 8 and Table 7). In fact, CH is known to be a non-polar solvent where no spectral relaxation due to solvation can be observed¹⁸ and previous studies,¹⁹ carried out in this solvent, suggested that the dynamics observed on a ∼10 ps time scale can be associated with vibrational relaxation processes. The two longer living transients can be both assigned to the decay of the locally excited (LE) S₁ state: they show indeed the same SAS. In fact, the deactivation of the lowest excited singlet state in the highly concentrated solutions used for the up-conversion measurements of these compounds (characterized by a small Stokes shift) may be complicated by self-absorption of the emitted light (see Fig. S24, ESI†). Analogous results were obtained from the up-conversion study of the other anthracene derivative, NA in CH (Fig. S20, ESI† left).

The ultrafast fluorescence data for AF in Tol (see Fig. S20, ESI† right) pointed out the occurrence of solvent relaxation of this non-dipolar solvent (1.2 ps, evidenced by the fast rise/decay dynamics in the kinetics acquired at different wavelengths) followed by vibrational cooling (18 ps) and decay of the LE S₁ state (270 ps). Differently, a much more important spectral evolution in time was revealed for AF in the more polar DCM (see Fig. 4, right). The rise/decay dynamics in the recorded kinetics takes place in this case in a much longer time scale. The Global Fitting of these data revealed the presence of four exponential components whose lifetimes are 0.69, 3.8, 100 and 995 ps and whose assignment was supported by the Time Resolved Area Normalised Emission Spectra (TRANES) analysis (see Fig. 5).^20–22 During the first 10 ps after photoexcitation (when the 0.69 and 3.8 ps components dominate the excited state deactivation, see Fig. S23, ESI†) a continuous shift is observed in the TRANES analysis indicating that solvation dynamics is occurring in this temporal interval. In contrast, an isoemissive point at 525 nm is observed in the 10–260 ps range, when the most abundant transients are the 100 and 995 ps components (see Fig. S23, ESI†). These species are therefore assigned to distinct local emissive minima on the potential energy surface of S₁: the LE state produced upon light absorption and the ICT state stabilized in a polar solvent such as DCM. These results give a deep insight about the origin of the shoulder detected in the stationary fluorescence spectra of AF in DCM at low temperature (see Fig. S2, ESI†) that the finding can be now rationalized considering the dual emission from the LE and ICT states populated for this compound in a mildly polar solvent.^23,24

Fig. 5 TRANES spectral evolution of AF in DCM recorded between 0.85 and 262 ps after the laser pulse excitation.

Experimental

The compounds investigated are shown in Chart 1. They were re-synthesized following previously reported procedures.^1,8 The solvents used (Fluka, spectroscopic grade) were cyclohexane (CH), methylcyclohexane (MCH), toluene (Tol) and dichloromethane (DCM).

The absorption measurements were carried out using a Perkin-Elmer Lambda 800 spectrophotometer. The fluorescence spectra, corrected for the instrumental response, were recorded using a SPEX Fluorolog-2 F112AI spectrofluorometer. The fluorescence lifetimes (τ_F) were measured on an Edinburgh Instrument 199S spectrofluorometer equipped with a LED source (centered at 370 nm in the case of fluorene-derivatives and at 455 nm for the anthracene-derivatives), using the single-photon counting method with a resolution time of 0.2 ns. The values reported in the tables are averages of three independent experiments with a mean deviation of ≅10%. The kinetic traces were also analyzed by means of the Maximum Entropy Method (MEM) using MEMEXP Software available online.²⁵ The method, which is based on the simultaneous maximization of the entropy (S) and the minimization of the Poisson deviance (C), gives back the distribution of lifetimes, as well as the number of transients (family of distributions around an average time) needed to describe the spectroscopic kinetics. These distributions describe the probability of having different lifetimes. The widths of the distributions are sensitive to the heterogeneous microenvironment and temperature a molecule can experience in the sample.²⁶ Further details of MEM analysis are given in the ESI.†

The triplet properties were measured by a previously described laser flash photolysis setup²⁷ exciting at 355 nm using the third harmonic of a Continuum (Surelite II) Nd:YAG laser (pulse width 7 ns and laser energy <5 mJ per pulse).

A cryostat (Oxford Instruments DN 1704) was used to control temperature in the 195–358 K range. For the measurements of the quantum yields of the photophysical processes as a function of temperature, the values at 293 K (from ref. 8) were used as reference, taking into account the changes in absorbance and refractive index with temperature. For all measurements, the solutions were deoxygenated by purging with nitrogen.

A detailed description of the experimental setup for ultrafast measurements was previously reported.^28,29 The 400 nm excitation pulses of ca. 60 fs were generated by an amplified Ti:Sapphire laser system (Spectra Physics, Mountain View, CA). In the up-conversion set-up (Halcyone, Ultrafast Systems) the 400 nm pulses excite the sample whereas the remaining fundamental laser beam at 800 nm plays the role of the “optical gate” after passing through a delay line. The temporal resolution of the up-conversion equipment is about 250 fs, whereas the spectral resolution is 5 nm. Ultrafast spectroscopic data were fitted by Global and Target Analysis using Glotaran software.³⁰

Quantum-mechanical calculations were carried out using the Gaussian 09 package.³¹ Density functional theory (DFT) based on the CAM-B3LYP method was used.^32,33 In particular, for the ground state of the substrates HF optimization was employed, while transition energies and probabilities of the excited singlet states were obtained by TD-DFT CAM-B3LYP/6-31G(d). Tol and DCM solvation effects were included in the calculations by means of the conductor-like polarizable continuum model CPCM.³⁴ A relaxed potential energy surface scan was performed by HF to investigate the relative stability of conformations originated by rotation around the single bond adjacent to the aromatic moieties.

Conclusions

The temperature effect on the photophysical properties confirmed the presence of an activated ISC to an upper triplet state only in the case of fluorene derivatives in non-polar Tol, as previously hypothesized on the basis of TDDFT quantum-mechanical calculations and ultrafast absorption spectroscopy. In fact, replacing the fluorene aromatic structure with an anthracene moiety, the energy of the (π,π*) states is significantly lowered (as also indicated by the bathochromic shift of the UV-Vis absorption spectrum of anthryl compounds compared with those of fluorene-derivatives) while the (n,π*) states are much less affected. Thus, the energy of the ³(n,π*) state of NA and AA is significantly higher than that of the lowest excited ¹(π,π*) state in agreement with no activated ISC found to occur for the latter.

An unexpected activated IC was also found to be operative in the deactivation of S₁ competing with fluorescence and ISC, in the case of fluorene-derivatives in Tol. It was also present in the photorelaxation of AA in Tol above room T. With the help of the distributions of lifetimes (larger at high T and narrower at low T) obtained by MEM analysis of the fluorescence decay data, this behavior was explained by the presence in fluid solution at high T of different sets of twisted conformations (responsible for the increased IC yield) while more planar and stable forms (mainly deactivated through fluorescence) are selected at low T.

An interesting opposite trend of ϕ_F with T was observed for AF on going from Tol to the more polar DCM and explained on the basis of the increase of the ¹(π,π*)–³(n,π*) energy gap predicted by DFT calculations in the more polar solvent (that makes the activated ISC less efficient) and the role of ICT states in solvents of higher dielectric constant. In fact, the increase of ϕ_F at low T measured in Tol is explained by the decreased efficiency of the activated ISC and IC processes, while the decrease of ϕ_F at low T measured in DCM is due to the increase of the dielectric constant that causes an increased CT character of S₁, the latter implying an increase of k_IC and a decrease of k_F.

The presence of dual fluorescence of AF in DCM, suggested by the appearance of a hypsochromic shoulder in its emission spectrum at low T, was nicely confirmed by femtosecond up-conversion measurements that allowed reconstructing time resolved spectra revealing different emission bands. The detection of an isoemissive point in the TRANES analysis unequivocally indicated the presence of two excited states (¹LE* and ¹ICT*) for AF in DCM.

Acknowledgements

The authors acknowledge support from the Italian “Ministero per l'Università e la Ricerca Scientifica e Tecnologica”, MIUR (Rome, Italy) under the FIRB “Futuro in Ricerca” 2013, no. RBFR13PSB6 and PRIN ‘‘Programmi di Ricerca di Interesse Nazionale’’ 2010–2011, no. 2010FM738P.

References

1. R. Flamini, I. Tomasi, A. Marrocchi, B. Carlotti and A. Spalletti, J. Photochem. Photobiol., A, 2011, 223, 140.
2. A. Marrocchi, A. Spalletti, S. Ciorba, M. Seri, F. Elisei and A. Taticchi, J. Photochem. Photobiol., A, 2010, 211, 162.
3. For recent representative examples see: (a) M. Inagaki and F. Kang, J. Mater. Chem. A, 2014, 2, 13193; (b) A. Broggi, I. Tomasi, L. Bianchi, A. Marrocchi and L. Vaccaro, ChemPlusChem, 2014, 79, 486; (c) C. Kastner, S. Rathgeber, D. A. M. Egbe and H. Hoppe, J. Mater. Chem. A, 2013, 1, 3961; (d) S. Rochat and T. M. Swager, ACS Appl. Mater. Interfaces, 2013, 5, 4488; (e) E. Bartollini, M. Seri, S. Tortorella, A. Facchetti, T. J. Marks, A. Marrocchi and L. Vaccaro, RSC Adv., 2013, 3, 9288; (f) A. Marrocchi, I. Tomasi and L. Vaccaro, Isr. J. Chem., 2012, 52, 41; (g) A. Operamolla, R. Ragni, O. H. Omar, G. Iacobellis, A. Cardone, F. Babudri and G. Farinola, Curr. Org. Synth., 2012, 9, 764; (h) T. L. Andrew and T. M. Swager, J. Polym. Sci., Part B: Polym. Phys., 2011, 49, 476; (i) A. Operamolla and G. Farinola, Eur. J. Org. Chem., 2011, 423; (j) F. Silvestri and A. Marrocchi, Int. J. Mol. Sci., 2010, 11, 1471.
4. R. Flamini, B. Carlotti, A. Spalletti and A. Marrocchi, Org. Photonics Photovolt., 2014, 2, 1.
5. R. Flamini, A. Marrocchi and A. Spalletti, Photochem. Photobiol. Sci., 2014, 13, 1031.
6. (a) B. Carlotti, A. Spalletti, M. Šindler-Kulyk and F. Elisei, Phys. Chem. Chem. Phys., 2011, 13, 4519; (b) I. Kikaš, B. Carlotti, I. Škorić, M. Šindler-Kulyk, U. Mazzucato and A. Spalletti, J. Photochem. Photobiol., A, 2012, 244, 38; (c) B. Carlotti, I. Kikaš, I. Škorić, A. Spalletti and F. Elisei, ChemPhysChem, 2013, 14, 970; (d) B. Carlotti, F. Elisei, U. Mazzucato and A. Spalletti, Phys. Chem. Chem. Phys., 2015, 17, 14740; (e) B. Carlotti, G. Consiglio, F. Elisei, C. G. Fortuna, U. Mazzucato and A. Spalletti, J. Phys. Chem. A, 2014, 118, 3580; (f) B. Carlotti, E. Benassi, V. Barone, G. Consiglio, F. Elisei, A. Mazzoli and A. Spalletti, ChemPhysChem, 2015, 16, 1440.
7. For representative examples see: (a) P. Sonar, S. P. Singh, Y. Li, Z. Ooi, T. Ha, I. Wong, M. S. Soh and A. Dodabalapur, Energy Environ. Sci., 2011, 4, 2288; (b) C. Aurisicchio, D. Bonifazi, B. Ventura and A. Barbieri, J. Phys. Chem. C, 2009, 113, 17927; (c) Y. Gong, X. Guo, S. Wang, H. Su and A. Xia, J. Phys. Chem. A, 2007, 111, 5806.
8. B. Carlotti, R. Flamini, A. Spalletti, A. Marrocchi and F. Elisei, ChemPhysChem, 2012, 13, 724.
9. B. Carlotti, R. Flamini, I. Kikaš, U. Mazzucato and A. Spalletti, Chem. Phys., 2012, 407, 9.
10. B. Carlotti, F. Elisei and A. Spalletti, Phys. Chem. Chem. Phys., 2011, 13, 20787.
11. K. Suzuki, A. Demeter, W. Kühnle, E. Tauer, K. A. Zachariasse, S. Tobita and H. Shizuka, Phys. Chem. Chem. Phys., 2000, 2, 981 , and references therein.
12. G. Marconi, G. Bartocci, U. Mazzucato, A. Spalletti, F. Abbate, L. Angeloni and E. Castellucci, Chem. Phys., 1995, 196, 383.
13. G. Marconi, G. Bartocci, U. Mazzucato, A. Spalletti, F. Abbate, L. Angeloni and E. Castellucci, Am. Inst. Phys., Conf. Proc., 1996, 364, 175 and references therein.
14. E. C. Lim, J. Phys. Chem., 1986, 90, 6770.
15. R. Flamini, A. Marrocchi and A. Spalletti, ChemPlusChem, 2015, 80, 1045.
16. P. L. Gentili, F. Ortica, A. Romani and G. Favaro, J. Phys. Chem. A, 2007, 111, 193.
17. E. Benassi, C. Cappelli, B. Carlotti and V. Barone, Phys. Chem. Chem. Phys., 2014, 16, 26963.
18. L. Reynolds, J. A. Gardecki, S. J. V. Frankland, M. L. Horng and M. Maroncelli, J. Phys. Chem., 1996, 100, 10337.
19. M. L. Horng, J. A. Gardecki, A. Papazyan and M. Maroncelli, J. Phys. Chem., 1995, 99, 17311.
20. N. Periasamy and A. S. R. Koti, Proc. Indiana Acad. Sci., 2003, 69, 41.
21. L. Gutierrez-Arzaluz, C. A. Guarin, W. Rodrigez-Cordoba and J. Peon, J. Phys. Chem. B, 2013, 117, 12175.
22. M. Park, D. Im, Y. Ho Rhee and T. Joo, J. Phys. Chem. A, 2014, 118, 5125.
23. B. Carlotti, A. Cesaretti, C. G. Fortuna, A. Spalletti and F. Elisei, Phys. Chem. Chem. Phys., 2015, 17, 1877.
24. E. Benassi, B. Carlotti, M. Segado, A. Cesaretti, A. Spalletti, F. Elisei and V. Barone, J. Phys. Chem. B, 2015, 119, 6035.
25. (a) P. J. Steinbach, R. Ionescu and C. R. Matthews, Biophys. J., 2002, 82, 2244; (b) P. J. Steinbach, J. Chem. Inf. Comput. Sci., 2002, 42, 1476.
26. (a) M. Penconi, P. L. Gentili, G. Massaro, F. Elisei and F. Ortica, Photochem. Photobiol. Sci., 2014, 13, 48; (b) P. L. Gentili, C. Clementi and A. Romani, Appl. Spectrosc., 2010, 64, 923; (c) P. L. Gentili, Dyes Pigm., 2014, 110, 235.
27. H. Gorner, F. Elisei and G. G. Aloisi, J. Chem. Soc., Faraday Trans., 1992, 88, 29.
28. B. Carlotti, E. Benassi, A. Cesaretti, C. G. Fortuna, A. Spalletti, V. Barone and F. Elisei, Phys. Chem. Chem. Phys., 2015, 17, 20981.
29. A. Cesaretti, B. Carlotti, P. L. Gentili, C. Clementi, R. Germani and F. Elisei, J. Phys. Chem. B, 2014, 118, 8601.
30. J. J. Snellenburg, S. Laptenok, R. Seger, K. M. Mullen and I. H. M. van Stokkum, J. Stat. Soft., 2012, 49, 1.
31. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian 09, Revision B.01, Gaussian, Inc., Wallingford, CT, 2010.
32. T. Yanai, D. P. Tew and N. C. Handy, Chem. Phys. Lett., 2004, 393, 51.
33. T. Yanai, R. J. Harrison and N. C. Handy, Mol. Phys., 2005, 103, 413.
34. V. Barone and M. Cossi, J. Phys. Chem. A, 1998, 102, 1995.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5cp06025g

---