ARTICLE pubs.acs.org/JPCA
Photophysical Study of Bis(naphthalimide)-Amine Conjugates: Toward Molecular Design of Excimer Emission Switching R. Ferreira,† C. Baleiz~ao,‡ J. M. Mu~noz-Molina,† M. N. Berberan-Santos,*,‡ and U. Pischel*,† † ‡
Centro de Investigacion en Química Sostenible, Universidad de Huelva, Campus de El Carmen s/n, E-21071 Huelva, Spain Centro de Química-Física Molecular and Institute of Nanoscience and Nanotechnology, Instituto Superior Tecnico, Universidade Tecnica de Lisboa, 1049-001 Lisboa, Portugal
bS Supporting Information ABSTRACT: The fluorescence properties of two bis(1,8-naphthalimides) with amino-containing spacers are investigated, giving special emphasis to the observation of excimer emission. It is found that a minor elongation of the spacer by two methylene units gives rise to a quantitative shut-down of the broad and red-shifted excimer emission. Furthermore, a switching of this emission is established through manipulation of a photoinduced electron transfer process, which involves the amino spacer. Protons as well as protic solvents lead to substantial excimer emission with lifetimes of 12 to 27 ns. The excimer quantum yield takes a maximum value of Φf = 0.07 (acetonitrile with 1 equiv trifluoroacetic acid). The increased virtual Stokes shifts (ca. 150 nm) as compared to the fluorescence of monomeric 1,8-naphthalimides are an alternative approach to obtain colored, significant, and long-lived fluorescence from these chromophores. As an additional excited state pathway, the occurrence of homo-F€orster resonance energy transfer (homo-FRET) is established by fluorescence polarization measurements and calculation of the corresponding critical F€orster radius (R0 ca. 13 Å). The average interchromophore distance between the naphthalimides is estimated as 7.5 Å and 9.5 Å for the dyad with the shorter and the longer spacer, respectively. These observations and the absence of a rise time component for excimer emission are in agreement with the formation of a “loose” ground state dimer, which upon excitation undergoes a fast geometrical adjustment to the excimer structure where the chromophores are at contact distance.
’ INTRODUCTION 1,8-Naphthalimides have attracted general interest for the design of molecular fluorescent devices, for example, for chemosensing, 1-10 switching,11,12 and logic operations.13-15 The parent fluorophore motif is characterized by a rather low fluorescence quantum yield.16,17 However, its strong electronaccepting properties, and thus the possibility of photoinduced electron transfer (PET), have been shown to constitute an advantage for the above-mentioned applications.9,13,14,18,19 Furthermore, electron-transfer processes are also the basis for the use of naphthalimides as photoactive reagents for nucleotide oxidation.20 On the other hand, PET with 4-amino-1,8-naphthalimides has been demonstrated to depend sensitively on the regiochemistry of substitution with the corresponding electron donor,21-23 which imposes restrictions for the molecular design. Beside electron transfer, electronic energy transfer,14,24 excimer formation,18,25-28 and exciplex formation29,30 have been found to contribute equally to the rich photophysics of the parent chromophore. A few studies were dedicated to the intramolecular excimer formation of bis(1,8-naphthalimide) conjugates with r 2011 American Chemical Society
flexible oligomethylene spacers.25,28 It was found that an increasing spacer length results in a blue shift of the excimer emission maximum.28 Further, the emission quantum yield of the excimer was observed to be largest for six methylene units. The general trends regarding the dependence on the spacer length and conformation are in agreement with the numerous studies reported on the formation of excimers in dyads with terminal aromatic hydrocarbon moieties.31-35 Two other studies dealt with the excimer emission of the bis(1,8-naphthalimide)spermine conjugate.18,27 The formation of excimers of the naphthalimide chromophore turns out to be also of interest from another and so far rather neglected point of view. Namely, much effort have been made to shift the emission of naphthalimides to the visible wavelength range (specifically, λ > 450 nm). This can be achieved by careful molecular design via the fine-tuned placement of electron-donating Received: November 2, 2010 Revised: December 30, 2010 Published: January 26, 2011 1092
dx.doi.org/10.1021/jp110470h | J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A Chart 1. Molecular Structures of Dyads 1 and 2 and Previously Reported Dyads with Excimer Emission 3 and 4(n)
ARTICLE
spacer length, which connects the terminal naphthalimide chromophores, leads to the total absence of excimer emission. Upon comparison of these results with classical intramolecular excimer studies, where generally a rather gradual decrease of excimer formation with increasing spacer length was reported, this clearcut on-off situation is rather surprising.
’ EXPERIMENTAL METHODS
and -accepting substituents. Thus, it is generally realized that the introduction of an amino group at the aromatic 4-position of the parent chromophore yields a strongly red-shifted emission, which originates from the charge-transfer character of the involved excited state. Heagy and co-workers36,37 came up with a different and quite elegant approach, which consists in the introduction of electron-donating and -accepting substituents at the aromatic core of the imide and at an N-aryl moiety. This design allows the prediction of dual fluorescence (locally excited state and charge-transfer emission) via a balanced “seesaw” model. As we will show in this work, an alternative worth being considered in the context of generating visible fluorescence of the parent 1,8-naphthalimide is the rational design of excimeremitting dyads. This leads ultimately to large virtual Stokes shifts (>100 nm), measured by the difference between absorption and excimer emission maxima. An additional advantage for applications like chemosensing is the generally longer lifetime of excimers (tens of nanoseconds)18,27 as compared to the subnanosecond values measured for 1,8-naphthalimide monomer fluorescence.17 Herein we are especially interested in the possibility of excimer formation in bis(1,8-naphthalimides) with an amino-containing spacer. This molecular design is adopted in order to open the possibility of switching excimer emission. This is anticipated to be possible through protonation or hydrogen bonding of the amino nitrogen, and thus by addressing the competition between PET and excimer formation and by induction of conformational changes of the spacer. A related photophysical scenario was established for naphthalene-polyamine and pyrene-polyamine conjugates.33-35,38,39 We decided to study in detail the bichromophoric dyads 1 and 2 (see structures in Chart 1), which contain a spacer with an integrated secondary amino group. As will be shown in this work, already a minimal variation of the
Materials. 1,8-Naphthalic anhydride, bis(2-aminoethyl)amine, and bis(3-aminopropyl)amine were from Fluka in highest available purity (>98%). All solvents were of spectroscopic or HPLC grade. General procedure40 for the preparation of dyads 1 and 2: 1,8Naphthalic anhydride (1.0 mmol) was refluxed for 4 h with 0.5 mmol of bis(2-aminoethyl)amine or bis(3-aminopropyl)amine in 50 mL of ethanol. The solution was allowed to cool down to room temperature, upon which a precipitate formed. This was collected by filtration, washed with copious amounts of cold ethanol, and dried under vacuum. If necessary, the compounds were further purified by recrystallization from ethanol. Bis(N-ethyl-1,8-naphthalimidyl)amine (1). 1H NMR (CDCl3, 400 MHz): δ 3.10 (t, J = 6.2 Hz, 4H), 4.33 (t, J = 6.2 Hz, 4H), 7.66 (t, J = 7.8 Hz, 4H), 8.17 (d, J = 7.8 Hz, 4H), 8.39 ppm (d, J = 6.2 Hz, 4H). HRMS (CI): calcd for C28H22N3O4 (1 þ Hþ), 464.1610; found, 464.1624. Bis(N-propyl-1,8-naphthalimidyl)amine (2). 1H NMR (CDCl3, 400 MHz): δ 1.96 (q, J = 7.0 Hz, 4H), 2.73 (t, J = 7.0 Hz, 4H), 4.26 (t, J = 7.0 Hz, 4H), 7.74 (t, J = 7.8 Hz, 4H), 8.19 (d, J = 7.8 Hz, 4H), 8.58 ppm (d, J = 6.2 Hz, 4H). HRMS (CI): calcd for C30H26N3O4 (2 þ Hþ), 492.1923; found, 492.1929. Spectroscopic Measurements. Absorption measurements were performed with a UV-1603 spectrometer from Shimadzu. Fluorescence spectra were recorded with a Cary Eclipse instrument from Varian. The measurements were done at ambient temperature in 1 cm quartz cuvettes with diluted (5-10 μM) airequilibrated dyad solutions. The fluorescence spectra were recorded with an excitation wavelength of 330 or 332 nm and corrected with the corresponding instrument response function. Quantum yields were calculated relative to the corrected emission spectrum of an optically matched solution of N-propyl-1,8naphthalimide (available from an earlier project;13 Φf = 0.016, in aerated acetonitrile).18 The values are corrected for differences in the refractive index of the solvent of the reference compound and the actual solvent of the dyad.41 In order to obtain separate values for the quantum yields of monomer and excimer emission, the fluorescence spectrum of Npropyl-1,8-naphthalimide was measured in each specific solvent. Under the reasonable assumption of its correspondence to pure monomer emission, this spectrum was normalized to the intensity of the monomer emission maximum of dyad 1 and then subtracted from the global emission of the dyad. This procedure yielded the spectrum of the excimer and therefore enabled the determination of its quantum yield. The monomer emission quantum yield was then calculated with this value and the global emission quantum yield under the assumption of the validity of the relation Φf,global = Φf,monomer þ Φf,excimer. Time-resolved picosecond fluorescence measurements were performed using the single-photon timing method with laser excitation. The setup consisted of a mode-locked Coherent Innova 400-10 argon-ion laser that synchronously pumped a 1093
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A
ARTICLE
cavity-dumped Coherent 701-2 dye [rhodamine 6G or 4-dicyanomethylene-2-methyl-6-(p-dimethylamino)styryl-4H-pyran, DCM] laser, delivering fundamental or frequency-doubled 5 ps pulses at a repetition rate of 3.4 MHz. Alternatively, a pulse picked Spectra-Physics Tsunami Ti:sapphire laser, which was pumped by a Spectra-Physics Millenia Xs Nd:YVO4 diodepumped laser, was used as excitation source. This configuration delivered 100 fs frequency-doubled pulses at a repetition rate of 4 MHz. The intensity decay measurements were done by alternate collection of impulse and decay, with the emission polarizer
positioned at the magic angle. The anisotropy decay measurements were done by collection of sequences of impulse and decays with the emission polarizer set at the vertical and horizontal positions. The instrument impulse was recorded slightly away from the excitation wavelength with a scattering suspension. For the decays, a cutoff filter was used, effectively removing all excitation light. Excitation/emission wavelengths were 335/385 and 335/490 nm for monomer and excimer, respectively. The emission signal passed through a depolarizer and a Jobin-Yvon HR320 monochromator with a grating of 100 lines/mm and was detected with a Hamamatsu 2809U-01 microchannel plate photomultiplier. The instrument response function had an effective full width at half-maximum (fwhm) of 35 ps. The fitting of the decay traces was done by deconvolution, with the instrument impulse taken into account.
’ RESULTS Absorption Spectra. The absorption and fluorescence properties of dyads 1 and 2 are investigated in the following solvents: 2,2,2-trifluoroethanol, water, methanol, ethanol, acetonitrile, and a 1/1 (v/v) mixture of methanol and acetonitrile. The normalized spectra of dyad 1 in each solvent are shown in Figure 1. For both dyads the typical naphthalimide UV absorption between 300 and 400 nm with a maximum at ca. 331-334 nm is observed (Table 1). Exceptions are 2,2,2-trifluoroethanol and water, for which slightly red-shifted maxima are measured (Δλ ca. 10 nm). These results are in line with the value (λmax = 331 nm in acetonitrile) reported for N-propyl-1,8-naphthalimide, a compound that will be considered as a model throughout this study.18 The molar absorption coefficients (ε) for 1 and 2 are found between 19000 and 26000 M-1 3 cm-1, depending on the solvent medium. For N-propyl-1,8-naphthalimide, a value of 12 700 M-1 3 cm-1 was
Figure 1. Absorption and fluorescence spectra of dyad 1 (10 μM) in six solvent media: TFE = 2,2,2-trifluoroethanol; EtOH = ethanol; water; MeOH = methanol; MeOH/MeCN = 1/1 mixture of methanol and acetonitrile; MeCN = acetonitrile (in the presence of 1 equiv of trifluoroacetic acid, TFA). The spectra are normalized to 1 at λmax of absorption and short-wavelength emission.
Table 1. Absorption and Fluorescence Properties of Dyads 1 and 2 in Different Aerated Solvents at Room Temperature (T = 295 K) short-wavelength emission ηb (mPa 3 s)
εc
Rd
λabs (nm)
ε (M-1 3 cm-1)
TFE
1.995
26.67
1.51
340
water
1.002
80.16
1.17
MeOH
0.593
32.66
EtOH MeCN
1.200 0.345
24.55 35.94
long-wavelength emission
λf (nm)
Φfe/10-2
τav (ns)
λf (nm)
Φfe/10-2
τav (ns)
25 200
392
2.5
0.53
494
3.5
22.2
343
f
397
2.1
2.11
505
2.5
26.8
0.93
333
20 600
386
1.0
0.52
491
0.5
24.4
0.83 0.19
333 331
26 200 24 000
386 378
1.9 0.1
0.54 0.48
491
0.7
25.2
MeCN þ TFA
333
24 400
385
3.7
0.35
487
7.0
11.9
MeCN/MeOH
333
23 400
384
0.5
0.33
489
0.4
25.6
487
1.5
g
solventa
Dyad 1
Dyad 2 TFE
1.995
26.67
1.51
343
18 500
393
21.0
1.52
water
1.002
80.16
1.17
345
f
397
8.3
1.63
MeOH
0.593
32.66
0.93
334
23 000
387
5.3
0.70
EtOH MeCN
1.200 0.345
24.55 35.94
0.83 0.19
335 332
22 500 20 800
385 378
4.0 0.2
0.70 1.07
MeCN þ TFA
333
20 700
386
11.2
0.76
MeCN/MeOH
334
24 000
386
5.6
0.66
a
TFE = 2,2,2-trifluoroethanol; MeOH/MeCN = 1/1 mixture of methanol and acetonitrile; MeCN = acetonitrile (þTFA, in the presence of 1 equiv of trifluoroacetic acid). b Viscosity at 20 C (except acetonitrile, 25 C).46 c Dielectric constant at 25 C (except water and 2,2,2-trifluoroethanol, 20 C).46 d Kamlet-Taft R parameter.50 e Measured with N-propyl-1,8-naphthalimide as reference (Φf = 0.016 in aerated acetonitrile).18 f Not determined, because of low solubility of 1 and 2 in water. g Not measured. 1094
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A
Figure 2. Corrected fluorescence spectra of 1 (10 μM) in acetonitrile in the absence of TFA (a) and in the presence of 1 equiv of TFA (b).
Figure 3. Absorption spectrum (abs, - 3 -), corrected fluorescence spectrum (fluo, ;), phosphorescence spectrum (phos, ---), and corrected excitation spectra (a, λobs = 390 nm; b, λobs = 485 nm; 3 3 3 ) of 1 (10 μM) in 2,2,2-trifluoroethanol or in tetrahydrofuran (phos). The excitation spectra are horizontally displaced for better visibility.
reported in acetonitrile,18 approximately half the value of the bichromophoric dyads under study. Emission Properties. While both dyads, 1 and 2, show very comparable absorption spectra, they behave quite differently with respect to their fluorescence properties (see Table 1). On the one hand, for bis(naphthalimide) 1, two emission bands are noted in all solvents (see Figure 1), except for acetonitrile in the absence of acid (see Figure 2). In detail, a short-wavelength band (λmax between 378 and 397 nm), typically observed for 1,8naphthalimide derivatives, is accompanied by a broad and redshifted band (λmax between 487 and 505 nm). Similar spectral signatures were reported for the bis(1,8-naphthalimides) 3 and 4(n) (see structures in Chart 1).18,27,28 The excitation spectra obtained by monitoring the emission maxima of each of the two bands coincide practically with the absorption spectrum (see Figure 3). On the other hand, the fluorescence measurements of dyad 2 lead to the observation of only the short-wavelength band (λmax between 378 and 397 nm), except for water, where an additional red-shifted band with minor relative intensity and a maximum at 487 nm is noted. Importantly, all measurements are performed in dilute solutions (ca. 5-10 μM), and therefore it can be reasonably assumed that the observed fluorescence phenomena originate from intramolecular processes.42
ARTICLE
The maxima of the short- and long-wavelength fluorescence bands of 1 and the short-wavelength band of 2 are solventdependent. In comparison to acetonitrile, protic solvents such as alcohols and water cause bathochromic shifts between 10 and 20 nm, being most pronounced for water and 2,2,2-trifluoroethanol. This is in accordance with the effects noted for the UV absorption (see above). Furthermore, similar trends were reported earlier for the fluorescence of the related derivatives 3 and 4(n) in water and acetonitrile/water mixtures (1/1 v/v), respectively.18,28 The fluorescence quantum yields (Φf) in neat solvents (2,2,2trifluoroethanol, water, methanol, ethanol, and acetonitrile) are listed in Table 1 for the short-wavelength emission (1 and 2) and the long-wavelength emission (mainly 1). A closer look at these values reveals that the short-wavelength fluorescence quantum yields measured for 2 are generally higher than those for 1. For both dyads, Φf of this emission is highest in 2,2,2trifluoroethanol or water and smallest in acetonitrile. The longwavelength emission of 1 follows a similar trend with respect to the solvent dependence: the highest Φf values are obtained in 2,2,2-trifluoroethanol and water. However, in the case of 2, the quantum yield in water, the only solvent for which a minor redshifted long-wavelength emission is observed for this dyad, is smaller than in the case of 1. It is also instructive to compare the ratio of Φf for the two emissions of 1 in the different solvents. Again in 2,2,2-trifluoroethanol and water, the long-wavelength emission is favored over the short-wavelength emission (ratio of ca. 1.4). It is noteworthy that, for 2 in water, the same ratio is just 0.2. The fluorescence decay curves are found to follow multiexponential kinetics but often with a dominant component. Conformational issues of the spacer and their relations with distance-dependent quenching mechanisms (such as PET, see below) are assumed to lead to this observation. However, for our purposes it is enough to report average fluorescence lifetimes (τav, eq 1). The values for the short-wavelength emission are found around 0.5 ns for 1, except for water, where τav is 2.1 ns. For dyad 2, average lifetimes between 0.7 and 1.6 ns result, being highest for water and 2,2,2-trifluoroethanol. The long-wavelength emission of 1 shows an average lifetime of ca. 22-27 ns, being significantly shorter only for acetonitrile (11.9 ns). τav ¼
n X i¼1
fi τi
Ri τi with fi ¼ P n Ri τ i
ð1Þ
i¼1
Acetonitrile is the least favorable solvent for observation of significant emission of both dyads. Furthermore, the longwavelength fluorescence of 1 is virtually absent in this solvent medium. Upon addition of 1 equiv of trifluoroacetic acid (TFA), which is sufficient to protonate the spacer-integrated amino function (see fluorescence titration of 2 in Figure 4),43 the picture changes completely.13,14 Now much stronger shortwavelength emissions are obtained for 1 (Φf = 0.037) and 2 (Φf = 0.112). This corresponds to fluorescence enhancement with factors of 37 and 56 for 1 and 2, respectively. In addition, also quite intense long-wavelength emission is now obtained for 1, in fact showing the highest quantum yield (Φf = 0.07) in the entire series and the best ratio (ca. 1.9; see Figure 2) of longto short-wavelength emission. Short-wavelength fluorescence enhancement is also observed upon addition of methanol to acetonitrile solutions of 1 and 2, although to a lesser extent 1095
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A
Figure 4. Fluorescence titration of dyad 2 (10 μM) with TFA in acetonitrile. The inset shows the corresponding titration curve.
(factors of 5 and 28, respectively). In the case of dyad 1, again also the long-wavelength emission appears. For the sake of a complete spectroscopic characterization of the dyads, phosphorescence measurements in tetrahydrofuran glass at 77 K are also performed. The dyads 1 and 2 show nearly identical spectra with λmax = 538 and 540 nm, respectively; see Figure 3 for dyad 1. Also the phosphorescence lifetimes (τp) are very similar (530 ms for 1 and 460 ms for 2).
’ DISCUSSION Assignment of Excimer Emission. The observation of an unstructured and long-wavelength emission band for 1 in most solvents (except in pure acetonitrile) can be attributed to the formation of an excimer. The always observed short-wavelength band for 1 and 2 is indicative of emission from a locally excited state (monomer). These assignments are in agreement with the observations made for the dyads 4(n).28 Also for the spermine conjugate 3 a dual emission was described.18,27 In the latter case the long-wavelength emission was interpreted as originating from the direct excitation of a ground-state dimer. As shown in Figure 3 for dyad 1 in 2,2,2-trifluoroethanol, the excitation spectra monitoring the maxima of both emissions are very similar, which lends support to the idea that the red-shifted emission originates from an excimer, possibly resulting from the excitation of a “loose” ground-state dimer (see below). Such preformed excimers require only minimal excited-state adjustments regarding the interchromophore distance and orientation.44 They are also characterized by very short rise times of the excimer emission, which are verified in our experiments (see below). The absorption spectra of 1 and the model N-propyl-1,8-naphthalimide in the same solvent do not coincide totally (see Supporting Information), with the dyad showing an absorption band without the typical fine structure and slightly red-shifted maxima (Δλ = þ7 nm). This also points to some ground-state interaction between both chromophores. However, a similar red shift of the absorption band is also observed for dyad 2, which lacks significant long-wavelength emission. The notion that excimer formation originates from a preformed “loose” dimer, which, however, is not tightly preorganized (this requires parallel orientation of both naphthalimide units at contact distance of 3-4 Å), is further supported by measurements at 77 K in ethanol/methanol (4/1) glass. As shown in Figure 5, no long-wavelength emission is observed for dyad 1 under these
ARTICLE
Figure 5. Emission spectra (λexc = 330 nm) of dyad 1 in ethanol/ methanol (4/1) at T = 295 K (;) and at T = 77 K (---). The emission at 77 K for λ > 500 nm is phosphorescence.
conditions. In case of the existence of a “tight” ground-state dimer, this conformation would be frozen and strong excimer emission should be observed, which is not the case. However, the hindered diffusion in the glass is expected to prevent the required geometrical adjustments of the “loose” dimer in the excited state and leads to the absence of excimer formation. It can be concluded that there is some degree of ground-state interaction, but the medium distance between both chromophores is too long and further approximation in the excited state is required to yield the observed excimer emission (see discussion below).45 Time Dependence of Excimer Emission. No rise component (negative amplitude) is observed in the kinetic traces for excimer emission of dyad 1, which leads us to establish a lower limit of ca. 1012 s-1 for the rate constant of excimer formation, meaning that excimers are of the preformed type (see also above).44 Excitation of the fraction of dyads in the “loose” dimer configuration always yields excimer emission in less than a few picoseconds; hence, no monomer fluorescence emission can be observed from them. As monomer fluorescence is nevertheless detected, a second population of dyad conformations must exist, which cannot engage in excimer formation (because no rise time is observed). There are therefore two different excited-state populations separated by a kinetic barrier, one corresponding to preformed excimers and the other corresponding to conformational pairs unable to form excimers during the monomer excited-state lifetime (Scheme 1). Spacer Length Dependence of Excimer Emission. For dyad 2 only short-wavelength emission originating from the monomer is detected, except for a minor excimer emission in water. The only structural difference between 1 and 2 is the slightly longer spacer (two more methylene units) of the latter dyad. Therefore, the distinct behavior of 2 with respect to excimer formation may be related to conformational issues of the spacer, which hinder an efficient approach of both chromophoric units (see also discussion of homo-F€orster resonance energy transfer below). This has been verified for classical systems with aromatic chromophores (such as pyrene) as well as for dyads with terminal naphthalimide units, for example, dyads 4(n).28,32 For example, the maximum excimer emission quantum yield of 4(n) was found for a spacer length of six methylene groups. For comparison, 1 has five units (counting the NH as isoelectronic CH2), while the spacer of 2 is composed of 1096
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A
ARTICLE
Scheme 1. Mechanistic Picture That Takes into Account the Simultaneous Observation of Fluorescence from the Monomer and Excimera
Figure 6. Fluorescence anisotropy kinetic curves at 77 K in ethanol/ methanol (4/1) glass at λobs = 390 nm of (a) N-propyl-1,8-naphthalimide, (b) dyad 1 (black), and (c) dyad 2 (light gray). IRF = instrument response function.
a
demonstrated by the formation of excimers. Z ¥ FD ðλÞεA ðλÞλ4 dλ Jd- d ¼
seven units. However, the abrupt absence of excimer formation for an increase of only two length units of the spacer is surprising and, for example, not reflected in the series of dyads 4(n), where a rather gradual decrease of long-wavelength emission with increasing spacer length was observed. In a recent study about 1,n-di(1-pyrenyl)alkanes, a similar sharp decrease of excimer emission around n = 7 was reported.32 This observation was reasoned with the argument that intrachain H/H interactions between the methylene groups next to the chromophore become effective. As a consequence, the orbital overlap of the terminal chromophore units required for efficient excimer formation is prevented. Actually, in the present case (dyads 1 and 2), an analogous argument may be applied to explain the observed drop in excimer emission upon elongation of the spacer. As shown for the 1,n-di(1-pyrenyl)alkanes (see above), the existence of methylene groups next to the excimerforming chromophores is a crucial factor that also applies to our dyads. Interchromophore Distance and Involvement of Homo€ rster Resonance Energy Transfer. In order to gain more Fo insight into the interchromophoric distance in the dyads, homoF€orster resonance energy transfer (homo-FRET) is included in the mechanistic discussion. This process is commonly understood in terms of the interaction between two identical chromophores (one in the excited state and one in the ground state), under the precondition that the absorption and fluorescence show significant spectral overlap. The corresponding overlap integral Jd-d is calculated by using the absorption and fluorescence spectra of the model compound N-propyl-1,8-naphthalimide (eq 2). In ethanol, a value of Jd-d = 5.6 10-13 cm6 3 mol-1 results, which with the quantum yield Φdonor = 0.056 (this work) of the unquenched model donor N-propyl-1,8-naphthalimide yields a critical F€orster radius of R0 ca. 13 Å (eq 3, with the orientation factor κ2 = 2/3 and the refractive index of ethanol n = 1.3614).46 The critical F€orster radius is defined as the distance between energy donor and acceptor for which the energy transfer quantum yield is equal to 50%. Expectedly, shorter distances are possible for both dyads, which especially in the case of 1 is
ð2Þ
0
M = monomer; EXC = excimer; S = proton or protic solvent.
9 ln 10k2 Φdonor Jd- d ð3Þ 128π5 n4 NA Further evidence for the possibility of medium-range chromophore interaction in 1 was obtained from time-resolved fluorescence polarization measurements in ethanol/methanol (4/1) glass at 77 K. On the one hand, for the model compound N-propyl-1,8-naphthalimide, the fluorescence anisotropy r is constant at a value of 0.23. On the other hand, for the bis(naphthalimides) 1 and 2, the anisotropy levels off at a value of 0.12 (see Figure 6). This depolarization is expected for homo-FRET and the drop to half the value for the fundamental anisotropy of N-propyl-1,8-naphthalimide corroborates the random mutual orientation of both chromophore units.47 The fluorescence anisotropy decay times θ for dyads 1 and 2 are obtained as 7.6 and 39 ps, respectively. If, for simplicity, a single distance and isotropic transfer are considered,48 the decay times θ are related to the F€orster rate constant by eq 4.48 The rate constants for the homoFRET process in dyads 1 and 2 are thus estimated as 6.6 1010 s-1 and 1.3 1010 s-1, respectively. It should be pointed that the competitive homo-FRET does not alter the possibility of PET (see below) or excimer formation (when possible), because the nominal situation remains essentially the same due to the two indistinguishable chromophores in the symmetric dyads: one ground-state naphthalimide and one excited-state naphthalimide. 1 ð4Þ khomo-FRET ¼ 2θ 1 R0 6 khomo-FRET ¼ ð5Þ τ0 R R0 6 ¼
With eq 5, the critical F€orster radius R0 of 13 Å (see above), and τ0 = 475 ps (for unquenched N-methyl-1,8-naphthalimide in ethanol),17 the average interchromophore distances in 1 and 2 are obtained as ca. 7.5 Å and ca. 9.5 Å, respectively. Although these values refer to 77 K, they corroborate our notion that the chromophores in 1 are, on average, closer than in dyad 2. However, a distance of ca. 7.5 Å is still larger than expected for an intimate contact in a “tight” dimer (ca. 3-4 Å). On the other hand, this supports our interpretation of the involvement of a “loose” ground-state dimer for 1. 1097
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A Calculation of the average number of excited-state cycles, eq 6,49 is also possible: 12 R0 R n ¼ ð6Þ 6 R0 1þ2 R and yields hn = 13.3 and 3.1 for 1 and 2, respectively, which are quite different. The excitation energy hops between chromophores much more frequently in dyad 1. A final argument in favor of a closer distance between the chromophore units in 1 comes from comparison of the 1H NMR chemical shifts in the aromatic region with those of 2 (see Supporting Information) and monomeric N-propyl-1,8naphthalimide. While the chemical shifts for dyad 2 coincide virtually with those reported for N-propyl-1,8-naphthalimide in the same solvent (CDCl3),18 significant upfield shifts (-Δδ = 0.02-0.18 ppm) are noted for dyad 1. These chemical shift differences are expected for the protons of naphthalimide rings in closer spatial distance. Solvent Dependence of Excimer Emission. A detailed analysis of the relative contribution of excimer emission to the total emission in dependence on the nature of the solvent provides additional insights in the excited-state processes of 1. On the one hand, there seems no clear trend of variation of excimer emission upon changing solvent polarity or viscosity (see values in Table 1). For example, ethanol and 2,2,2-trifluoroethanol have comparable viscosities (η) and dielectric constants (ε), with the latter being even somewhat more viscous, but only the fluorinated alcohol promotes significant excimer emission. On the other hand, the Kamlet-Taft R parameter, which is a measure for the hydrogenbond donator strength of a solvent, is largely different for these two solvents. 2,2,2-Trifluoroethanol, which shows intense excimer emission, is a much stronger hydrogen-bond-donating solvent than ethanol. Hence, it is concluded that the amino function in the bridge is involved in hydrogen bonding, which, depending on the solvent, promotes efficient excimer formation for dyad 1. Two effects may be held responsible for this observation: (a) a conformational change of the spacer, bringing the two chromophores closer together and thereby promoting the formation of a “loose” dimer, and (b) the switch-off of PET from the amine to the strong electron-accepting naphthalimide as competitive excitedstate pathway (see below). Dyad 2 shows no significant excimer emission. However, the monomer emission is equally dependent on the solvent. The same is true for the monomer emission of dyad 1. A significant enhancement of fluorescence quantum yields and lifetimes is observed in protic solvents as compared to acetonitrile. The blocking of PET through hydrogen bonding with the spacer amino function is one factor that may contribute to the observed trends (see below). However, as discussed earlier by Wintgens et al.,17 the short-wavelength monomer emission of 1,8-naphthalimides (such as N-methyl-1,8-naphthalimide) may be inherently enhanced through interaction of protic solvents with the carbonyl oxygens of the imide. As a consequence, the energy gap between the emitting π,π* singlet state and the upper n,π* triplet state is increased. The resulting decreased mixing between both states confers mainly π,π* character to the fluorescent state, which results in increased fluorescence quantum yields and lifetimes. The modulation of the monomer emission of the investigated dyads should also be influenced by this solvent
ARTICLE
effect: For 2 it is observed that 2,2,2-trifluoroethanol (a strong hydrogen-bond-donating solvent) yields Φf = 0.21 and τav = 1.52 ns, which has to be compared to Φf = 0.112 and τav = 0.76 ns for protonated 2 in the nonprotic acetonitrile. The calculation of radiative and nonradiative rate constants would certainly reveal more insights in the solvent-mediated photophysical trends. However, the rather complex kinetic situation (multiexponential monomer decay) obscures such quantification. Photoinduced Electron Transfer versus Excimer Formation. Support for the involvement of photoinduced electron transfer (PET) in non-hydrogen-bonding solvents and in the absence of protonation by acid is provided by the thermodynamic feasibility of this process, which is readily verified with the Rehm-Weller equation. For the related dyad 3, a value of ΔGPET = -0.86 eV in acetonitrile was reported for the monomer excited singlet state PET quenching. Due to the electronic resemblance of our dyads and 3, a similar energetics for the corresponding PET processes is expected for 1 and 2.18 The effect of fluorescence enhancement upon blocking of PET is further verified by protonation in acetonitrile (see Figure 2 for dyad 1 and Figure 4 for dyad 2). In a similar manner, the interaction of the amino lone pair with hydrogen-bond-donating solvents decreases the electron donor strength of the amine and leads to partial (e.g., for the weaker hydrogen-bond donor ethanol) or complete cancellation of PET (in the strong hydrogen-bond-donating 2,2,2-trifluoroethanol). This is experimentally verified by the dramatic increase of the total fluorescence quantum yield for both dyads. It is noteworthy that, although PET is hindered in protonated 2 in acetonitrile, the spacer conformation does not allow a sufficient spatial approximation between both chromophores to yield excimer formation. The only exception is water as solvent, which may be the result of a hydrophobic effect forcing the two naphthalimides to closer distances.
’ CONCLUSIONS A minimal difference in the molecular structure of two bis(1,8naphthalimides) is observed to yield drastic changes in the photophysical behavior, especially with respect to the formation of excimers. It is found that the blocking of competitive PET processes involving an electron-donating amino function, integrated in the spacer, enables the formation of excimers and the observation of dual fluorescence with large virtual Stokes shifts (ca. 150 nm) and reasonable quantum yields (up to Φf = 0.07 for the long-wavelength emission). Hydrogen-bond-donating solvents as well as protons may be used for the modulation of excimer emission. However, this emission is almost exclusively observed for dyad 1, which contains a spacer with five isoelectronic units. The elongation of the spacer with two more methylene groups (dyad 2) gives rise to only a simple monomer fluorescence switching. The discussion of homo-FRET yields some further insights in the interchromophore distance in each dyad and reveals the involvement of a “loose” ground-state dimer. ’ ASSOCIATED CONTENT
bS
Five figures showing 1H NMR spectra of 1 and 2 and additional spectroscopic data (UV/vis absorption and fluorescence). This material is available free of charge via the Internet at http://pubs.acs.org.
1098
Supporting Information.
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
The Journal of Physical Chemistry A
’ AUTHOR INFORMATION Corresponding Author
*E-mail
[email protected] (M.N.B.-S.);
[email protected] (U.P.).
’ ACKNOWLEDGMENT Financial support by the Ministerio de Ciencia e Innovacion, Madrid (Grant CTQ2008-06777-C02-02/BQU for U.P.), the Consejería de Innovacion, Ciencia y Empresa de la Junta de Andalucía (Grant P08-FQM-3685 for U.P.), and the Fundac-~ao para a Ci^encia e Tecnologia, Lisbon (Fellowship SFRH/BPD/ 34384/2006 for R.F., Grant PTDC/ENR/64909/2006 for M.N. B.-S.) is gratefully acknowledged. Furthermore, we thank Marek Kluciar for assistance in the synthesis of 1 and 2. ’ REFERENCES (1) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.; McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. Rev. 1997, 97, 1515–1566. (2) Daffy, L. M.; de Silva, A. P.; Gunaratne, H. Q. N.; Huber, C.; Lynch, P. L. M.; Werner, T.; Wolfbeis, O. S. Chem.;Eur. J. 1998, 4, 1810–1815. (3) Ramachandram, B.; Saroja, G.; Sankaran, N. B.; Samanta, A. J. Phys. Chem. B 2000, 104, 11824–11832. (4) Gunnlaugsson, T.; Lee, T. C.; Parkesh, R. Org. Biomol. Chem. 2003, 1, 3265–3267. (5) Guo, X.; Qian, X.; Jia, L. J. Am. Chem. Soc. 2004, 126, 2272–2273. (6) Esteban-Gomez, D.; Fabbrizzi, L.; Licchelli, M.; Sacchi, D. J. Mater. Chem. 2005, 15, 2670–2675. (7) Tusa, J. K.; He, H. J. Mater. Chem. 2005, 15, 2640–2647. (8) Chovelon, J.-M.; Grabchev, I. Spectrochim. Acta A 2007, 67, 87–91. (9) Koner, A. L.; Schatz, J.; Nau, W. M.; Pischel, U. J. Org. Chem. 2007, 72, 3889–3895. (10) Nandhikonda, P.; Begaye, M. P.; Heagy, M. D. Tetrahedron Lett. 2009, 50, 2459–2461. (11) Jiang, G.; Wang, S.; Yuan, W.; Jiang, L.; Song, Y.; Tian, H.; Zhu, D. Chem. Mater. 2006, 18, 235–237. (12) Li, Y.; Cao, L.; Tian, H. J. Org. Chem. 2006, 71, 8279–8282. (13) Kluciar, M.; Ferreira, R.; de Castro, B.; Pischel, U. J. Org. Chem. 2008, 73, 6079–6085. (14) Ferreira, R.; Remon, P.; Pischel, U. J. Phys. Chem. C 2009, 113, 5805–5811. (15) Pischel, U.; Uzunova, V. D.; Remon, P.; Nau, W. M. Chem. Commun. 2010, 46, 2635–2637. (16) Barros, T. C.; Molinari, G. R.; Filho, P. B.; Toscano, V. G.; Politi, M. J. J. Photochem. Photobiol. A: Chem. 1993, 76, 55–60. (17) Wintgens, V.; Valat, P.; Kossanyi, J.; Biczok, L.; Demeter, A.; Berces, T. J. Chem. Soc., Faraday Trans. 1994, 90, 411–421. (18) Jones, G., II; Kumar, S. J. Photochem. Photobiol. A: Chem. 2003, 160, 139–149. (19) Li, J.-Q.; Li, X.-Y. J. Phys. Chem. A 2007, 111, 13061–13068. (20) Rogers, J. E.; Weiss, S. J.; Kelly, L. A. J. Am. Chem. Soc. 2000, 122, 427–436. (21) de Silva, A. P.; Gunaratne, H. Q. N.; Habib-Jiwan, J.-L.; McCoy, C. P.; Rice, T. E.; Soumillion, J.-P. Angew. Chem., Int. Ed. 1995, 34, 1728–1731. (22) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Lynch, P. L. M. New J. Chem. 1996, 20, 871–880. (23) de Silva, A. P.; Vance, T. P.; West, M. E. S.; Wright, G. D. Org. Biomol. Chem. 2008, 6, 2468–2480. (24) Abad, S.; Kluciar, M.; Miranda, M. A.; Pischel, U. J. Org. Chem. 2005, 70, 10565–10568. (25) Barros, T. C.; Filho, P. B.; Toscano, V. G.; Politi, M. J. J. Photochem. Photobiol. A: Chem. 1995, 89, 141–146.
ARTICLE
(26) Licchelli, M.; Biroli, A. O.; Poggi, A.; Sacchi, D.; Sangermani, C.; Zema, M. Dalton Trans. 2003, 4537–4545. (27) McMasters, S.; Kelly, L. A. J. Phys. Chem. B 2006, 110, 1046– 1055. (28) Cho, D. W.; Fujitsuka, M.; Sugimoto, A.; Majima, T. J. Phys. Chem. A 2008, 112, 7208–7213. (29) Cho, D. W.; Fujitsuka, M.; Choi, K. H.; Park, M. J.; Yoon, U. C.; Majima, T. J. Phys. Chem. B 2006, 110, 4576–4582. (30) Cho, D. W.; Fujitsuka, M.; Yoon, U. C.; Majima, T. J. Photochem. Photobiol. A: Chem. 2007, 190, 101–109. (31) Hirayama, F. J. Chem. Phys. 1965, 42, 3163–3171. (32) Zachariasse, K. A.; Mac-anita, A. L.; K€uhnle, W. J. Phys. Chem. B 1999, 103, 9356–9365. (33) Albelda, M. T.; Bernardo, M. A.; Díaz, P.; García-Espa~ na, E.; Seixas de Melo, J.; Pina, F.; Soriano, C.; Luis, S. V. Chem. Commun. 2001, 1520–1521. (34) Seixas de Melo, J.; Albelda, M. T.; Díaz, P.; García-Espa~ na, E.; Lodeiro, C.; Alves, S.; Lima, J. C.; Pina, F.; Soriano, C. J. Chem. Soc., Perkin Trans. 2 2002, 991–998. (35) Albelda, M. T.; García-Espa~ na, E.; Gil, L.; Lima, J. C.; Lodeiro, C.; Seixas de Melo, J.; Melo, M. J.; Parola, A. J.; Pina, F.; Soriano, C. J. Phys. Chem. B 2003, 107, 6573–6578. (36) Cao, H.; Chang, V.; Hernandez, R.; Heagy, M. D. J. Org. Chem. 2005, 70, 4929–4934. (37) Nandhikonda, P.; Begaye, M. P.; Cao, Z.; Heagy, M. D. Chem. Commun. 2009, 4941–4943. (38) Pina, F.; Lima, J. C.; Lodeiro, C.; Seixas de Melo, J.; Díaz, P.; Albelda, M. T.; García-Espa~ na, E. J. Phys. Chem. A 2002, 106, 8207– 8212. (39) Zhang, X.; Rehm, S.; Safont-Sempere, M. M.; W€urthner, F. Nat. Chem. 2009, 1, 623–629. (40) Grabchev, I.; Petkov, C.; Bojinov, V. Dyes Pigm. 2001, 48, 239– 244. (41) Valeur, B. Molecular Fluorescence: Principles and Applications, 1st ed.; Wiley-VCH: Weinheim, Germany, 2001. (42) N-Propyl-1,8-naphthalimide shows a broad red-shifted emission with a maximum at 465 nm in the solid state at room temperature. (43) Dyad 2 was chosen for this titration, because a simple monomer fluorescence OFF-ON switching without complicating excimer formation as for dyad 1 applies. However, due to the similar structural and electronic situation of the amino function in 1, it is expected that 1 equiv of TFA (saturation point of the titration of 2) is as well sufficient for the protonation of this dyad. (44) Berberan-Santos, M. N.; Canceill, J.; Brochon, J.-C.; Jullien, L.; Lehn, J.-M.; Pouget, J.; Tauc, P.; Valeur, B. J. Am. Chem. Soc. 1992, 114, 6427–6436. (45) The long-wavelength emission lifetime is measured in the nanosecond range and this excludes its formation by triplet-triplet annihilation. It is noteworthy that, due to the presence of an amino function in the spacer, an exciplex-related emission cannot be excluded a priori. However, in such a case the use of polar solvents and the protonation of the electron-donating amino nitrogen should yield the disappearance of the long-wavelength emission due to stabilization of the charge-separated pair or to blocking of charge-transfer processes, respectively, which is not observed. (46) Montalti, M.; Credi, A.; Prodi, L.; Gandolfi, M. T. Handbook of Photochemistry, 3rd ed.; CRC Press/Taylor & Francis Group: Boca Raton, FL, 2006. (47) Berberan-Santos, M. N.; Choppinet, P.; Fedorov, A.; Jullien, L.; Valeur, B. J. Am. Chem. Soc. 2000, 122, 11876–11886. (48) Berberan-Santos, M. N.; Valeur, B. J. Chem. Phys. 1991, 95, 8048–8055. (49) Baleiz~ao, C.; Berberan-Santos, M. N. ChemPhysChem 2009, 10, 199–205. (50) Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H.; Taft, R. W. J. Org. Chem. 1983, 48, 2877–2887.
1099
dx.doi.org/10.1021/jp110470h |J. Phys. Chem. A 2011, 115, 1092–1099
