EÜect of molecular structure and hydrogen bonding on the —uorescence of hydroxy-substituted naphthalimides Laç szloç Biczoç k,*a Pierre Valatb and Veç ronique Wintgensb a Chemical Research Center, Hungarian Academy of Sciences, P.O. Box 17, 1525 Budapest, Hungary b L aboratoire des C.N.R.S., E.R. 241, 2, 8 rue H. Dunant, 94320 T hiais, Mateç riaux Moleç culaires, France Received 7th June 1999, Accepted 23rd August 1999 Fluorescence properties of hydroxy-naphthalimides were studied in methylene chloride in the absence and the presence of hydrogen-bonding additives.The position of the HO-substituent only slightly aÜects the radiative rate, however, the triplet yield and the rate of the radiationless processes are considerably higher for the 3-hydroxy derivative. Addition of nitrogen-heterocyclic compounds leads not only to hydrogen-bonding in the ground state but also —uorescence quenching. The parallel change throughout the series of the hydrogen-bond acceptors between the proton affinity and the rate constants of dynamic quenching indicates that proton displacement plays a crucial role in the excited hydrogen-bonded complexes.Interaction of hydroxy-naphthalimides with pyridine and benzoxazole results in rapid radiationless deactivation from the singlet excited state, whereas intense emission as well as long —uorescence lifetime characterize imidazole and pyrazole complexes. The dual emission of the imidazole complexes observed in solvents of medium polarity is assigned to two conformers which diÜer in the extent of the proton shift along the hydrogen-bond.Introduction The molecular mechanism of the excited state relaxation induced by intermolecular hydrogen-bond formation is of great interest because it belongs to the most fundamental processes of photochemistry. Most of the studies in this –eld have dealt with the eÜect of the hydrogen-bond donors on the —uorescent properties of aromatic heterocyclic1 and carbonyl compounds.2h6 These molecules form hydrogen-bonded complexes with alcohols and hydroperoxides in the excited state and the hydrogen-bond acts as an efficient vibronic dissipative mode in the nonradiative transition.2h6 On the other hand, coupled electron»proton movement was found to play a dominant role in the deactivation of excited hydrogen-bonded porphyrins7,8 and Ru(II)polypyridyl complexes8 as well as in the interaction of excited ketones with phenols.4,9 Picosecond laser photolysis studies established that excited hydrogen-bond donors, such as aromatic TN»H or »O»H compounds, are efficiently quenched by pyridine derivatives via non-—uorescent hydrogen-bonded complex, in which associated electron and proton displacement facilitates the charge transfer interaction between the two conjugate pelectronic systems.10 However, when phenols form hydrogenbonded complexes with aliphatic amines no charge delocalization is possible along the hydrogen-bond and photoexcitation induces proton transfer.11 The present paper focuses on the question of how the variation of the molecular structure of the hydrogen-bonding additive in—uences the —uorescent properties and the deactivation mechanism of the excited molecules.In order to reveal the role of aromaticity and proton affinity in the hydrogenbonding induced quenching process, aromatic heterocyclic compounds were used as hydrogen-bond acceptors. We show examples where the excited hydrogen-bonded complexes emit dual —uorescence, which is assigned to two conformers diÜering in the extent of the proton shift along the hydrogen-bond.Hydroxy-substituted naphthalimides were chosen as model compounds in these studies because they have both hydrogenbond donor and acceptor moieties, and, on the basis of our previous studies,12 they are expected to be strongly —uorescent. In addition, the electron withdrawing character of the imide group probably enhances the acidity in the excited state and light absorption may serve as an ultrafast trigger for the proton transfer reaction.Recent studies of related compounds demonstrated that substitution of naphthols with cyano or methanesulfonyl groups markedly increases the photoacidity. 13h15 Another main goal of this work was to examine how the introduction of a hydroxy substituent into the 1,8-naphthalimide moiety alters the dominant energy dissipating pathways occurring from the singlet excited state. The investigated compounds are given in Scheme 1.Experimental Acetonitrile, methylene chloride (Prolabo, HPLC grade), dimethylsulfoxide (Merck, spectroscopic grade) and tri- —uoroethanol (Aldrich) were used as received. Benzoxazole, imidazole, isoxazole, pyrazole and pyridine were purchased from Aldrich (highest quality available). N-Methyl-1,8-naphthalimide (NI), also called 2-methyl-1Hbenz[ d,e]isoquinoline-1,3(2H)-dione, was available from our previous study.16 4-Hydroxy-N-methyl-1,8-naphthalimide (4- HONI), also called 6-hydroxy-2-methyl-1H-benz[d,e] Scheme 1 Phys.Chem. Chem. Phys., 1999, 1, 4759»4766 4759 This journal is The Owner Societies 1999 (isoquinoline-1,3(2H)-dione, was prepared by demethylation of 4-methoxy-N-methyl-1,8-naphthalimide,17 following a procedure similar to one described previously.18 One equivalent of the methoxy derivative mixed with 40 equivalents of pyridine hydrochloride was heated at 190 °C under argon for 35 min.After cooling, the solid reaction mixture was added to an aqueous HCl solution (1 M). The solid was –ltered, washed with aqueous HCl solution (1 M) and water. The crude product was puri–ed by dissolution in an aqueous Na2CO3 solution, followed by extraction with ether and precipitation by the addition of perchloric acid in the aqueous phase (45% yield), m.p. 298»302 °C (lit.18 303»305 °C). 3-Hydroxy-N-methyl-1,8-naphthalimide (3-HONI), also called 5-hydroxy-2-methyl-1H-benz[d,e]isoquinoline-1,3(2H)- dione, was synthesized via four reaction steps.First, 3-nitro- N-methyl-1,8-naphthalimide was prepared by condensation of 3-nitro-1,8-naphthalic anhydride (Aldrich) with methylamine hydrochloride in acetic acid. Then 1 equivalent of 3-nitro-Nmethyl- 1,8-naphthalimide was reduced by 4 equivalents of tin(II)chloride in hot hydrochloric acid.19 The obtained amino derivative was diazotized and the diazonium salt was hydrolyzed by aqueous HCl solution.The –nal product was puri–ed by the method described for 4-HONI (15% yield, m.p. 249» 251 °C). The UV»visible absorption spectra were obtained with a Varian»Cary model 50 Bio apparatus. Fluorescence spectra were recorded with an SLM-Aminco model 8100 device. Fluorescence quantum yields of 3-HONI and 4-HONI were determined by comparison with that of 4-methoxy-N-methyl-1,8- naphthalimide in acetonitrile solution, for which a reference yield of was taken.12 Singlet lifetimes were mea- UF\0.88 sured by excitation with a B.M.Industries frequency-tripled Nd»YAG laser (pulse duration 30 ps FWHM), using the experimental set-up already described.20 Laser —ash photolysis experiments were carried out with 8 ns FWHM pulse of a Nd»YAG laser and the monitoring light from Applied Photophysics xenon lamp passed through the sample perpendicular to the excitation. Intersystem crossing (ISC) quantum yields were determined in oxygen-free solutions relative to triplet benzophenone.We compared the initial triplet»triplet absorbances of the investigated compound (A) at 480 nm with that of the benzophenone reference at 530 nm. The solutions (Aref) had matched absorbances at the excitation wavelength (355 nm). The triplet yields were obtained based on the equa- (UISC) tion UISC\3Uref(A/Aref)(eref/e) (1) using the well-established yield and molar (3Uref\1.00)21 absorption coefficient M~1 cm~1 at 530 nm)22 for (eref\7200 triplet benzophenone, whereas e\10 000 M~1 cm~1 was taken for the triplet molar absorption coefficient of naphthalimides at 480 nm.16 Results and discussion I.Photophysical properties of hydroxy-naphthalimides Absorption and —uorescence spectra. Fig. 1 presents the absorption and —uorescence spectra of hydroxy-naphthalimides in methylene chloride. The absorption spectra resemble those of the corresponding naphthols, however, the introduction of the imide moiety results in a remarkable bathochromic shift of the maxima, which becomes most apparent for the low energy bands.For the energy of the lowest excited singlet states, 308 kJ mol~1 and 300 kJ mol~1 were obtained from the locations of the intersections of the normalized absorption and —uorescence spectra in the case of 3-HONI and 4-HONI, respectively. The energy diÜerence between the –rst absorption band of the hydroxy-naphthalimide and the corresponding naphthol was found to be more considerable for the 4-HO derivative.This clearly indi- Fig. 1 Absorption and —uorescence spectra of 3-HONI (»») and 4-HONI (… … …) in CH2Cl2 . cates the larger extent of conjugation between the electron donating OH and the electron withdrawing imide moiety in 4-HONI. The Stokes-shift of the —uorescence spectrum is more pronounced (78 nm) for 4-HONI compared with that of 3-HONI (35 nm). Changing the solvent from methylene chloride to acetonitrile leads to a 6 nm and 2 nm —uorescence maximum displacement for the former and the latter compounds, respectively.These small solvatochromic shifts together with the small decrease of the state energy with S1 increasing solvent polarity (vide infra) suggest that the lowest excited singlet states have limited charge transfer character for both HONI isomers. Fig. 2 shows the excitation and the —uorescence spectra of the hydroxy-naphthalimides and their conjugated bases in acetonitrile. In this solvent we used 1.5]10~4 M perchloric acid to prevent the dissociation of the OH-moiety and to record solely the spectra corresponding to the phenolic form.In order to deprotonate the OH group, 2 ll of 1 M KOH in methanol was added into 2 ml of hydroxy-naphthalimide solution. Under these conditions, the spectra are assigned to the naphtholate anion and the bands are located at lower energies. Based on the intersection of the normalized excitation and —uorescence spectra, 305 kJ mol~1 and 210 kJ mol~1 were calculated for the energies of the –rst excited singlet state of 3-HONI and its conjugated base, respectively.A much Fig. 2 Fluorescence and excitation spectra of (A) 3-HONI and (B) 4-HONI in acetonitrile. Phenolic form in the presence of 1.5]10~4 M perchloric acid (… … …) and deprotonated anion form in the presence of 1]10~3 M KOH (»»). 4760 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766smaller diÜerence was observed between the energies of S1 4-HONI and its deprotonated form (297 kJ mol~1 and 229 kJ mol~1, respectively).According to the Foé rster cycle,23 these results suggest that 4-HONI undergoes a smaller acidity enhancement upon electronic excitation. The lack of the long wavelength emission in clearly indicates that none of CH2Cl2 the singlet excited hydroxy-naphthalimides are sufficiently acidic to transfer proton to this solvent. Photophysical properties. Table 1 demonstrates that substitution of N-methyl-1,8-naphthalimide (NI) with a hydroxy moiety leads to a considerable change in —uorescence yield —uorescence lifetime and triplet yield Fluores- (UF), (qF) (UISC).cence yields and lifetimes increase more than one order of magnitude through the series of NI, 3-HONI, 4-HONI, whereas the variation of the triplet yields exhibits the opposite tendency; they decrease parallel with the energy of the lowest excited singlet In order to get a deeper insight into the (E(S1)). factors controlling the rate of the energy dissipating pathways, the rate constants for —uorescence intersystem crossing (kF), and internal conversion were derived using the (kISC) (kIC) expressions given below: kF\UF/qF (2) kISC\UISC/qF (3) kIC\(1[UISC[UF)/qF (4) It is seen from the rate constants presented in Table 1 that —uorescence emission is the dominant process from the singlet excited state of 4-HONI, however, radiationless transitions prevail for 3-HONI.In contrast with the very fast intersystem crossing observed for unsubstituted NI,16 triplet formation is fairly slow for both hydroxy-naphthalimides.This is especially true of the 4-HO derivative, where almost negligible was kISC found and the phosphorescence is extremely weak at 77 K in organic glass. In the case of 3-HONI the more efficient intersystem crossing permits the determination of the triplet»triplet absorption and the phosphorescence spectra. The triplet» triplet absorption maxima detected by a laser —ash photolysis technique appear at 450 and 480 nm at ambient temperature in whereas the phosphorescence peaks can be found CH2Cl2 , at 590 and 640 nm at 77 K in 95 : 5 butyronitrile : butyl acetate mixture.From the 0»0 transition of the phosphorescence the energy of the –rst triplet excited state is calculated to be 213 kJ mol~1. This value is lower than the one found for N-methyl-1,8-naphthalimide (221 kJ mol~1) indicating that substitution with an OH group not only decreases the energy of the state but also that of the state.S1 T1 The efficient triplet formation for N-methyl-1,8-naphthalimide was rationalized in terms of the transition (UISC\0.94) from the lowest singlet excited state to a close-lying higher triplet state As we established in a previous paper,16 (Tn).16 this is a thermally enhanced process in moderately and strongly polar solvents, which results in a temperature dependent —uorescent behavior. However, the —uorescence lifetimes of 3-HONI and 4-HONI were found to be temperature independent in the 296»198 K range.Based on these results, we can exclude the thermally activated intersystem crossing pathway for hydroxy-naphthalimides. The electron donating hydroxy group decreases the state energy and thereby, S1 increases the energy gap. With such an increased energy S1»Tn gap, thermal activation is not able to initiate the endothermic transition, consequently, triplet formation can occur S1 ]Tn only by the slower transitions to lower triplet levels.Similar eÜects were observed for 4-methoxy-N-methyl-1,8-naphthalimide, which also has an electron donating moiety.12 The signi–cant diÜerence in the triplet formation rate constant between 3-HONI and 4-HONI (Table 1) can be rationalized based on the relative position of the singlet and the triplet energy levels. Semi-empirical calculation24 showed that the energy gap is larger for 4-HONI compared with S1»T1 that of 3-HONI (144 kJ mol~1 and 132 kJ mol~1 were obtained, respectively).This larger energy gap may lessen the magnitude of the spin»orbit coupling between the and S1 T1 states leading to a slower intersystem crossing process for 4-HONI. It is evident from the data summarized in Table 1 that the introduction of a hydroxy group into the 1,8-naphthalimide moiety in position 3 hardly in—uences the rate constant for internal conversion however, the radiationless (kIC), S1 ]S0 transition is markedly decelerated for 4-HONI.It seems to be a general tendency that the 4-substitution of the 1,8-naphthalimide skeleton with an electron donating group leads to a reduced This eÜect can be observed for HO», and kIC. CH3 O» derivatives alike. Prado et al. suggested25 that the mol- NH2» ecule has a quinoid resonance form if an electron donating group is attached to position 4. This kind of electron displacement, which can occur both in the and the states, may S1 S0 aÜect the vibronic coupling between these states and consequently, may lead to slower internal conversion.II. EÜect of hydrogen-bonding additives In order to reveal how hydrogen-bonding in—uences the rate and the mechanism of the deactivation processes originating from the singlet excited state, we systematically varied the proton affinity, the hydrogen-bonding ability and the aromaticity of the additives using both hydrogen-bond donors and acceptors. Tri—uorethanol (TFE).We have previously shown that the —uorescence lifetime and quantum yield of N-methyl-1,8- naphthalimide considerably increase16 in TFE, a solvent which has high hydrogen-bonding power. In the present work, we extend these studies to hydroxy-naphthalimides, In contrast with that found for the unsubstituted N-methyl-1,8- naphthalimide, addition of 0.035 M TFE does not in—uence the —uorescent behavior of its hydroxy-derivatives in CH2Cl2 . A ten times higher amount of TFE caused a ca. 9 nm bathochromic shift of the —uorescence maxima for both the 3- and the 4-substituted derivatives, but neither —uorescence quenching nor appearance of a new emission band was observed. The excitation and ground state absorption spectra exhibited a red-shift as a function of the TFE concentration. The clear isosbestic points in the absorption spectra demonstrated that a 1 : 1 hydrogen-bonded complex formed with TFE. The equilibrium constants for hydrogen-bonding (K) were determined using the following relationship.26 [1[(A0/A)j]/[additive]\[K]K(eC/eA)j(A0/A)j (5) where is the ratio of the molar absorption coefficients (eC/eA)j for the complexed and free hydroxy-naphthalimides at a particular wavelength (j), and A denotes the absorbances in A0 Table 1 Photophysical properties of hydroxy-naphthalimides in CH2Cl2 jabs max/nm jFmax/nm E(SI)/kJ mol~1 UF qF/ns UISC kF/107 s~1 kIC/107 s~1 kISC/107 s~1 NIa 348 378 334 0.031 0.14 0.94b 22 21b 670b 3-HONI 376 411 308 0.13 1.9 0.50 6.8 19 26 4-HONI 362 440 300 0.85 9.0 0.03 9.4 1.3 0.3 a Ref. 16. b Based on triplet yield in acetonitrile. Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 4761the presence and the absence of TFE, respectively. Plotting the left-hand side of the function against gives good (A0/A)j linear correlation. From the intercepts, 1.5 M~1 and 1.1 M~1 hydrogen-bonding equilibrium constants were obtained for 3-HONI and 4-HONI, respectively. These values are comparable with that reported for —uorenone»TFE complex (0.7 M~1 in where TFE is also attached to a carbonyl CH2Cl2)4 moiety.Dimethylsulfoxide (DMSO). In contrast to the small eÜect of TFE hydrogen-bond acceptors, which interact with the HO-substituent, DMSO considerably alters the spectral behavior of hydroxy-naphthalimides. When weakly basic but strongly hydrogen-bonding DMSO is added to the hydroxynaphthalimide solutions, characteristic changes are observed, which are demonstrated for 3-HONI in Fig. 3. The red-shift of the long-wavelength absorption band and the clear isosbestic points at 337, 346.5, 380 nm for the 3-hydroxy and at 366.5 nm for the 4-hydroxy derivatives indicate 1 : 1 hydrogenbonded complex formation. From the eÜect of DMSO on the absorption spectra, the hydrogen-bonding equilibrium constants (K) were derived using eqn. (5). The average values of K calculated from the data measured at diÜerent wavelengths are 101^15 M~1 and 191^13 M~1 for the 3-HONI and 4-HONI complexes, respectively.The —uorescence maximum gradually shifts to lower energies with increasing DMSO concentration and an isoemissive point appears but the quantum yield of —uorescence does not change signi–cantly. As the red-shift is only ca. 11 nm, we assign the new —uorescence band to the hydrogen-bonded complex. This view is supported by the fact that DMSO does Fig. 3 Absorption and —uorescence spectra of 3-HONI in at CH2Cl2 diÜerent DMSO concentrations (0, 0.0014, 0.0035, 0.007, 0.014, 0.021, 0.035 M).not bring about any signi–cant change in the spectra of methoxy-naphthalimides, where no hydrogen-bonding is feasible. In addition, we can exclude proton transfer from the OHgroup to DMSO because no —uorescence was detected in the 500»800 nm range, where emission from the conjugate base (anion) of the hydroxy-naphthalimides is expected (Fig. 2). The excitation and the absorption spectra showed similar changes indicating that hydrogen-bonding with DMSO in the ground state is the dominant process.The variation of the —uorescence intensity (I) as the function of [DMSO] was analyzed using a relationship analogous to eqn. (5). [1[(I0/I)F]/[additive]\[K]K(eC/eA)j(UC/UA)F(I0/I)F (6) where represents the ratio of —uorescence intensities in (I0/I)F the presence and the absence of DMSO, and is the (UC/UA)F ratio of —uorescence efficiencies for the complex and the free —uorophore at the detection wavelength.26 Fitting this function to the experimental data leads to K values of 85^9 M~1 and 195^23 M~1 for the equilibrium constants for hydrogen-bonding of DMSO with 3-HONI and 4-HONI, respectively.The good agreement of these results with the corresponding values derived from absorption spectroscopic studies indicates that excitation of the complex does not cause signi–cant change in the hydrogen-bonding equilibrium constant, i.e., K is not signi–cantly diÜerent for the excited and the ground state complex.Time-resolved —uorescence measurements proved that the dynamic quenching by DMSO is negligible. The lifetimes of the singlet excited hydrogenbonded complexes of 3-HONI and 4-HONI are 2.2 ns and 9.3 ns, respectively. Since the lifetimes of the free and the complexed molecules agree very closely, we conclude that no energy dissipation takes place via the hydrogen-bond with DMSO. Pyridine. Addition of pyridine to the hydroxynaphthalimide solutions in leads not only to CH2Cl2 hydrogen-bonding in the ground state but also to —uorescence quenching.The change in the absorption spectrum closely resembles that found in the case of DMSO. Plotting the absorbances according to eqn. (5), we determined the equilibrium constant of hydrogen-bonding (K). The average values of K derived from measurements at various wavelengths are summarized in Table 2. The increase of the pyridine concentration in the 0»0.06 M range results in considerable —uorescence quenching but neither the appearance of a new band nor a shift of the maximum can be seen in the —uorescence spectra.Based on these observations, we conclude that the hydroxynaphthalimide »pyridine hydrogen-bonded complexes have negligible —uorescence yield. The Stern»Volmer plot of the Table 2 Hydrogen-bonding equilibrium constants and rate constants of —uorescence quenching Proton Aromaticity Ka Kb kq c Quencher af–nity/kJ mol~1 index, Ix absorption/M~1 —uorescence/M~1 lifetime/ 109 M~1 s~1 3-Hydroxy-naphthalimide» Imidazole 942.8 64.0 240 183 13 Pyridine 930.0 85.7 46 47 8.8 Pyrazole 894.1 73.0 21 17 9.3 Benzoxazol 891.6 38.0 d d 0.75 DMSO 884.4 » 101 85 e Isoxazole 848.6 47.0 d d e 4-Hydroxy-naphthalimide» Imidazole 942.8 64.0 900 615 9.6 Pyridine 930.0 85.7 138 137 8.9 Pyrazole 894.1 73.0 44 26 7.6 Benzoxazol 891.6 38.0 d d 1.8 DMSO 884.4 191 195 e Isoxazole 848.6 47.0 d d e a From absorption spectra.b From —uorescence spectra. c From —uorescence lifetime measurements.d No hydrogen-bonding can be detected. e No quenching. 4762 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766steady-state —uorescence intensities in the absence and the (I0) presence of pyridine (I) shows an upward curvature (Fig. 4). This suggests that the —uorophore can be quenched both in dynamic and static processes. If the static quenching is attributed entirely to ground state hydrogen-bonding, the modi–ed form of the Stern»Volmer equation describes the variation of vs.quencher concentration.27 I0/I I0/I\(1]K[quencher])(1]kq q0[quencher]) (7) where refers to the lifetime of singlet excited hydroxy- q0 naphthalimide, is the rate constant of the dynamic quen- kq ching and K denotes the equilibrium constant of complex formation in the ground state. The quenching rate constants were determined by time-resolved —uorescence technique (kq) (vide infra) and the K values were calculated by the nonlinear least-square –t of eqn.(7) to the experimental data. It is apparent in Fig. 4 that the calculated curves describe very well the experimental results. Table 2 demonstrates that the hydrogenbonding equilibrium constants derived from both absorption and —uorescence measurements closely agree. Addition of pyridine to the solutions of hydroxynaphthalimides shortens the lifetime of the lowest singlet excited state. The —uorescence decays are well described by a single exponential function.The —uorescence lifetimes in the absence and the presence (q) of pyridine are plotted in Fig. (q0) 4 based on the following equation: q0/q\1]kq q0[quencher] (8) Since the —uorescence decay times are in—uenced only by dynamic quenching, linear correlation is found between q0/q and the quencher concentration. The quenching rate constants derived from the slopes are given in the last column of (kq) Table 2. Benzoxazole and isoxazole. In order to reveal the major factors controlling the rate of the hydrogen-bonding induced —uorescence quenching, we extended our studies to heterocyclic compounds containing a –ve-membered ring.Isoxazole aÜects neither the —uorescence decay nor the spectral characteristics of hydroxy-naphthalimides indicating that no interaction occurs between these compounds either in the ground or the excited state. However, the eÜect of benzoxazole resembles that observed for pyridine. Time-resolved —uorescence measurements proved that dynamic quenching takes place but the reaction rate is much lower than that of pyridine. No clear indication was found for ground state hydrogen-bonding because of the overlap between the benzoxazole and the hydroxy-naphthalimide absorption. Pyrazole and imidazole.Compounds containing two heterocyclic nitrogens in a –ve-membered ring induce a diÜerent type of —uorescent behavior. They not only quench the —uorescence of hydroxy-naphthalimides but also cause a new —uorescence in the 500»800 nm spectral range whose intensity increases with the quencher concentration.It is apparent in Fig. 5 that the new emission consists of two bands for 3- HONI]imidazole and 3-HONI]N-methylimidazole solutions, whereas in the other cases no such clear evidence can be observed for dual luminescence in the long wavelength band. The short wavelength (SW) emission around 370»480 nm originates from the excited hydroxy-naphthalimides. The —uorescence lifetimes measured in this band decrease with increasing concentration of pyrazole and imidazole.This proves that dynamic quenching occurs. Fig. 6 gives the Stern»Volmer plots of the data obtained by steady-state and time-resolved —uorescence techniques for HONI»imidazole systems. In contrast with the linear dependence obtained from lifetime measurements, the Stern»Volmer plots of the steady-state —uorescence intensities are concave indicating that static quenching has an important contribution as well.The redshifts and the isosbestic points in the absorption spectra suggest 1 : 1 hydrogen-bonding. The experimental data were analyzed as described above for the other additives and the results are included in Table 2. The most interesting feature of the spectra in Fig. 5 is the appearance of the long-wavelength (LW) emissions, which are attributed to the singlet excited complexes of hydroxynaphthalimide with pyrazole and imidazole. The extent of proton transfer within these complexes is probably very sensitive to the acid»base properties of the constituents and the local polarity of the solvate shell.In the case of 3-HONI, the SW and the LW bands are wellseparated, therefore, we could readily see the formation and the decay of the species emitting in these spectral ranges. The variation of the —uorescence intensity as the function of time is presented in Fig. 7 for the solution containing 3-HONI and 0.024 M imidazole. The dotted line in Fig. 7A exhibits the —uorescence decays detected at 405 and 640 nm, whereas the continuous lines represent the –tted curves (vide infra), which were calculated by a non-linear least-squares deconvolution method. The parameter describing the rise of the signal at 640 nm (1.3 ns) perfectly agrees with the decay parameter obtained at 405 nm. The excitation pulse pro–le (dotted line) and the growing in of the —uorescence at 640 nm are shown in Fig. 7B using a better time resolution.The calculated curves match the measured data so well that they are hardly distinguishable in the –gure. The —uorescence decay at 405 nm can be well described with a single exponential function. (The signal shown in Fig. 7A has 1.3 ns lifetime.) The time-resolved —uorescence of the 3-HONI»imidazole complex exhibits more complex kinetics. The data were analyzed with a double exponential function : C2 exp([t/q2)[C1 exp([t/q1) (9) where t denotes time, and are constants. Calculation C1 C2 resulted in ns and ns for the decay constants q1\1.3 q2\4.9 when 0.024 M imidazole concentration was used.The C1/C2 ratio is expected to be 1 if the species emitting at long wavelengths is produced only in the quenching reaction.28 We found which clearly indicates that both the C2/C1\1.57, direct excitation of the ground state hydrogen-bonded complex and the dynamic quenching of the singlet excited 3-HONI result in LW emission. As it is expected for a pseudo- –rst-order process, the growing in of the LW —uorescence and the decay of the SW —uorescence strongly depend on the concentration of the quencher.However, the decay time of the LW emission only slightly decreases with the quencher concentration. It is evident from the spectra displayed in Fig. 5 that the excited hydrogen-bonded complexes of pyrazole have diÜerent characteristics compared with that of imidazole and Nmethylimidazole. In the former case, the LW emissions have a Gaussian shape with maxima at 600 nm and 520 nm for 3-HONI and 4-HONI, respectively.Since these maxima are at higher energies than the —uorescence peak of the deprotonated hydroxy-naphthalimides (Fig. 2), we suggest that only a partial proton shift takes place along the hydrogen-bond in these excited complexes. It is especially noteworthy that the addition of imidazole to the 3-HONI solution leads to structured LW —uorescence, which can be resolved to two components (Fig. 5A).In order to exclude the possibility that one of the LW —uorescence components originates from the interaction of 3-HONI with imidazole dimer29 we studied the reactions of Nmethylimidazole as well. This compound is not able to form a hydrogen-bonded dimer because it does not contain an N»H moiety. Fig. 5A demonstrates that the structured LW emission appearing in the presence of N-methylimidazole resembles that obtained with imidazole. Thus, we can rule out that Phys. Chem. Chem.Phys., 1999, 1, 4759»4766 4763Fig. 4 Stern»Volmer plots of the results obtained by time-resolved and steady-state —uorescence technique for HONI»pyridine systems in (A) 3-HONI: steady-state measurements, time- CH2Cl2. >, Ö, resolved measurements. (B) 4-HONI: steady-state measurements, |, time-resolved measurements. L, dimerization of the additive causes the dual emission at long wavelengths. The relative intensity of the two —uorescence bands in the 500»800 nm spectral domain is temperature dependent both in ethyl acetate and in The substantial increase of CH2Cl2 .the higher energy component is particularly discernible in the 295»181 K temperature range in ethyl acetate where the two bands are better separated than in CH2Cl2 . It is readily seen in Fig. 8 that the intensity ratio of the two emissions in the 500»800 nm region strongly depends on the media. In acetonitrile, the band with a maximum around 550 nm disappears, and, likewise, addition of ethanol in the solution signi–cantly weakens this emission. Our CH2Cl2 results indicate that in solvents of medium polarity the 3- HONI»imidazole excited complexes have two dominant struc- Fig. 5 Fluorescence spectra in (A) 3-HONI in the presence CH2Cl2 . of 0.156 M pyrazole (heavy line), 0.024 M imidazole (dotted line) and 0.024 M N-methylimidazole (thin line) ; (B) 4-HONI in the presence of 0.057 M pyrazole (heavy line), 0.018 M imidazole (dotted line) and 0.018 M N-methylimidazole (thin line).Fig. 6 Stern»Volmer plots of the data obtained by steady-state and time-resolved —uorescence techniques for the HONI»imidazole systems in (A) 3-HONI: steady-state measurements; CH2Cl2. >, Ö, time-resolved measurements. (B) 4-HONI: steady-state measure- |, ments; time-resolved measurements. L, tures which diÜer in the extent of their proton shift. However, the complex that —uoresces at higher energies is bound with a hydrogen-bond and possesses only a limited proton transfer character when the proton is removed from the HO group toward the heterocyclic nitrogen in the species emitting at long wavelengths.The identical —uorescence decay times throughout the LW bands of the 3-HONI»imidazole complex in suggest that a fast equilibrium is established CH2Cl2 between the two types of complex. The polar solvents weaken the hydrogen-bond and promote proton transfer, therefore, no dual emission can be seen in acetonitrile. The ion-pair character of the complex in acetonitrile is supported by the fact that the —uorescence maximum of both the 3-HONI»imidazole complex (Fig. 8) and the 3-HONI anion (Fig. 2A) are located around 630 nm in this polar solvent. Fig. 7 Fluorescence decays in 3-HONI]0.024 M imidazole solution in (A) Fluorescence decay (dotted line) and –tted curve CH2Cl2 . (continuous line) at 405 nm and 640 nm. (B) Excitation pulse pro–le (dotted line), —uorescence growing in and –tted curve (continuous line) at 640 nm. 4764 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766Fig. 8 Fluorescence spectra of 3-HONI in the presence of 0.024 M imidazole: (A) in (B) in M EtOH, (C) in the CH2Cl2, CH2 Cl2]0.17 presence of 0.052 M imidazole in acetonitrile and (D) in ethyl acetate. Comparison of the eÜects of the various hydrogen-bond acceptors. It was demonstrated that the intermolecular hydrogen-bonding with alcohols in the singlet excited state acts as an eÜective accepting mode of radiationless deactivation for aromatic carbonyl compounds.2h6 We should comment on the question of why hydroxy-naphthalimides, which also contain carbonyl groups, are not quenched by the strong hydrogen-bond donor TFE.For the example of 2- substituted —uorenones we demonstrated that efficient hydrogen-bonding induced internal conversion can occur only if the carbonyl oxygen has high electron density in the excited state.30 As stated above, the absorption and the —uorescence spectra of hydroxy-naphthalimides exhibit small solvatochromic shifts because excitation leads to minor change in the dipole moment. Theoretical calculations also corroborated that there is no signi–cant diÜerence in the electron density of the carbonyl oxygen for the and the states of naphthali- S0 S1 mides,12,31 therefore, hydrogen-bonding with TFE does not accelerate the internal conversion process.Table 2 lists the hydrogen-bonding equilibrium constants and the quenching rate constants obtained by diÜerent methods. The proton affinities32 and the Bird aromaticity indices33 for the quenchers are also included.Katritzky et al. showed that the Bird aromaticity index is the best measure of the classical aromaticity,34 therefore, we chose this among the various aromaticity parameters available in the literature. The rate constant of the excited hydroxy-naphthalimide quenching varies remarkably with the molecular structure of the hydrogen-bond acceptor ; no quenching takes place with isoxazole but the reaction is diÜusion controlled in the case of imidazole. A parallel change can be seen between the proton affinity and the rate constants throughout the series of the kq quenchers shown in Table 2.The reactants that have low proton affinity do not promote the deactivation of the singlet excited hydroxy-naphthalimides. The hydrogen-bonding power of the additives, as measured directly by the ground state hydrogen-bonding equilibrium constants (K), does not play a rate determining role because no correlation can be found between the K and quantities.For example, DMSO kq does not quench the —uorescence of hydroxy-naphthalimides in spite of the fairly large hydrogen-bonding equilibrium constant in the ground state. These results suggest that proton displacement plays a crucial role in the interaction of excited hydroxy-naphthalimides and hydrogen-bond acceptors. It is not surprising that no correlation appears between the ground state hydrogen-bonding equilibrium constants and the proton affinities listed in Table 2.Gurka and Taft established that hydrogen-bonding and basicity are unrelated.35 For example, using a common hydrogen-bond donor, they showed that the of a carbonyl compound is 13 powers of ten less pKa than that of the corresponding amine for equal values of hydrogen-bonding equilibrium constant.36 Moreover, the data reported by Abraham et al. demonstrate that DMSO forms stronger hydrogen-bonds than the much more basic pyridine derivatives.37 Coupled electron»proton movement was suggested to promote the radiationless deactivation when a heterocyclic molecule containing an aromatic p-electronic system is connected to excited hydroxyarenes directly via a hydrogenbond. 10 The extremely weak —uorescence for the hydrogen-bonded complexes of benzoxazole and pyridine with hydroxy-naphthalimides is probably due to a rapid internal conversion via a similar process.The proton shift toward the hydrogen-bond acceptor induces efficient nonradiative energy dissipation. However, the intensive emission as well as the long lifetime (ca. 4»9 ns) of the excited hydrogen-bonded complexes containing imidazole and pyrazole obviously indicate slow internal conversion value of ca. 108 s~1 can be (kIC deduced from the experimental data). We did not –nd a correlation between the aromaticity index of the hydrogen-bond acceptor and the radiationless deactivation rate of the excited hydrogen-bonded complex.This seems to indicate that the energy dissipation mechanism suggested for the excited hydroxyarene»pyridine species does not play a dominant role if other types of nitrogen heterocyclics serve as the hydrogenbond acceptor. Acknowledgements very much appreciate the support of this work by the We Hungarian Science Foundation (OTKA, Grant T 023428) and the scienti–c exchange program between the French Ministry of Foreign AÜairs and the Hungarian Committee for Technological Development (Balaton Project F-10/97).References 1 J. Herbich, C.-Y. Hung, R. P. Thummel and J. Waluk, J. Am. Chem. Soc., 1996, 118, 3508 and references therein. 2 H. Inoue, M. Hida, N. Nakashima and K. Yoshihara, J. Phys. Chem., 1982, 86, 3184. 3 R. S. Moog, N. A. Burozski, M. M. Desai, W. R. Good, C. D. Silvers, P. A. Thompson and J. D. Simon, J. Phys. Chem., 1991, 95, 8466; J. Ritter, H. U. Borst, T. Lindner, M. Hauser, S.Brosig, K. Bredereck, U. E. Steiner, D. Kué hn, J. Kelemen and H. E. A. Kramer, J. Photochem. Photobiol. A, 1988, 41, 227; H. U. Borst, J. Kelemen, J. Fabian, M. Nepras and H. E. A. Kramer, J. Photochem. Photobiol. A, 1992, 69, 97. 4 L. Biczoç k, T. Beç rces and H. Linschitz, J. Am. Chem. Soc., 1997, 119, 11071. 5 T. Yatsuhashi and H. Inoue, J. Phys. Chem. A, 1997, 101, 8166. 6 T. Yatsuhashi, Y. Nakajima, T. Shimada, H. Tachibana and H. Inoue, J. Phys. Chem.A, 1998, 102, 8657. 7 C. Turroç , C. K. Chang, G. E. Leroi, R. I. Cukier and D. G. Nocera, J. Am. Chem. Soc., 1992, 114, 4013. 8 R. I. Cukier and D. G. Nocera, Annu. Rev. Phys. Chem., 1998, 49, 337. 9 W. J. Liegh, E. C. Lathioor and M. J. St Pierre, J. Am. Chem. Soc., 1996, 118, 12339. 10 N. Mataga and H. Miyasaka, Prog. React. Kinet., 1994, 19, 317 and references therein. 11 H. Miyasaka, K. Wada, S. Ojima and N. Mataga, Isr. J. Chem., 1993, 33, 183. 12 V. Wintgens, P. Valat, J.Kossanyi, A. Demeter, L. Biczoç k and T. Beç rces, New J. Chem., 1996, 20, 1149. 13 L. M. Tolbert and J. E. Haubrich, J. Am. Chem. Soc., 1990, 112, 8163. 14 L. M. Tolbert and J. E. Haubrich, J. Am. Chem. Soc., 1994, 116, 10593. 15 D. Huppert, L. M. Tolbert and S. Linares-Samaniego, J. Phys. Chem. A, 1997, 101, 4602. 16 V. Wintgens, P. Valat, J. Kossanyi, L. Biczoç k, A. Demeter and T. Beç rces, J. Chem. Soc., Faraday T rans., 1994, 90, 411. 17 W. Adam, X. Quian and C.R. Saha-Moé ller, T etrahedron, 1993, 49, 417. Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 476518 R. Royer, J. P. Buisson, P. Demerseman and J. P. Lechartier, Bull. Soc. Chim. Fr., 1969, 2792. 19 W. M. Rodionow and A. M. Fedorowa, Bull. Soc. Chim. Fr., 1939, 479. 20 P. Valat, V. Wintgens, J. Kossanyi, L. Biczoç k, A. Demeter and T. Beç rces, J. Am. Chem. Soc., 1992, 114, 947. 21 S. L. Murov, G. L. Carmichael and I. Hug, Handbook of Photochemistry, Marcel Dekker, New York, 2nd edn., 1993. 22 J. K. Hurley, N. Sinai and H. Linschitz, Photochem. Photobiol., 1983, 38, 9. 23 T. Foé rster, Z. Electrochem., 1950, 54, 531. 24 The HyperChem program package was used. The geometry of the molecule was optimized by AM1 calculation and the energy levels were obtained by ZINDO/S method. 25 A. Prado, J. Campanario, J. M. L. Poyato, J. J. Camacho, D. Reyman, E. Martin, T HEOCHEM., 1988, 166, 463. 26 N. Mataga and S. Tsuno, Bull. Chem. Soc. Jpn., 1957, 30, 368. 27 J.R. Lakowicz, Principles of —uorescence spectroscopy, Plenum Press, New York, 1983, p. 266. 28 D. V. OœConnor, L. Chewter and D. Phillips, J. Phys. Chem., 1982, 86, 3400. 29 E. Fischer, Ber. Bunsen-Ges. Phys. Chem., 1969, 73, 1007. 30 L. Biczoç k, T. Beç rces and H. Inoue, J. Phys. Chem. A., 1999, 103, 3837. 31 M. Adachi, Y. Murata and S. Nakamura, J. Phys. Chem., 1995, 99, 14240. 32 E. P. L. Hunter and S. G. Lias, J. Phys. Chem., Ref. Data, 1998, 27, 707. 33 C. W. Bird, T etrahedron, 1985, 41, 1409; ibid., T etrahedron, 1986, 42, 89. 34 A. R. Katritzky, M. Karelson and N. Malhotra, Heterocycles, 1991, 32, 127. 35 D. Gurka and R. W. Taft, J. Am. Chem. Soc., 1969, 91, 4794. 36 R. W. Taft, D. Gurka, L. Joris, P. von R. Schleyer and J. W. Rakshys, J. Am. Chem. Soc., 1969, 91, 4801. 37 M. H. Abraham, P. P. Duce, D. V. Prior, D. G. Barrett, J. J. Morris and P. J. Taylor, J. Chem. Soc., Perkin T rans. 2, 1989. 1355. Paper 9/04520A 4766 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 EÜect of molecular structure and hydrogen bonding on the —uorescence of hydroxy-substituted naphthalimides Laç szloç Biczoç k,*a Pierre Valatb and Veç ronique Wintgensb a Chemical Research Center, Hungarian Academy of Sciences, P.O.Box 17, 1525 Budapest, Hungary b L aboratoire des C.N.R.S., E.R. 241, 2, 8 rue H. Dunant, 94320 T hiais, Mateç riaux Moleç culaires, France Received 7th June 1999, Accepted 23rd August 1999 Fluorescence properties of hydroxy-naphthalimides were studied in methylene chloride in the absence and the presence of hydrogen-bonding additives.The position of the HO-substituent only slightly aÜects the radiative rate, however, the triplet yield and the rate of the radiationless processes are considerably higher for the 3-hydroxy derivative. Addition of nitrogen-heterocyclic compounds leads not only to hydrogen-bonding in the ground state but also —uorescence quenching. The parallel change throughout the series of the hydrogen-bond acceptors between the proton affinity and the rate constants of dynamic quenching indicates that proton displacement plays a crucial role in the excited hydrogen-bonded complexes.Interaction of hydroxy-naphthalimides with pyridine and benzoxazole results in rapid radiationless deactivation from the singlet excited state, whereas intense emission as well as long —uorescence lifetime characterize imidazole and pyrazole complexes. The dual emission of the imidazole complexes observed in solvents of medium polarity is assigned to two conformers which diÜer in the extent of the proton shift along the hydrogen-bond.Introduction The molecular mechanism of the excited state relaxation induced by intermolecular hydrogen-bond formation is of great interest because it belongs to the most fundamental processes of photochemistry. Most of the studies in this –eld have dealt with the eÜect of the hydrogen-bond donors on the —uorescent properties of aromatic heterocyclic1 and carbonyl compounds.2h6 These molecules form hydrogen-bonded complexes with alcohols and hydroperoxides in the excited state and the hydrogen-bond acts as an efficient vibronic dissipative mode in the nonradiative transition.2h6 On the other hand, coupled electron»proton movement was found to play a dominant role in the deactivation of excited hydrogen-bonded porphyrins7,8 and Ru(II)polypyridyl complexes8 as well as in the interaction of excited ketones with phenols.4,9 Picosecond laser photolysis studies established that excited hydrogen-bond donors, such as aromatic TN»H or »O»H compounds, are efficiently quenched by pyridine derivatives via non-—uorescent hydrogen-bonded complex, in which associated electron and proton displacement facilitates the charge transfer interaction between the two conjugate pelectronic systems.10 However, when phenols form hydrogenbonded complexes with aliphatic amines no charge delocalization is possible along the hydrogen-bond and photoexcitation induces proton transfer.11 The present paper focuses on the question of how the variation of the molecular structure of the hydrogen-bonding additive in—uences the —uorescent properties and the deactivation mechanism of the excited molecules.In order to reveal the role of aromaticity and proton affinity in the hydrogenbonding induced quenching process, aromatic heterocyclic compounds were used as hydrogen-bond acceptors.We show examples where the excited hydrogen-bonded complexes emit dual —uorescence, which is assigned to two conformers diÜering in the extent of the proton shift along the hydrogen-bond. Hydroxy-substituted naphthalimides were chosen as model compounds in these studies because they have both hydrogenbond donor and acceptor moieties, and, on the basis of our previous studies,12 they are expected to be strongly —uorescent. In addition, the electron withdrawing character of the imide group probably enhances the acidity in the excited state and light absorption may serve as an ultrafast trigger for the proton transfer reaction.Recent studies of related compounds demonstrated that substitution of naphthols with cyano or methanesulfonyl groups markedly increases the photoacidity. 13h15 Another main goal of this work was to examine how the introduction of a hydroxy substituent into the 1,8-naphthalimide moiety alters the dominant energy dissipating pathways occurring from the singlet excited state.The investigated compounds are given in Scheme 1. Experimental Acetonitrile, methylene chloride (Prolabo, HPLC grade), dimethylsulfoxide (Merck, spectroscopic grade) and tri- —uoroethanol (Aldrich) were used as received. Benzoxazole, imidazole, isoxazole, pyrazole and pyridine were purchased from Aldrich (highest quality available). N-Methyl-1,8-naphthalimide (NI), also called 2-methyl-1Hbenz[ d,e]isoquinoline-1,3(2H)-dione, was available from our previous study.16 4-Hydroxy-N-methyl-1,8-naphthalimide (4- HONI), also called 6-hydroxy-2-methyl-1H-benz[d,e] Scheme 1 Phys.Chem. Chem. Phys., 1999, 1, 4759»4766 4759 This journal is The Owner Societies 1999 (isoquinoline-1,3(2H)-dione, was prepared by demethylation of 4-methoxy-N-methyl-1,8-naphthalimide,17 following a procedure similar to one described previously.18 One equivalent of the methoxy derivative mixed with 40 equivalents of pyridine hydrochloride was heated at 190 °C under argon for 35 min.After cooling, the solid reaction mixture was added to an aqueous HCl solution (1 M). The solid was –ltered, washed with aqueous HCl solution (1 M) and water. The crude product was puri–ed by dissolution in an aqueous Na2CO3 solution, followed by extraction with ether and precipitation by the addition of perchloric acid in the aqueous phase (45% yield), m.p. 298»302 °C (lit.18 303»305 °C). 3-Hydroxy-N-methyl-1,8-naphthalimide (3-HONI), also called 5-hydroxy-2-methyl-1H-benz[d,e]isoquinoline-1,3(2H)- dione, was synthesized via four reaction steps. First, 3-nitro- N-methyl-1,8-naphthalimide was prepared by condensation of 3-nitro-1,8-naphthalic anhydride (Aldrich) with methylamine hydrochloride in acetic acid.Then 1 equivalent of 3-nitro-Nmethyl- 1,8-naphthalimide was reduced by 4 equivalents of tin(II)chloride in hot hydrochloric acid.19 The obtained amino derivative was diazotized and the diazonium salt was hydrolyzed by aqueous HCl solution.The –nal product was puri–ed by the method described for 4-HONI (15% yield, m.p. 249» 251 °C). The UV»visible absorption spectra were obtained with a Varian»Cary model 50 Bio apparatus. Fluorescence spectra were recorded with an SLM-Aminco model 8100 device. Fluorescence quantum yields of 3-HONI and 4-HONI were determined by comparison with that of 4-methoxy-N-methyl-1,8- naphthalimide in acetonitrile solution, for which a reference yield of was taken.12 Singlet lifetimes were mea- UF\0.88 sured by excitation with a B.M. Industries frequency-tripled Nd»YAG laser (pulse duration 30 ps FWHM), using the experimental set-up already described.20 Laser —ash photolysis experiments were carried out with 8 ns FWHM pulse of a Nd»YAG laser and the monitoring light from Applied Photophysics xenon lamp passed through the sample perpendicular to the excitation.Intersystem crossing (ISC) quantum yields were determined in oxygen-free solutions relative to triplet benzophenone.We compared the initial triplet»triplet absorbances of the investigated compound (A) at 480 nm with that of the benzophenone reference at 530 nm. The solutions (Aref) had matched absorbances at the excitation wavelength (355 nm). The triplet yields were obtained based on the equa- (UISC) tion UISC\3Uref(A/Aref)(eref/e) (1) using the well-established yield and molar (3Uref\1.00)21 absorption coefficient M~1 cm~1 at 530 nm)22 for (eref\7200 triplet benzophenone, whereas e\10 000 M~1 cm~1 was taken for the triplet molar absorption coefficient of naphthalimides at 480 nm.16 Results and discussion I.Photophysical properties of hydroxy-naphthalimides Absorption and —uorescence spectra. Fig. 1 presents the absorption and —uorescence spectra of hydroxy-naphthalimides in methylene chloride. The absorption spectra resemble those of the corresponding naphthols, however, the introduction of the imide moiety results in a remarkable bathochromic shift of the maxima, which becomes most apparent for the low energy bands.For the energy of the lowest excited singlet states, 308 kJ mol~1 and 300 kJ mol~1 were obtained from the locations of the intersections of the normalized absorption and —uorescence spectra in the case of 3-HONI and 4-HONI, respectively. The energy diÜerence between the –rst absorption band of the hydroxy-naphthalimide and the corresponding naphthol was found to be more considerable for the 4-HO derivative.This clearly indi- Fig. 1 Absorption and —uorescence spectra of 3-HONI (»») and 4-HONI (… … …) in CH2Cl2 . cates the larger extent of conjugation between the electron donating OH and the electron withdrawing imide moiety in 4-HONI. The Stokes-shift of the —uorescence spectrum is more pronounced (78 nm) for 4-HONI compared with that of 3-HONI (35 nm). Changing the solvent from methylene chloride to acetonitrile leads to a 6 nm and 2 nm —uorescence maximum displacement for the former and the latter compounds, respectively.These small solvatochromic shifts together with the small decrease of the state energy with S1 increasing solvent polarity (vide infra) suggest that the lowest excited singlet states have limited charge transfer character for both HONI isomers. Fig. 2 shows the excitation and the —uorescence spectra of the hydroxy-naphthalimides and their conjugated bases in acetonitrile.In this solvent we used 1.5]10~4 M perchloric acid to prevent the dissociation of the OH-moiety and to record solely the spectra corresponding to the phenolic form. In order to deprotonate the OH group, 2 ll of 1 M KOH in methanol was added into 2 ml of hydroxy-naphthalimide solution. Under these conditions, the spectra are assigned to the naphtholate anion and the bands are located at lower energies. Based on the intersection of the normalized excitation and —uorescence spectra, 305 kJ mol~1 and 210 kJ mol~1 were calculated for the energies of the –rst excited singlet state of 3-HONI and its conjugated base, respectively.A much Fig. 2 Fluorescence and excitation spectra of (A) 3-HONI and (B) 4-HONI in acetonitrile. Phenolic form in the presence of 1.5]10~4 M perchloric acid (… … …) and deprotonated anion form in the presence of 1]10~3 M KOH (»»). 4760 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766smaller diÜerence was observed between the energies of S1 4-HONI and its deprotonated form (297 kJ mol~1 and 229 kJ mol~1, respectively).According to the Foé rster cycle,23 these results suggest that 4-HONI undergoes a smaller acidity enhancement upon electronic excitation. The lack of the long wavelength emission in clearly indicates that none of CH2Cl2 the singlet excited hydroxy-naphthalimides are sufficiently acidic to transfer proton to this solvent. Photophysical properties. Table 1 demonstrates that substitution of N-methyl-1,8-naphthalimide (NI) with a hydroxy moiety leads to a considerable change in —uorescence yield —uorescence lifetime and triplet yield Fluores- (UF), (qF) (UISC).cence yields and lifetimes increase more than one order of magnitude through the series of NI, 3-HONI, 4-HONI, whereas the variation of the triplet yields exhibits the opposite tendency; they decrease parallel with the energy of the lowest excited singlet In order to get a deeper insight into the (E(S1)).factors controlling the rate of the energy dissipating pathways, the rate constants for —uorescence intersystem crossing (kF), and internal conversion were derived using the (kISC) (kIC) expressions given below: kF\UF/qF (2) kISC\UISC/qF (3) kIC\(1[UISC[UF)/qF (4) It is seen from the rate constants presented in Table 1 that —uorescence emission is the dominant process from the singlet excited state of 4-HONI, however, radiationless transitions prevail for 3-HONI.In contrast with the very fast intersystem crossing observed for unsubstituted NI,16 triplet formation is fairly slow for both hydroxy-naphthalimides. This is especially true of the 4-HO derivative, where almost negligible was kISC found and the phosphorescence is extremely weak at 77 K in organic glass. In the case of 3-HONI the more efficient intersystem crossing permits the determination of the triplet»triplet absorption and the phosphorescence spectra. The triplet» triplet absorption maxima detected by a laser —ash photolysis technique appear at 450 and 480 nm at ambient temperature in whereas the phosphorescence peaks can be found CH2Cl2 , at 590 and 640 nm at 77 K in 95 : 5 butyronitrile : butyl acetate mixture.From the 0»0 transition of the phosphorescence the energy of the –rst triplet excited state is calculated to be 213 kJ mol~1. This value is lower than the one found for N-methyl-1,8-naphthalimide (221 kJ mol~1) indicating that substitution with an OH group not only decreases the energy of the state but also that of the state.S1 T1 The efficient triplet formation for N-methyl-1,8-naphthalimide was rationalized in terms of the transition (UISC\0.94) from the lowest singlet excited state to a close-lying higher triplet state As we established in a previous paper,16 (Tn).16 this is a thermally enhanced process in moderately and strongly polar solvents, which results in a temperature dependent —uorescent behavior.However, the —uorescence lifetimes of 3-HONI and 4-HONI were found to be temperature independent in the 296»198 K range. Based on these results, we can exclude the thermally activated intersystem crossing pathway for hydroxy-naphthalimides. The electron donating hydroxy group decreases the state energy and thereby, S1 increases the energy gap. With such an increased energy S1»Tn gap, thermal activation is not able to initiate the endothermic transition, consequently, triplet formation can occur S1 ]Tn only by the slower transitions to lower triplet levels.Similar eÜects were observed for 4-methoxy-N-methyl-1,8-naphthalimide, which also has an electron donating moiety.12 The signi–cant diÜerence in the triplet formation rate constant between 3-HONI and 4-HONI (Table 1) can be rationalized based on the relative position of the singlet and the triplet energy levels. Semi-empirical calculation24 showed that the energy gap is larger for 4-HONI compared with S1»T1 that of 3-HONI (144 kJ mol~1 and 132 kJ mol~1 were obtained, respectively).This larger energy gap may lessen the magnitude of the spin»orbit coupling between the and S1 T1 states leading to a slower intersystem crossing process for 4-HONI. It is evident from the data summarized in Table 1 that the introduction of a hydroxy group into the 1,8-naphthalimide moiety in position 3 hardly in—uences the rate constant for internal conversion however, the radiationless (kIC), S1 ]S0 transition is markedly decelerated for 4-HONI.It seems to be a general tendency that the 4-substitution of the 1,8-naphthalimide skeleton with an electron donating group leads to a reduced This eÜect can be observed for HO», and kIC. CH3 O» derivatives alike. Prado et al. suggested25 that the mol- NH2» ecule has a quinoid resonance form if an electron donating group is attached to position 4. This kind of electron displacement, which can occur both in the and the states, may S1 S0 aÜect the vibronic coupling between these states and consequently, may lead to slower internal conversion.II. EÜect of hydrogen-bonding additives In order to reveal how hydrogen-bonding in—uences the rate and the mechanism of the deactivation processes originating from the singlet excited state, we systematically varied the proton affinity, the hydrogen-bonding ability and the aromaticity of the additives using both hydrogen-bond donors and acceptors.Tri—uorethanol (TFE). We have previously shown that the —uorescence lifetime and quantum yield of N-methyl-1,8- naphthalimide considerably increase16 in TFE, a solvent which has high hydrogen-bonding power. In the present work, we extend these studies to hydroxy-naphthalimides, In contrast with that found for the unsubstituted N-methyl-1,8- naphthalimide, addition of 0.035 M TFE does not in—uence the —uorescent behavior of its hydroxy-derivatives in CH2Cl2 . A ten times higher amount of TFE caused a ca. 9 nm bathochromic shift of the —uorescence maxima for both the 3- and the 4-substituted derivatives, but neither —uorescence quenching nor appearance of a new emission band was observed. The excitation and ground state absorption spectra exhibited a red-shift as a function of the TFE concentration. The clear isosbestic points in the absorption spectra demonstrated that a 1 : 1 hydrogen-bonded complex formed with TFE. The equilibrium constants for hydrogen-bonding (K) were determined using the following relationship.26 [1[(A0/A)j]/[additive]\[K]K(eC/eA)j(A0/A)j (5) where is the ratio of the molar absorption coefficients (eC/eA)j for the complexed and free hydroxy-naphthalimides at a particular wavelength (j), and A denotes the absorbances in A0 Table 1 Photophysical properties of hydroxy-naphthalimides in CH2Cl2 jabs max/nm jFmax/nm E(SI)/kJ mol~1 UF qF/ns UISC kF/107 s~1 kIC/107 s~1 kISC/107 s~1 NIa 348 378 334 0.031 0.14 0.94b 22 21b 670b 3-HONI 376 411 308 0.13 1.9 0.50 6.8 19 26 4-HONI 362 440 300 0.85 9.0 0.03 9.4 1.3 0.3 a Ref. 16. b Based on triplet yield in acetonitrile. Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 4761the presence and the absence of TFE, respectively. Plotting the left-hand side of the function against gives good (A0/A)j linear correlation. From the intercepts, 1.5 M~1 and 1.1 M~1 hydrogen-bonding equilibrium constants were obtained for 3-HONI and 4-HONI, respectively.These values are comparable with that reported for —uorenone»TFE complex (0.7 M~1 in where TFE is also attached to a carbonyl CH2Cl2)4 moiety. Dimethylsulfoxide (DMSO). In contrast to the small eÜect of TFE hydrogen-bond acceptors, which interact with the HO-substituent, DMSO considerably alters the spectral behavior of hydroxy-naphthalimides. When weakly basic but strongly hydrogen-bonding DMSO is added to the hydroxynaphthalimide solutions, characteristic changes are observed, which are demonstrated for 3-HONI in Fig. 3. The red-shift of the long-wavelength absorption band and the clear isosbestic points at 337, 346.5, 380 nm for the 3-hydroxy and at 366.5 nm for the 4-hydroxy derivatives indicate 1 : 1 hydrogenbonded complex formation. From the eÜect of DMSO on the absorption spectra, the hydrogen-bonding equilibrium constants (K) were derived using eqn. (5). The average values of K calculated from the data measured at diÜerent wavelengths are 101^15 M~1 and 191^13 M~1 for the 3-HONI and 4-HONI complexes, respectively.The —uorescence maximum gradually shifts to lower energies with increasing DMSO concentration and an isoemissive point appears but the quantum yield of —uorescence does not change signi–cantly. As the red-shift is only ca. 11 nm, we assign the new —uorescence band to the hydrogen-bonded complex. This view is supported by the fact that DMSO does Fig. 3 Absorption and —uorescence spectra of 3-HONI in at CH2Cl2 diÜerent DMSO concentrations (0, 0.0014, 0.0035, 0.007, 0.014, 0.021, 0.035 M).not bring about any signi–cant change in the spectra of methoxy-naphthalimides, where no hydrogen-bonding is feasible. In addition, we can exclude proton transfer from the OHgroup to DMSO because no —uorescence was detected in the 500»800 nm range, where emission from the conjugate base (anion) of the hydroxy-naphthalimides is expected (Fig. 2). The excitation and the absorption spectra showed similar changes indicating that hydrogen-bonding with DMSO in the ground state is the dominant process.The variation of the —uorescence intensity (I) as the function of [DMSO] was analyzed using a relationship analogous to eqn. (5). [1[(I0/I)F]/[additive]\[K]K(eC/eA)j(UC/UA)F(I0/I)F (6) where represents the ratio of —uorescence intensities in (I0/I)F the presence and the absence of DMSO, and is the (UC/UA)F ratio of —uorescence efficiencies for the complex and the free —uorophore at the detection wavelength.26 Fitting this function to the experimental data leads to K values of 85^9 M~1 and 195^23 M~1 for the equilibrium constants for hydrogen-bonding of DMSO with 3-HONI and 4-HONI, respectively.The good agreement of these results with the corresponding values derived from absorption spectroscopic studies indicates that excitation of the complex does not cause signi–cant change in the hydrogen-bonding equilibrium constant, i.e., K is not signi–cantly diÜerent for the excited and the ground state complex.Time-resolved —uorescence measurements proved that the dynamic quenching by DMSO is negligible. The lifetimes of the singlet excited hydrogenbonded complexes of 3-HONI and 4-HONI are 2.2 ns and 9.3 ns, respectively. Since the lifetimes of the free and the complexed molecules agree very closely, we conclude that no energy dissipation takes place via the hydrogen-bond with DMSO. Pyridine. Addition of pyridine to the hydroxynaphthalimide solutions in leads not only to CH2Cl2 hydrogen-bonding in the ground state but also to —uorescence quenching.The change in the absorption spectrum closely resembles that found in the case of DMSO. Plotting the absorbances according to eqn. (5), we determined the equilibrium constant of hydrogen-bonding (K). The average values of K derived from measurements at various wavelengths are summarized in Table 2. The increase of the pyridine concentration in the 0»0.06 M range results in considerable —uorescence quenching but neither the appearance of a new band nor a shift of the maximum can be seen in the —uorescence spectra.Based on these observations, we conclude that the hydroxynaphthalimide »pyridine hydrogen-bonded complexes have negligible —uorescence yield. The Stern»Volmer plot of the Table 2 Hydrogen-bonding equilibrium constants and rate constants of —uorescence quenching Proton Aromaticity Ka Kb kq c Quencher af–nity/kJ mol~1 index, Ix absorption/M~1 —uorescence/M~1 lifetime/ 109 M~1 s~1 3-Hydroxy-naphthalimide» Imidazole 942.8 64.0 240 183 13 Pyridine 930.0 85.7 46 47 8.8 Pyrazole 894.1 73.0 21 17 9.3 Benzoxazol 891.6 38.0 d d 0.75 DMSO 884.4 » 101 85 e Isoxazole 848.6 47.0 d d e 4-Hydroxy-naphthalimide» Imidazole 942.8 64.0 900 615 9.6 Pyridine 930.0 85.7 138 137 8.9 Pyrazole 894.1 73.0 44 26 7.6 Benzoxazol 891.6 38.0 d d 1.8 DMSO 884.4 191 195 e Isoxazole 848.6 47.0 d d e a From absorption spectra.b From —uorescence spectra.c From —uorescence lifetime measurements. d No hydrogen-bonding can be detected. e No quenching. 4762 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766steady-state —uorescence intensities in the absence and the (I0) presence of pyridine (I) shows an upward curvature (Fig. 4). This suggests that the —uorophore can be quenched both in dynamic and static processes. If the static quenching is attributed entirely to ground state hydrogen-bonding, the modi–ed form of the Stern»Volmer equation describes the variation of vs.quencher concentration.27 I0/I I0/I\(1]K[quencher])(1]kq q0[quencher]) (7) where refers to the lifetime of singlet excited hydroxy- q0 naphthalimide, is the rate constant of the dynamic quen- kq ching and K denotes the equilibrium constant of complex formation in the ground state. The quenching rate constants were determined by time-resolved —uorescence technique (kq) (vide infra) and the K values were calculated by the nonlinear least-square –t of eqn.(7) to the experimental data. It is apparent in Fig. 4 that the calculated curves describe very well the experimental results. Table 2 demonstrates that the hydrogenbonding equilibrium constants derived from both absorption and —uorescence measurements closely agree. Addition of pyridine to the solutions of hydroxynaphthalimides shortens the lifetime of the lowest singlet excited state. The —uorescence decays are well described by a single exponential function.The —uorescence lifetimes in the absence and the presence (q) of pyridine are plotted in Fig. (q0) 4 based on the following equation: q0/q\1]kq q0[quencher] (8) Since the —uorescence decay times are in—uenced only by dynamic quenching, linear correlation is found between q0/q and the quencher concentration. The quenching rate constants derived from the slopes are given in the last column of (kq) Table 2. Benzoxazole and isoxazole. In order to reveal the major factors controlling the rate of the hydrogen-bonding induced —uorescence quenching, we extended our studies to heterocyclic compounds containing a –ve-membered ring.Isoxazole aÜects neither the —uorescence decay nor the spectral characteristics of hydroxy-naphthalimides indicating that no interaction occurs between these compounds either in the ground or the excited state. However, the eÜect of benzoxazole resembles that observed for pyridine. Time-resolved —uorescence measurements proved that dynamic quenching takes place but the reaction rate is much lower than that of pyridine.No clear indication was found for ground state hydrogen-bonding because of the overlap between the benzoxazole and the hydroxy-naphthalimide absorption. Pyrazole and imidazole. Compounds containing two heterocyclic nitrogens in a –ve-membered ring induce a diÜerent type of —uorescent behavior. They not only quench the —uorescence of hydroxy-naphthalimides but also cause a new —uorescence in the 500»800 nm spectral range whose intensity increases with the quencher concentration.It is apparent in Fig. 5 that the new emission consists of two bands for 3- HONI]imidazole and 3-HONI]N-methylimidazole solutions, whereas in the other cases no such clear evidence can be observed for dual luminescence in the long wavelength band. The short wavelength (SW) emission around 370»480 nm originates from the excited hydroxy-naphthalimides.The —uorescence lifetimes measured in this band decrease with increasing concentration of pyrazole and imidazole. This proves that dynamic quenching occurs. Fig. 6 gives the Stern»Volmer plots of the data obtained by steady-state and time-resolved —uorescence techniques for HONI»imidazole systems. In contrast with the linear dependence obtained from lifetime measurements, the Stern»Volmer plots of the steady-state —uorescence intensities are concave indicating that static quenching has an important contribution as well.The redshifts and the isosbestic points in the absorption spectra suggest 1 : 1 hydrogen-bonding. The experimental data were analyzed as described above for the other additives and the results are included in Table 2. The most interesting feature of the spectra in Fig. 5 is the appearance of the long-wavelength (LW) emissions, which are attributed to the singlet excited complexes of hydroxynaphthalimide with pyrazole and imidazole.The extent of proton transfer within these complexes is probably very sensitive to the acid»base properties of the constituents and the local polarity of the solvate shell. In the case of 3-HONI, the SW and the LW bands are wellseparated, therefore, we could readily see the formation and the decay of the species emitting in these spectral ranges. The variation of the —uorescence intensity as the function of time is presented in Fig. 7 for the solution containing 3-HONI and 0.024 M imidazole. The dotted line in Fig. 7A exhibits the —uorescence decays detected at 405 and 640 nm, whereas the continuous lines represent the –tted curves (vide infra), which were calculated by a non-linear least-squares deconvolution method. The parameter describing the rise of the signal at 640 nm (1.3 ns) perfectly agrees with the decay parameter obtained at 405 nm. The excitation pulse pro–le (dotted line) and the growing in of the —uorescence at 640 nm are shown in Fig. 7B using a better time resolution. The calculated curves match the measured data so well that they are hardly distinguishable in the –gure. The —uorescence decay at 405 nm can be well described with a single exponential function. (The signal shown in Fig. 7A has 1.3 ns lifetime.) The time-resolved —uorescence of the 3-HONI»imidazole complex exhibits more complex kinetics. The data were analyzed with a double exponential function : C2 exp([t/q2)[C1 exp([t/q1) (9) where t denotes time, and are constants.Calculation C1 C2 resulted in ns and ns for the decay constants q1\1.3 q2\4.9 when 0.024 M imidazole concentration was used. The C1/C2 ratio is expected to be 1 if the species emitting at long wavelengths is produced only in the quenching reaction.28 We found which clearly indicates that both the C2/C1\1.57, direct excitation of the ground state hydrogen-bonded complex and the dynamic quenching of the singlet excited 3-HONI result in LW emission. As it is expected for a pseudo- –rst-order process, the growing in of the LW —uorescence and the decay of the SW —uorescence strongly depend on the concentration of the quencher.However, the decay time of the LW emission only slightly decreases with the quencher concentration. It is evident from the spectra displayed in Fig. 5 that the excited hydrogen-bonded complexes of pyrazole have diÜerent characteristics compared with that of imidazole and Nmethylimidazole. In the former case, the LW emissions have a Gaussian shape with maxima at 600 nm and 520 nm for 3-HONI and 4-HONI, respectively. Since these maxima are at higher energies than the —uorescence peak of the deprotonated hydroxy-naphthalimides (Fig. 2), we suggest that only a partial proton shift takes place along the hydrogen-bond in these excited complexes. It is especially noteworthy that the addition of imidazole to the 3-HONI solution leads to structured LW —uorescence, which can be resolved to two components (Fig. 5A). In order to exclude the possibility that one of the LW —uorescence components originates from the interaction of 3-HONI with imidazole dimer29 we studied the reactions of Nmethylimidazole as well. This compound is not able to form a hydrogen-bonded dimer because it does not contain an N»H moiety. Fig. 5A demonstrates that the structured LW emission appearing in the presence of N-methylimidazole resembles that obtained with imidazole. Thus, we can rule out that Phys.Chem. Chem. Phys., 1999, 1, 4759»4766 4763Fig. 4 Stern»Volmer plots of the results obtained by time-resolved and steady-state —uorescence technique for HONI»pyridine systems in (A) 3-HONI: steady-state measurements, time- CH2Cl2. >, Ö, resolved measurements. (B) 4-HONI: steady-state measurements, |, time-resolved measurements. L, dimerization of the additive causes the dual emission at long wavelengths. The relative intensity of the two —uorescence bands in the 500»800 nm spectral domain is temperature dependent both in ethyl acetate and in The substantial increase of CH2Cl2 .the higher energy component is particularly discernible in the 295»181 K temperature range in ethyl acetate where the two bands are better separated than in CH2Cl2 . It is readily seen in Fig. 8 that the intensity ratio of the two emissions in the 500»800 nm region strongly depends on the media. In acetonitrile, the band with a maximum around 550 nm disappears, and, likewise, addition of ethanol in the solution signi–cantly weakens this emission.Our CH2Cl2 results indicate that in solvents of medium polarity the 3- HONI»imidazole excited complexes have two dominant struc- Fig. 5 Fluorescence spectra in (A) 3-HONI in the presence CH2Cl2 . of 0.156 M pyrazole (heavy line), 0.024 M imidazole (dotted line) and 0.024 M N-methylimidazole (thin line) ; (B) 4-HONI in the presence of 0.057 M pyrazole (heavy line), 0.018 M imidazole (dotted line) and 0.018 M N-methylimidazole (thin line).Fig. 6 Stern»Volmer plots of the data obtained by steady-state and time-resolved —uorescence techniques for the HONI»imidazole systems in (A) 3-HONI: steady-state measurements; CH2Cl2. >, Ö, time-resolved measurements. (B) 4-HONI: steady-state measure- |, ments; time-resolved measurements. L, tures which diÜer in the extent of their proton shift. However, the complex that —uoresces at higher energies is bound with a hydrogen-bond and possesses only a limited proton transfer character when the proton is removed from the HO group toward the heterocyclic nitrogen in the species emitting at long wavelengths. The identical —uorescence decay times throughout the LW bands of the 3-HONI»imidazole complex in suggest that a fast equilibrium is established CH2Cl2 between the two types of complex.The polar solvents weaken the hydrogen-bond and promote proton transfer, therefore, no dual emission can be seen in acetonitrile. The ion-pair character of the complex in acetonitrile is supported by the fact that the —uorescence maximum of both the 3-HONI»imidazole complex (Fig. 8) and the 3-HONI anion (Fig. 2A) are located around 630 nm in this polar solvent. Fig. 7 Fluorescence decays in 3-HONI]0.024 M imidazole solution in (A) Fluorescence decay (dotted line) and –tted curve CH2Cl2 . (continuous line) at 405 nm and 640 nm. (B) Excitation pulse pro–le (dotted line), —uorescence growing in and –tted curve (continuous line) at 640 nm. 4764 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766Fig. 8 Fluorescence spectra of 3-HONI in the presence of 0.024 M imidazole: (A) in (B) in M EtOH, (C) in the CH2Cl2, CH2 Cl2]0.17 presence of 0.052 M imidazole in acetonitrile and (D) in ethyl acetate. Comparison of the eÜects of the various hydrogen-bond acceptors. It was demonstrated that the intermolecular hydrogen-bonding with alcohols in the singlet excited state acts as an eÜective accepting mode of radiationless deactivation for aromatic carbonyl compounds.2h6 We should comment on the question of why hydroxy-naphthalimides, which also contain carbonyl groups, are not quenched by the strong hydrogen-bond donor TFE.For the example of 2- substituted —uorenones we demonstrated that efficient hydrogen-bonding induced internal conversion can occur only if the carbonyl oxygen has high electron density in the excited state.30 As stated above, the absorption and the —uorescence spectra of hydroxy-naphthalimides exhibit small solvatochromic shifts because excitation leads to minor change in the dipole moment.Theoretical calculations also corroborated that there is no signi–cant diÜerence in the electron density of the carbonyl oxygen for the and the states of naphthali- S0 S1 mides,12,31 therefore, hydrogen-bonding with TFE does not accelerate the internal conversion process. Table 2 lists the hydrogen-bonding equilibrium constants and the quenching rate constants obtained by diÜerent methods.The proton affinities32 and the Bird aromaticity indices33 for the quenchers are also included. Katritzky et al. showed that the Bird aromaticity index is the best measure of the classical aromaticity,34 therefore, we chose this among the various aromaticity parameters available in the literature. The rate constant of the excited hydroxy-naphthalimide quenching varies remarkably with the molecular structure of the hydrogen-bond acceptor ; no quenching takes place with isoxazole but the reaction is diÜusion controlled in the case of imidazole.A parallel change can be seen between the proton affinity and the rate constants throughout the series of the kq quenchers shown in Table 2. The reactants that have low proton affinity do not promote the deactivation of the singlet excited hydroxy-naphthalimides. The hydrogen-bonding power of the additives, as measured directly by the ground state hydrogen-bonding equilibrium constants (K), does not play a rate determining role because no correlation can be found between the K and quantities. For example, DMSO kq does not quench the —uorescence of hydroxy-naphthalimides in spite of the fairly large hydrogen-bonding equilibrium constant in the ground state.These results suggest that proton displacement plays a crucial role in the interaction of excited hydroxy-naphthalimides and hydrogen-bond acceptors. It is not surprising that no correlation appears between the ground state hydrogen-bonding equilibrium constants and the proton affinities listed in Table 2.Gurka and Taft established that hydrogen-bonding and basicity are unrelated.35 For example, using a common hydrogen-bond donor, they showed that the of a carbonyl compound is 13 powers of ten less pKa than that of the corresponding amine for equal values of hydrogen-bonding equilibrium constant.36 Moreover, the data reported by Abraham et al.demonstrate that DMSO forms stronger hydrogen-bonds than the much more basic pyridine derivatives.37 Coupled electron»proton movement was suggested to promote the radiationless deactivation when a heterocyclic molecule containing an aromatic p-electronic system is connected to excited hydroxyarenes directly via a hydrogenbond. 10 The extremely weak —uorescence for the hydrogen-bonded complexes of benzoxazole and pyridine with hydroxy-naphthalimides is probably due to a rapid internal conversion via a similar process.The proton shift toward the hydrogen-bond acceptor induces efficient nonradiative energy dissipation. However, the intensive emission as well as the long lifetime (ca. 4»9 ns) of the excited hydrogen-bonded complexes containing imidazole and pyrazole obviously indicate slow internal conversion value of ca. 108 s~1 can be (kIC deduced from the experimental data). We did not –nd a correlation between the aromaticity index of the hydrogen-bond acceptor and the radiationless deactivation rate of the excited hydrogen-bonded complex. This seems to indicate that the energy dissipation mechanism suggested for the excited hydroxyarene»pyridine species does not play a dominant role if other types of nitrogen heterocyclics serve as the hydrogenbond acceptor.Acknowledgements very much appreciate the support of this work by the We Hungarian Science Foundation (OTKA, Grant T 023428) and the scienti–c exchange program between the French Ministry of Foreign AÜairs and the Hungarian Committee for Technological Development (Balaton Project F-10/97).References 1 J. Herbich, C.-Y. Hung, R. P. Thummel and J. Waluk, J. Am. Chem. Soc., 1996, 118, 3508 and references therein. 2 H. Inoue, M. Hida, N. Nakashima and K. Yoshihara, J. Phys. Chem., 1982, 86, 3184. 3 R. S. Moog, N. A. Burozski, M. M. Desai, W. R. Good, C. D. Silvers, P. A. Thompson and J. D. Simon, J. Phys. Chem., 1991, 95, 8466; J.Ritter, H. U. Borst, T. Lindner, M. Hauser, S. Brosig, K. Bredereck, U. E. Steiner, D. Kué hn, J. Kelemen and H. E. A. Kramer, J. Photochem. Photobiol. A, 1988, 41, 227; H. U. Borst, J. Kelemen, J. Fabian, M. Nepras and H. E. A. Kramer, J. Photochem. Photobiol. A, 1992, 69, 97. 4 L. Biczoç k, T. Beç rces and H. Linschitz, J. Am. Chem. Soc., 1997, 119, 11071. 5 T. Yatsuhashi and H. Inoue, J. Phys. Chem. A, 1997, 101, 8166. 6 T. Yatsuhashi, Y. Nakajima, T. Shimada, H. Tachibana and H.Inoue, J. Phys. Chem. A, 1998, 102, 8657. 7 C. Turroç , C. K. Chang, G. E. Leroi, R. I. Cukier and D. G. Nocera, J. Am. Chem. Soc., 1992, 114, 4013. 8 R. I. Cukier and D. G. Nocera, Annu. Rev. Phys. Chem., 1998, 49, 337. 9 W. J. Liegh, E. C. Lathioor and M. J. St Pierre, J. Am. Chem. Soc., 1996, 118, 12339. 10 N. Mataga and H. Miyasaka, Prog. React. Kinet., 1994, 19, 317 and references therein. 11 H. Miyasaka, K. Wada, S. Ojima and N. Mataga, Isr. J. Chem., 1993, 33, 183. 12 V. Wintgens, P. Valat, J. Kossanyi, A. Demeter, L. Biczoç k and T. Beç rces, New J. Chem., 1996, 20, 1149. 13 L. M. Tolbert and J. E. Haubrich, J. Am. Chem. Soc., 1990, 112, 8163. 14 L. M. Tolbert and J. E. Haubrich, J. Am. Chem. Soc., 1994, 116, 10593. 15 D. Huppert, L. M. Tolbert and S. Linares-Samaniego, J. Phys. Chem. A, 1997, 101, 4602. 16 V. Wintgens, P. Valat, J. Kossanyi, L. Biczoç k, A. Demeter and T. Beç rces, J. Chem. Soc., Faraday T rans., 1994, 90, 411. 17 W. Adam, X.Quian and C. R. Saha-Moé ller, T etrahedron, 1993, 49, 417. Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 476518 R. Royer, J. P. Buisson, P. Demerseman and J. P. Lechartier, Bull. Soc. Chim. Fr., 1969, 2792. 19 W. M. Rodionow and A. M. Fedorowa, Bull. Soc. Chim. Fr., 1939, 479. 20 P. Valat, V. Wintgens, J. Kossanyi, L. Biczoç k, A. Demeter and T. Beç rces, J. Am. Chem. Soc., 1992, 114, 947. 21 S. L. Murov, G. L. Carmichael and I. Hug, Handbook of Photochemistry, Marcel Dekker, New York, 2nd edn., 1993. 22 J. K. Hurley, N. Sinai and H. Linschitz, Photochem. Photobiol., 1983, 38, 9. 23 T. Foé rster, Z. Electrochem., 1950, 54, 531. 24 The HyperChem program package was used. The geometry of the molecule was optimized by AM1 calculation and the energy levels were obtained by ZINDO/S method. 25 A. Prado, J. Campanario, J. M. L. Poyato, J. J. Camacho, D. Reyman, E. Martin, T HEOCHEM., 1988, 166, 463. 26 N. Mataga and S. Tsuno, Bull. Chem. Soc. Jpn., 1957, 30, 368. 27 J. R. Lakowicz, Principles of —uorescence spectroscopy, Plenum Press, New York, 1983, p. 266. 28 D. V. OœConnor, L. Chewter and D. Phillips, J. Phys. Chem., 1982, 86, 3400. 29 E. Fischer, Ber. Bunsen-Ges. Phys. Chem., 1969, 73, 1007. 30 L. Biczoç k, T. Beç rces and H. Inoue, J. Phys. Chem. A., 1999, 103, 3837. 31 M. Adachi, Y. Murata and S. Nakamura, J. Phys. Chem., 1995, 99, 14240. 32 E. P. L. Hunter and S. G. Lias, J. Phys. Chem., Ref. Data, 1998, 27, 707. 33 C.W. Bird, T etrahedron, 1985, 41, 1409; ibid., T etrahedron, 1986, 42, 89. 34 A. R. Katritzky, M. Karelson and N. Malhotra, Heterocycles, 1991, 32, 127. 35 D. Gurka and R. W. Taft, J. Am. Chem. Soc., 1969, 91, 4794. 36 R. W. Taft, D. Gurka, L. Joris, P. von R. Schleyer and J. W. Rakshys, J. Am. Chem. Soc., 1969, 91, 4801. 37 M. H. Abraham, P. P. Duce, D. V. Prior, D. G. Barrett, J. J. Morris and P. J. Taylor, J. Chem. Soc., Perkin T rans. 2, 1989. 1355. Paper 9/04520A 4766 Phys.Chem. Chem. Phys., 1999, 1, 4759»4766 EÜect of molecular structure and hydrogen bonding on the —uorescence of hydroxy-substituted naphthalimides Laç szloç Biczoç k,*a Pierre Valatb and Veç ronique Wintgensb a Chemical Research Center, Hungarian Academy of Sciences, P.O. Box 17, 1525 Budapest, Hungary b L aboratoire des C.N.R.S., E.R. 241, 2, 8 rue H. Dunant, 94320 T hiais, Mateç riaux Moleç culaires, France Received 7th June 1999, Accepted 23rd August 1999 Fluorescence properties of hydroxy-naphthalimides were studied in methylene chloride in the absence and the presence of hydrogen-bonding additives.The position of the HO-substituent only slightly aÜects the radiative rate, however, the triplet yield and the rate of the radiationless processes are considerably higher for the 3-hydroxy derivative. Addition of nitrogen-heterocyclic compounds leads not only to hydrogen-bonding in the ground state but also —uorescence quenching. The parallel change throughout the series of the hydrogen-bond acceptors between the proton affinity and the rate constants of dynamic quenching indicates that proton displacement plays a crucial role in the excited hydrogen-bonded complexes.Interaction of hydroxy-naphthalimides with pyridine and benzoxazole results in rapid radiationless deactivation from the singlet excited state, whereas intense emission as well as long —uorescence lifetime characterize imidazole and pyrazole complexes. The dual emission of the imidazole complexes observed in solvents of medium polarity is assigned to two conformers which diÜer in the extent of the proton shift along the hydrogen-bond.Introduction The molecular mechanism of the excited state relaxation induced by intermolecular hydrogen-bond formation is of great interest because it belongs to the most fundamental processes of photochemistry. Most of the studies in this –eld have dealt with the eÜect of the hydrogen-bond donors on the —uorescent properties of aromatic heterocyclic1 and carbonyl compounds.2h6 These molecules form hydrogen-bonded complexes with alcohols and hydroperoxides in the excited state and the hydrogen-bond acts as an efficient vibronic dissipative mode in the nonradiative transition.2h6 On the other hand, coupled electron»proton movement was found to play a dominant role in the deactivation of excited hydrogen-bonded porphyrins7,8 and Ru(II)polypyridyl complexes8 as well as in the interaction of excited ketones with phenols.4,9 Picosecond laser photolysis studies established that excited hydrogen-bond donors, such as aromatic TN»H or »O»H compounds, are efficiently quenched by pyridine derivatives via non-—uorescent hydrogen-bonded complex, in which associated electron and proton displacement facilitates the charge transfer interaction between the two conjugate pelectronic systems.10 However, when phenols form hydrogenbonded complexes with aliphatic amines no charge delocalization is possible along the hydrogen-bond and photoexcitation induces proton transfer.11 The present paper focuses on the question of how the variation of the molecular structure of the hydrogen-bonding additive in—uences the —uorescent properties and the deactivation mechanism of the excited molecules.In order to reveal the role of aromaticity and proton affinity in the hydrogenbonding induced quenching process, aromatic heterocyclic compounds were used as hydrogen-bond acceptors.We show examples where the excited hydrogen-bonded complexes emit dual —uorescence, which is assigned to two conformers diÜering in the extent of the proton shift along the hydrogen-bond. Hydroxy-substituted naphthalimides were chosen as model compounds in these studies because they have both hydrogenbond donor and acceptor moieties, and, on the basis of our previous studies,12 they are expected to be strongly —uorescent. In addition, the electron withdrawing character of the imide group probably enhances the acidity in the excited state and light absorption may serve as an ultrafast trigger for the proton transfer reaction.Recent studies of related compounds demonstrated that substitution of naphthols with cyano or methanesulfonyl groups markedly increases the photoacidity. 13h15 Another main goal of this work was to examine how the introduction of a hydroxy substituent into the 1,8-naphthalimide moiety alters the dominant energy dissipating pathways occurring from the singlet excited state.The investigated compounds are given in Scheme 1. Experimental Acetonitrile, methylene chloride (Prolabo, HPLC grade), dimethylsulfoxide (Merck, spectroscopic grade) and tri- —uoroethanol (Aldrich) were used as received. Benzoxazole, imidazole, isoxazole, pyrazole and pyridine were purchased from Aldrich (highest quality available). N-Methyl-1,8-naphthalimide (NI), also called 2-methyl-1Hbenz[ d,e]isoquinoline-1,3(2H)-dione, was available from our previous study.16 4-Hydroxy-N-methyl-1,8-naphthalimide (4- HONI), also called 6-hydroxy-2-methyl-1H-benz[d,e] Scheme 1 Phys.Chem. Chem. Phys., 1999, 1, 4759»4766 4759 This journal is The Owner Societies 1999 (isoquinoline-1,3(2H)-dione, was prepared by demethylation of 4-methoxy-N-methyl-1,8-naphthalimide,17 following a procedure similar to one described previously.18 One equivalent of the methoxy derivative mixed with 40 equivalents of pyridine hydrochloride was heated at 190 °C under argon for 35 min.After cooling, the solid reaction mixture was added to an aqueous HCl solution (1 M). The solid was –ltered, washed with aqueous HCl solution (1 M) and water. The crude product was puri–ed by dissolution in an aqueous Na2CO3 solution, followed by extraction with ether and precipitation by the addition of perchloric acid in the aqueous phase (45% yield), m.p. 298»302 °C (lit.18 303»305 °C). 3-Hydroxy-N-methyl-1,8-naphthalimide (3-HONI), also called 5-hydroxy-2-methyl-1H-benz[d,e]isoquinoline-1,3(2H)- dione, was synthesized via four reaction steps. First, 3-nitro- N-methyl-1,8-naphthalimide was prepared by condensation of 3-nitro-1,8-naphthalic anhydride (Aldrich) with methylamine hydrochloride in acetic acid. Then 1 equivalent of 3-nitro-Nmethyl- 1,8-naphthalimide was reduced by 4 equivalents of tin(II)chloride in hot hydrochloric acid.19 The obtained amino derivative was diazotized and the diazonium salt was hydrolyzed by aqueous HCl solution.The –nal product was puri–ed by the method described for 4-HONI (15% yield, m.p. 249» 251 °C). The UV»visible absorption spectra were obtained with a Varian»Cary model 50 Bio apparatus. Fluorescence spectra were recorded with an SLM-Aminco model 8100 device. Fluorescence quantum yields of 3-HONI and 4-HONI were determined by comparison with that of 4-methoxy-N-methyl-1,8- naphthalimide in acetonitrile solution, for which a reference yield of was taken.12 Singlet lifetimes were mea- UF\0.88 sured by excitation with a B.M.Industries frequency-tripled Nd»YAG laser (pulse duration 30 ps FWHM), using the experimental set-up already described.20 Laser —ash photolysis experiments were carried out with 8 ns FWHM pulse of a Nd»YAG laser and the monitoring light from Applied Photophysics xenon lamp passed through the sample perpendicular to the excitation. Intersystem crossing (ISC) quantum yields were determined in oxygen-free solutions relative to triplet benzophenone.We compared the initial triplet»triplet absorbances of the investigated compound (A) at 480 nm with that of the benzophenone reference at 530 nm. The solutions (Aref) had matched absorbances at the excitation wavelength (355 nm). The triplet yields were obtained based on the equa- (UISC) tion UISC\3Uref(A/Aref)(eref/e) (1) using the well-established yield and molar (3Uref\1.00)21 absorption coefficient M~1 cm~1 at 530 nm)22 for (eref\7200 triplet benzophenone, whereas e\10 000 M~1 cm~1 was taken for the triplet molar absorption coefficient of naphthalimides at 480 nm.16 Results and discussion I.Photophysical properties of hydroxy-naphthalimides Absorption and —uorescence spectra. Fig. 1 presents the absorption and —uorescence spectra of hydroxy-naphthalimides in methylene chloride. The absorption spectra resemble those of the corresponding naphthols, however, the introduction of the imide moiety results in a remarkable bathochromic shift of the maxima, which becomes most apparent for the low energy bands.For the energy of the lowest excited singlet states, 308 kJ mol~1 and 300 kJ mol~1 were obtained from the locations of the intersections of the normalized absorption and —uorescence spectra in the case of 3-HONI and 4-HONI, respectively. The energy diÜerence between the –rst absorption band of the hydroxy-naphthalimide and the corresponding naphthol was found to be more considerable for the 4-HO derivative.This clearly indi- Fig. 1 Absorption and —uorescence spectra of 3-HONI (»») and 4-HONI (… … …) in CH2Cl2 . cates the larger extent of conjugation between the electron donating OH and the electron withdrawing imide moiety in 4-HONI. The Stokes-shift of the —uorescence spectrum is more pronounced (78 nm) for 4-HONI compared with that of 3-HONI (35 nm). Changing the solvent from methylene chloride to acetonitrile leads to a 6 nm and 2 nm —uorescence maximum displacement for the former and the latter compounds, respectively.These small solvatochromic shifts together with the small decrease of the state energy with S1 increasing solvent polarity (vide infra) suggest that the lowest excited singlet states have limited charge transfer character for both HONI isomers. Fig. 2 shows the excitation and the —uorescence spectra of the hydroxy-naphthalimides and their conjugated bases in acetonitrile.In this solvent we used 1.5]10~4 M perchloric acid to prevent the dissociation of the OH-moiety and to record solely the spectra corresponding to the phenolic form. In order to deprotonate the OH group, 2 ll of 1 M KOH in methanol was added into 2 ml of hydroxy-naphthalimide solution. Under these conditions, the spectra are assigned to the naphtholate anion and the bands are located at lower energies. Based on the intersection of the normalized excitation and —uorescence spectra, 305 kJ mol~1 and 210 kJ mol~1 were calculated for the energies of the –rst excited singlet state of 3-HONI and its conjugated base, respectively.A much Fig. 2 Fluorescence and excitation spectra of (A) 3-HONI and (B) 4-HONI in acetonitrile. Phenolic form in the presence of 1.5]10~4 M perchloric acid (… … …) and deprotonated anion form in the presence of 1]10~3 M KOH (»»). 4760 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766smaller diÜerence was observed between the energies of S1 4-HONI and its deprotonated form (297 kJ mol~1 and 229 kJ mol~1, respectively).According to the Foé rster cycle,23 these results suggest that 4-HONI undergoes a smaller acidity enhancement upon electronic excitation. The lack of the long wavelength emission in clearly indicates that none of CH2Cl2 the singlet excited hydroxy-naphthalimides are sufficiently acidic to transfer proton to this solvent. Photophysical properties. Table 1 demonstrates that substitution of N-methyl-1,8-naphthalimide (NI) with a hydroxy moiety leads to a considerable change in —uorescence yield —uorescence lifetime and triplet yield Fluores- (UF), (qF) (UISC). cence yields and lifetimes increase more than one order of magnitude through the series of NI, 3-HONI, 4-HONI, whereas the variation of the triplet yields exhibits the opposite tendency; they decrease parallel with the energy of the lowest excited singlet In order to get a deeper insight into the (E(S1)).factors controlling the rate of the energy dissipating pathways, the rate constants for —uorescence intersystem crossing (kF), and internal conversion were derived using the (kISC) (kIC) expressions given below: kF\UF/qF (2) kISC\UISC/qF (3) kIC\(1[UISC[UF)/qF (4) It is seen from the rate constants presented in Table 1 that —uorescence emission is the dominant process from the singlet excited state of 4-HONI, however, radiationless transitions prevail for 3-HONI.In contrast with the very fast intersystem crossing observed for unsubstituted NI,16 triplet formation is fairly slow for both hydroxy-naphthalimides. This is especially true of the 4-HO derivative, where almost negligible was kISC found and the phosphorescence is extremely weak at 77 K in organic glass. In the case of 3-HONI the more efficient intersystem crossing permits the determination of the triplet»triplet absorption and the phosphorescence spectra.The triplet» triplet absorption maxima detected by a laser —ash photolysis technique appear at 450 and 480 nm at ambient temperature in whereas the phosphorescence peaks can be found CH2Cl2 , at 590 and 640 nm at 77 K in 95 : 5 butyronitrile : butyl acetate mixture. From the 0»0 transition of the phosphorescence the energy of the –rst triplet excited state is calculated to be 213 kJ mol~1. This value is lower than the one found for N-methyl-1,8-naphthalimide (221 kJ mol~1) indicating that substitution with an OH group not only decreases the energy of the state but also that of the state.S1 T1 The efficient triplet formation for N-methyl-1,8-naphthalimide was rationalized in terms of the transition (UISC\0.94) from the lowest singlet excited state to a close-lying higher triplet state As we established in a previous paper,16 (Tn).16 this is a thermally enhanced process in moderately and strongly polar solvents, which results in a temperature dependent —uorescent behavior.However, the —uorescence lifetimes of 3-HONI and 4-HONI were found to be temperature independent in the 296»198 K range. Based on these results, we can exclude the thermally activated intersystem crossing pathway for hydroxy-naphthalimides. The electron donating hydroxy group decreases the state energy and thereby, S1 increases the energy gap. With such an increased energy S1»Tn gap, thermal activation is not able to initiate the endothermic transition, consequently, triplet formation can occur S1 ]Tn only by the slower transitions to lower triplet levels.Similar eÜects were observed for 4-methoxy-N-methyl-1,8-naphthalimide, which also has an electron donating moiety.12 The signi–cant diÜerence in the triplet formation rate constant between 3-HONI and 4-HONI (Table 1) can be rationalized based on the relative position of the singlet and the triplet energy levels. Semi-empirical calculation24 showed that the energy gap is larger for 4-HONI compared with S1»T1 that of 3-HONI (144 kJ mol~1 and 132 kJ mol~1 were obtained, respectively).This larger energy gap may lessen the magnitude of the spin»orbit coupling between the and S1 T1 states leading to a slower intersystem crossing process for 4-HONI. It is evident from the data summarized in Table 1 that the introduction of a hydroxy group into the 1,8-naphthalimide moiety in position 3 hardly in—uences the rate constant for internal conversion however, the radiationless (kIC), S1 ]S0 transition is markedly decelerated for 4-HONI.It seems to be a general tendency that the 4-substitution of the 1,8-naphthalimide skeleton with an electron donating group leads to a reduced This eÜect can be observed for HO», and kIC. CH3 O» derivatives alike. Prado et al. suggested25 that the mol- NH2» ecule has a quinoid resonance form if an electron donating group is attached to position 4. This kind of electron displacement, which can occur both in the and the states, may S1 S0 aÜect the vibronic coupling between these states and consequently, may lead to slower internal conversion.II. EÜect of hydrogen-bonding additives In order to reveal how hydrogen-bonding in—uences the rate and the mechanism of the deactivation processes originating from the singlet excited state, we systematically varied the proton affinity, the hydrogen-bonding ability and the aromaticity of the additives using both hydrogen-bond donors and acceptors.Tri—uorethanol (TFE). We have previously shown that the —uorescence lifetime and quantum yield of N-methyl-1,8- naphthalimide considerably increase16 in TFE, a solvent which has high hydrogen-bonding power. In the present work, we extend these studies to hydroxy-naphthalimides, In contrast with that found for the unsubstituted N-methyl-1,8- naphthalimide, addition of 0.035 M TFE does not in—uence the —uorescent behavior of its hydroxy-derivatives in CH2Cl2 .A ten times higher amount of TFE caused a ca. 9 nm bathochromic shift of the —uorescence maxima for both the 3- and the 4-substituted derivatives, but neither —uorescence quenching nor appearance of a new emission band was observed. The excitation and ground state absorption spectra exhibited a red-shift as a function of the TFE concentration. The clear isosbestic points in the absorption spectra demonstrated that a 1 : 1 hydrogen-bonded complex formed with TFE. The equilibrium constants for hydrogen-bonding (K) were determined using the following relationship.26 [1[(A0/A)j]/[additive]\[K]K(eC/eA)j(A0/A)j (5) where is the ratio of the molar absorption coefficients (eC/eA)j for the complexed and free hydroxy-naphthalimides at a particular wavelength (j), and A denotes the absorbances in A0 Table 1 Photophysical properties of hydroxy-naphthalimides in CH2Cl2 jabs max/nm jFmax/nm E(SI)/kJ mol~1 UF qF/ns UISC kF/107 s~1 kIC/107 s~1 kISC/107 s~1 NIa 348 378 334 0.031 0.14 0.94b 22 21b 670b 3-HONI 376 411 308 0.13 1.9 0.50 6.8 19 26 4-HONI 362 440 300 0.85 9.0 0.03 9.4 1.3 0.3 a Ref. 16. b Based on triplet yield in acetonitrile. Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 4761the presence and the absence of TFE, respectively. Plotting the left-hand side of the function against gives good (A0/A)j linear correlation. From the intercepts, 1.5 M~1 and 1.1 M~1 hydrogen-bonding equilibrium constants were obtained for 3-HONI and 4-HONI, respectively. These values are comparable with that reported for —uorenone»TFE complex (0.7 M~1 in where TFE is also attached to a carbonyl CH2Cl2)4 moiety.Dimethylsulfoxide (DMSO). In contrast to the small eÜect of TFE hydrogen-bond acceptors, which interact with the HO-substituent, DMSO considerably alters the spectral behavior of hydroxy-naphthalimides. When weakly basic but strongly hydrogen-bonding DMSO is added to the hydroxynaphthalimide solutions, characteristic changes are observed, which are demonstrated for 3-HONI in Fig. 3. The red-shift of the long-wavelength absorption band and the clear isosbestic points at 337, 346.5, 380 nm for the 3-hydroxy and at 366.5 nm for the 4-hydroxy derivatives indicate 1 : 1 hydrogenbonded complex formation. From the eÜect of DMSO on the absorption spectra, the hydrogen-bonding equilibrium constants (K) were derived using eqn. (5). The average values of K calculated from the data measured at diÜerent wavelengths are 101^15 M~1 and 191^13 M~1 for the 3-HONI and 4-HONI complexes, respectively.The —uorescence maximum gradually shifts to lower energies with increasing DMSO concentration and an isoemissive point appears but the quantum yield of —uorescence does not change signi–cantly. As the red-shift is only ca. 11 nm, we assign the new —uorescence band to the hydrogen-bonded complex. This view is supported by the fact that DMSO does Fig. 3 Absorption and —uorescence spectra of 3-HONI in at CH2Cl2 diÜerent DMSO concentrations (0, 0.0014, 0.0035, 0.007, 0.014, 0.021, 0.035 M).not bring about any signi–cant change in the spectra of methoxy-naphthalimides, where no hydrogen-bonding is feasible. In addition, we can exclude proton transfer from the OHgroup to DMSO because no —uorescence was detected in the 500»800 nm range, where emission from the conjugate base (anion) of the hydroxy-naphthalimides is expected (Fig. 2). The excitation and the absorption spectra showed similar changes indicating that hydrogen-bonding with DMSO in the ground state is the dominant process. The variation of the —uorescence intensity (I) as the function of [DMSO] was analyzed using a relationship analogous to eqn.(5). [1[(I0/I)F]/[additive]\[K]K(eC/eA)j(UC/UA)F(I0/I)F (6) where represents the ratio of —uorescence intensities in (I0/I)F the presence and the absence of DMSO, and is the (UC/UA)F ratio of —uorescence efficiencies for the complex and the free —uorophore at the detection wavelength.26 Fitting this function to the experimental data leads to K values of 85^9 M~1 and 195^23 M~1 for the equilibrium constants for hydrogen-bonding of DMSO with 3-HONI and 4-HONI, respectively.The good agreement of these results with the corresponding values derived from absorption spectroscopic studies indicates that excitation of the complex does not cause signi–cant change in the hydrogen-bonding equilibrium constant, i.e., K is not signi–cantly diÜerent for the excited and the ground state complex.Time-resolved —uorescence measurements proved that the dynamic quenching by DMSO is negligible. The lifetimes of the singlet excited hydrogenbonded complexes of 3-HONI and 4-HONI are 2.2 ns and 9.3 ns, respectively. Since the lifetimes of the free and the complexed molecules agree very closely, we conclude that no energy dissipation takes place via the hydrogen-bond with DMSO. Pyridine. Addition of pyridine to the hydroxynaphthalimide solutions in leads not only to CH2Cl2 hydrogen-bonding in the ground state but also to —uorescence quenching.The change in the absorption spectrum closely resembles that found in the case of DMSO. Plotting the absorbances according to eqn. (5), we determined the equilibrium constant of hydrogen-bonding (K). The average values of K derived from measurements at various wavelengths are summarized in Table 2. The increase of the pyridine concentration in the 0»0.06 M range results in considerable —uorescence quenching but neither the appearance of a new band nor a shift of the maximum can be seen in the —uorescence spectra.Based on these observations, we conclude that the hydroxynaphthalimide »pyridine hydrogen-bonded complexes have negligible —uorescence yield. The Stern»Volmer plot of the Table 2 Hydrogen-bonding equilibrium constants and rate constants of —uorescence quenching Proton Aromaticity Ka Kb kq c Quencher af–nity/kJ mol~1 index, Ix absorption/M~1 —uorescence/M~1 lifetime/ 109 M~1 s~1 3-Hydroxy-naphthalimide» Imidazole 942.8 64.0 240 183 13 Pyridine 930.0 85.7 46 47 8.8 Pyrazole 894.1 73.0 21 17 9.3 Benzoxazol 891.6 38.0 d d 0.75 DMSO 884.4 » 101 85 e Isoxazole 848.6 47.0 d d e 4-Hydroxy-naphthalimide» Imidazole 942.8 64.0 900 615 9.6 Pyridine 930.0 85.7 138 137 8.9 Pyrazole 894.1 73.0 44 26 7.6 Benzoxazol 891.6 38.0 d d 1.8 DMSO 884.4 191 195 e Isoxazole 848.6 47.0 d d e a From absorption spectra.b From —uorescence spectra. c From —uorescence lifetime measurements. d No hydrogen-bonding can be detected. e No quenching. 4762 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766steady-state —uorescence intensities in the absence and the (I0) presence of pyridine (I) shows an upward curvature (Fig. 4). This suggests that the —uorophore can be quenched both in dynamic and static processes. If the static quenching is attributed entirely to ground state hydrogen-bonding, the modi–ed form of the Stern»Volmer equation describes the variation of vs.quencher concentration.27 I0/I I0/I\(1]K[quencher])(1]kq q0[quencher]) (7) where refers to the lifetime of singlet excited hydroxy- q0 naphthalimide, is the rate constant of the dynamic quen- kq ching and K denotes the equilibrium constant of complex formation in the ground state. The quenching rate constants were determined by time-resolved —uorescence technique (kq) (vide infra) and the K values were calculated by the nonlinear least-square –t of eqn.(7) to the experimental data. It is apparent in Fig. 4 that the calculated curves describe very well the experimental results. Table 2 demonstrates that the hydrogenbonding equilibrium constants derived from both absorption and —uorescence measurements closely agree. Addition of pyridine to the solutions of hydroxynaphthalimides shortens the lifetime of the lowest singlet excited state. The —uorescence decays are well described by a single exponential function.The —uorescence lifetimes in the absence and the presence (q) of pyridine are plotted in Fig. (q0) 4 based on the following equation: q0/q\1]kq q0[quencher] (8) Since the —uorescence decay times are in—uenced only by dynamic quenching, linear correlation is found between q0/q and the quencher concentration. The quenching rate constants derived from the slopes are given in the last column of (kq) Table 2. Benzoxazole and isoxazole.In order to reveal the major factors controlling the rate of the hydrogen-bonding induced —uorescence quenching, we extended our studies to heterocyclic compounds containing a –ve-membered ring. Isoxazole aÜects neither the —uorescence decay nor the spectral characteristics of hydroxy-naphthalimides indicating that no interaction occurs between these compounds either in the ground or the excited state. However, the eÜect of benzoxazole resembles that observed for pyridine.Time-resolved —uorescence measurements proved that dynamic quenching takes place but the reaction rate is much lower than that of pyridine. No clear indication was found for ground state hydrogen-bonding because of the overlap between the benzoxazole and the hydroxy-naphthalimide absorption. Pyrazole and imidazole. Compounds containing two heterocyclic nitrogens in a –ve-membered ring induce a diÜerent type of —uorescent behavior. They not only quench the —uorescence of hydroxy-naphthalimides but also cause a new —uorescence in the 500»800 nm spectral range whose intensity increases with the quencher concentration.It is apparent in Fig. 5 that the new emission consists of two bands for 3- HONI]imidazole and 3-HONI]N-methylimidazole solutions, whereas in the other cases no such clear evidence can be observed for dual luminescence in the long wavelength band. The short wavelength (SW) emission around 370»480 nm originates from the excited hydroxy-naphthalimides.The —uorescence lifetimes measured in this band decrease with increasing concentration of pyrazole and imidazole. This proves that dynamic quenching occurs. Fig. 6 gives the Stern»Volmer plots of the data obtained by steady-state and time-resolved —uorescence techniques for HONI»imidazole systems. In contrast with the linear dependence obtained from lifetime measurements, the Stern»Volmer plots of the steady-state —uorescence intensities are concave indicating that static quenching has an important contribution as well.The redshifts and the isosbestic points in the absorption spectra suggest 1 : 1 hydrogen-bonding. The experimental data were analyzed as described above for the other additives and the results are included in Table 2. The most interesting feature of the spectra in Fig. 5 is the appearance of the long-wavelength (LW) emissions, which are attributed to the singlet excited complexes of hydroxynaphthalimide with pyrazole and imidazole.The extent of proton transfer within these complexes is probably very sensitive to the acid»base properties of the constituents and the local polarity of the solvate shell. In the case of 3-HONI, the SW and the LW bands are wellseparated, therefore, we could readily see the formation and the decay of the species emitting in these spectral ranges. The variation of the —uorescence intensity as the function of time is presented in Fig. 7 for the solution containing 3-HONI and 0.024 M imidazole.The dotted line in Fig. 7A exhibits the —uorescence decays detected at 405 and 640 nm, whereas the continuous lines represent the –tted curves (vide infra), which were calculated by a non-linear least-squares deconvolution method. The parameter describing the rise of the signal at 640 nm (1.3 ns) perfectly agrees with the decay parameter obtained at 405 nm. The excitation pulse pro–le (dotted line) and the growing in of the —uorescence at 640 nm are shown in Fig. 7B using a better time resolution. The calculated curves match the measured data so well that they are hardly distinguishable in the –gure. The —uorescence decay at 405 nm can be well described with a single exponential function. (The signal shown in Fig. 7A has 1.3 ns lifetime.) The time-resolved —uorescence of the 3-HONI»imidazole complex exhibits more complex kinetics. The data were analyzed with a double exponential function : C2 exp([t/q2)[C1 exp([t/q1) (9) where t denotes time, and are constants. Calculation C1 C2 resulted in ns and ns for the decay constants q1\1.3 q2\4.9 when 0.024 M imidazole concentration was used.The C1/C2 ratio is expected to be 1 if the species emitting at long wavelengths is produced only in the quenching reaction.28 We found which clearly indicates that both the C2/C1\1.57, direct excitation of the ground state hydrogen-bonded complex and the dynamic quenching of the singlet excited 3-HONI result in LW emission.As it is expected for a pseudo- –rst-order process, the growing in of the LW —uorescence and the decay of the SW —uorescence strongly depend on the concentration of the quencher. However, the decay time of the LW emission only slightly decreases with the quencher concentration. It is evident from the spectra displayed in Fig. 5 that the excited hydrogen-bonded complexes of pyrazole have diÜerent characteristics compared with that of imidazole and Nmethylimidazole.In the former case, the LW emissions have a Gaussian shape with maxima at 600 nm and 520 nm for 3-HONI and 4-HONI, respectively. Since these maxima are at higher energies than the —uorescence peak of the deprotonated hydroxy-naphthalimides (Fig. 2), we suggest that only a partial proton shift takes place along the hydrogen-bond in these excited complexes. It is especially noteworthy that the addition of imidazole to the 3-HONI solution leads to structured LW —uorescence, which can be resolved to two components (Fig. 5A). In order to exclude the possibility that one of the LW —uorescence components originates from the interaction of 3-HONI with imidazole dimer29 we studied the reactions of Nmethylimidazole as well. This compound is not able to form a hydrogen-bonded dimer because it does not contain an N»H moiety. Fig. 5A demonstrates that the structured LW emission appearing in the presence of N-methylimidazole resembles that obtained with imidazole.Thus, we can rule out that Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 4763Fig. 4 Stern»Volmer plots of the results obtained by time-resolved and steady-state —uorescence technique for HONI»pyridine systems in (A) 3-HONI: steady-state measurements, time- CH2Cl2. >, Ö, resolved measurements. (B) 4-HONI: steady-state measurements, |, time-resolved measurements. L, dimerization of the additive causes the dual emission at long wavelengths. The relative intensity of the two —uorescence bands in the 500»800 nm spectral domain is temperature dependent both in ethyl acetate and in The substantial increase of CH2Cl2 .the higher energy component is particularly discernible in the 295»181 K temperature range in ethyl acetate where the two bands are better separated than in CH2Cl2 . It is readily seen in Fig. 8 that the intensity ratio of the two emissions in the 500»800 nm region strongly depends on the media. In acetonitrile, the band with a maximum around 550 nm disappears, and, likewise, addition of ethanol in the solution signi–cantly weakens this emission. Our CH2Cl2 results indicate that in solvents of medium polarity the 3- HONI»imidazole excited complexes have two dominant struc- Fig. 5 Fluorescence spectra in (A) 3-HONI in the presence CH2Cl2 . of 0.156 M pyrazole (heavy line), 0.024 M imidazole (dotted line) and 0.024 M N-methylimidazole (thin line) ; (B) 4-HONI in the presence of 0.057 M pyrazole (heavy line), 0.018 M imidazole (dotted line) and 0.018 M N-methylimidazole (thin line).Fig. 6 Stern»Volmer plots of the data obtained by steady-state and time-resolved —uorescence techniques for the HONI»imidazole systems in (A) 3-HONI: steady-state measurements; CH2Cl2. >, Ö, time-resolved measurements. (B) 4-HONI: steady-state measure- |, ments; time-resolved measurements. L, tures which diÜer in the extent of their proton shift. However, the complex that —uoresces at higher energies is bound with a hydrogen-bond and possesses only a limited proton transfer character when the proton is removed from the HO group toward the heterocyclic nitrogen in the species emitting at long wavelengths.The identical —uorescence decay times throughout the LW bands of the 3-HONI»imidazole complex in suggest that a fast equilibrium is established CH2Cl2 between the two types of complex. The polar solvents weaken the hydrogen-bond and promote proton transfer, therefore, no dual emission can be seen in acetonitrile.The ion-pair character of the complex in acetonitrile is supported by the fact that the —uorescence maximum of both the 3-HONI»imidazole complex (Fig. 8) and the 3-HONI anion (Fig. 2A) are located around 630 nm in this polar solvent. Fig. 7 Fluorescence decays in 3-HONI]0.024 M imidazole solution in (A) Fluorescence decay (dotted line) and –tted curve CH2Cl2 . (continuous line) at 405 nm and 640 nm. (B) Excitation pulse pro–le (dotted line), —uorescence growing in and –tted curve (continuous line) at 640 nm. 4764 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766Fig. 8 Fluorescence spectra of 3-HONI in the presence of 0.024 M imidazole: (A) in (B) in M EtOH, (C) in the CH2Cl2, CH2 Cl2]0.17 presence of 0.052 M imidazole in acetonitrile and (D) in ethyl acetate. Comparison of the eÜects of the various hydrogen-bond acceptors. It was demonstrated that the intermolecular hydrogen-bonding with alcohols in the singlet excited state acts as an eÜective accepting mode of radiationless deactivation for aromatic carbonyl compounds.2h6 We should comment on the question of why hydroxy-naphthalimides, which also contain carbonyl groups, are not quenched by the strong hydrogen-bond donor TFE. For the example of 2- substituted —uorenones we demonstrated that efficient hydrogen-bonding induced internal conversion can occur only if the carbonyl oxygen has high electron density in the excited state.30 As stated above, the absorption and the —uorescence spectra of hydroxy-naphthalimides exhibit small solvatochromic shifts because excitation leads to minor change in the dipole moment. Theoretical calculations also corroborated that there is no signi–cant diÜerence in the electron density of the carbonyl oxygen for the and the states of naphthali- S0 S1 mides,12,31 therefore, hydrogen-bonding with TFE does not accelerate the internal conversion process.Table 2 lists the hydrogen-bonding equilibrium constants and the quenching rate constants obtained by diÜerent methods.The proton affinities32 and the Bird aromaticity indices33 for the quenchers are also included. Katritzky et al. showed that the Bird aromaticity index is the best measure of the classical aromaticity,34 therefore, we chose this among the various aromaticity parameters available in the literature. The rate constant of the excited hydroxy-naphthalimide quenching varies remarkably with the molecular structure of the hydrogen-bond acceptor ; no quenching takes place with isoxazole but the reaction is diÜusion controlled in the case of imidazole.A parallel change can be seen between the proton affinity and the rate constants throughout the series of the kq quenchers shown in Table 2. The reactants that have low proton affinity do not promote the deactivation of the singlet excited hydroxy-naphthalimides. The hydrogen-bonding power of the additives, as measured directly by the ground state hydrogen-bonding equilibrium constants (K), does not play a rate determining role because no correlation can be found between the K and quantities.For example, DMSO kq does not quench the —uorescence of hydroxy-naphthalimides in spite of the fairly large hydrogen-bonding equilibrium constant in the ground state. These results suggest that proton displacement plays a crucial role in the interaction of excited hydroxy-naphthalimides and hydrogen-bond acceptors.It is not surprising that no correlation appears between the ground state hydrogen-bonding equilibrium constants and the proton affinities listed in Table 2. Gurka and Taft established that hydrogen-bonding and basicity are unrelated.35 For example, using a common hydrogen-bond donor, they showed that the of a carbonyl compound is 13 powers of ten less pKa than that of the corresponding amine for equal values of hydrogen-bonding equilibrium constant.36 Moreover, the data reported by Abraham et al. demonstrate that DMSO forms stronger hydrogen-bonds than the much more basic pyridine derivatives.37 Coupled electron»proton movement was suggested to promote the radiationless deactivation when a heterocyclic molecule containing an aromatic p-electronic system is connected to excited hydroxyarenes directly via a hydrogenbond. 10 The extremely weak —uorescence for the hydrogen-bonded complexes of benzoxazole and pyridine with hydroxy-naphthalimides is probably due to a rapid internal conversion via a similar process. The proton shift toward the hydrogen-bond acceptor induces efficient nonradiative energy dissipation. However, the intensive emission as well as the long lifetime (ca. 4»9 ns) of the excited hydrogen-bonded complexes containing imidazole and pyrazole obviously indicate slow internal conversion value of ca. 108 s~1 can be (kIC deduced from the experimental data). We did not –nd a correlation between the aromaticity index of the hydrogen-bond acceptor and the radiationless deactivation rate of the excited hydrogen-bonded complex. This seems to indicate that the energy dissipation mechanism suggested for the excited hydroxyarene»pyridine species does not play a dominant role if other types of nitrogen heterocyclics serve as the hydrogenbond acceptor. Acknowledgements very much appreciate the support of this work by the We Hungarian Science Foundation (OTKA, Grant T 023428) and the scienti–c exchange program between the French Ministry of Foreign AÜairs and the Hungarian Committee for Technological Development (Balaton Project F-10/97). References 1 J. Herbich, C.-Y. Hung, R. P. Thummel and J. Waluk, J. Am. Chem. Soc., 1996, 118, 3508 and references therein. 2 H. Inoue, M. Hida, N. Nakashima and K. Yoshihara, J. Phys. Chem., 1982, 86, 3184. 3 R. S. Moog, N. A. Burozski, M. M. Desai, W. R. Good, C. D. Silvers, P. A. Thompson and J. D. Simon, J. Phys. Chem., 1991, 95, 8466; J. Ritter, H. U. Borst, T. Lindner, M. Hauser, S. Brosig, K. Bredereck, U. E. Steiner, D. Kué hn, J. Kelemen and H. E. A. Kramer, J. Photochem. Photobiol. A, 1988, 41, 227; H. U. Borst, J. Kelemen, J. Fabian, M. Nepras and H. E. A. Kramer, J. Photochem. Photobiol. A, 1992, 69, 97. 4 L. Biczoç k, T. Beç rces and H. Linschitz, J. Am. Chem. Soc., 1997, 119, 11071. 5 T. Yatsuhashi and H. Inoue, J. Phys. Chem. A, 1997, 101, 8166. 6 T. Yatsuhashi, Y. Nakajima, T. Shimada, H. Tachibana and H. Inoue, J. Phys. Chem. A, 1998, 102, 8657. 7 C. Turroç , C. K. Chang, G. E. Leroi, R. I. Cukier and D. G. Nocera, J. Am. Chem. Soc., 1992, 114, 4013. 8 R. I. Cukier and D. G. Nocera, Annu. Rev. Phys. Chem., 1998, 49, 337. 9 W. J. Liegh, E. C. Lathioor and M. J. St Pierre, J. Am. Chem. Soc., 1996, 118, 12339. 10 N. Mataga and H. Miyasaka, Prog. React. Kinet., 1994, 19, 317 and references therein. 11 H. Miyasaka, K. Wada, S. Ojima and N. Mataga, Isr. J. Chem., 1993, 33, 183. 12 V. Wintgens, P. Valat, J. Kossanyi, A. Demeter, L. Biczoç k and T. Beç rces, New J. Chem., 1996, 20, 1149. 13 L. M. Tolbert and J. E. Haubrich, J. Am. Chem. Soc., 1990, 112, 8163. 14 L. M. Tolbert and J. E. Haubrich, J. Am. Chem. Soc., 1994, 116, 10593. 15 D. Huppert, L. M. Tolbert and S. Linares-Samaniego, J. Phys. Chem. A, 1997, 101, 4602. 16 V. Wintgens, P. Valat, J. Kossanyi, L. Biczoç k, A. Demeter and T. Beç rces, J. Chem. Soc., Faraday T rans., 1994, 90, 411. 17 W. Adam, X. Quian and C. R. Saha-Moé ller, T etrahedron, 1993, 49, 417. Phys. Chem. Chem. Phys., 1999, 1, 4759»4766 476518 R. Royer, J. P. Buisson, P. Demerseman and J. P. Lechartier, Bull. Soc. Chim. Fr., 1969, 2792. 19 W. M. Rodionow and A. M. Fedorowa, Bull. Soc. Chim. Fr., 1939, 479. 20 P. Valat, V. Wintgens, J. Kossanyi, L. Biczoç k, A. Demeter and T. Beç rces, J. Am. Chem. Soc., 1992, 114, 947. 21 S. L. Murov, G. L. Carmichael and I. Hug, Handbook of Photochemistry, Marcel Dekker, New York, 2nd edn., 1993. 22 J. K. Hurley, N. Sinai and H. Linschitz, Photochem. Photobiol., 1983, 38, 9. 23 T. Foé rster, Z. Electrochem., 1950, 54, 531. 24 The HyperChem program package was used. The geometry of the molecule was optimized by AM1 calculation and the energy levels were obtained by ZINDO/S method. 25 A. Prado, J. Campanario, J. M. L. Poyato, J. J. Camacho, D. Reyman, E. Martin, T HEOCHEM., 1988, 166, 463. 26 N. Mataga and S. Tsuno, Bull. Chem. Soc. Jpn., 1957, 30, 368. 27 J. R. Lakowicz, Principles of —uorescence spectroscopy, Plenum Press, New York, 1983, p. 266. 28 D. V. OœConnor, L. Chewter and D. Phillips, J. Phys. Chem., 1982, 86, 3400. 29 E. Fischer, Ber. Bunsen-Ges. Phys. Chem., 1969, 73, 1007. 30 L. Biczoç k, T. Beç rces and H. Inoue, J. Phys. Chem. A., 1999, 103, 3837. 31 M. Adachi, Y. Murata and S. Nakamura, J. Phys. Chem., 1995, 99, 14240. 32 E. P. L. Hunter and S. G. Lias, J. Phys. Chem., Ref. Data, 1998, 27, 707. 33 C. W. Bird, T etrahedron, 1985, 41, 1409; ibid., T etrahedron, 1986, 42, 89. 34 A. R. Katritzky, M. Karelson and N. Malhotra, Heterocycles, 1991, 32, 127. 35 D. Gurka and R. W. Taft, J. Am. Chem. Soc., 1969, 91, 4794. 36 R. W. Taft, D. Gurka, L. Joris, P. von R. Schleyer and J. W. Rakshys, J. Am. Chem. Soc., 1969, 91, 4801. 37 M. H. Abraham, P. P. Duce, D. V. Prior, D. G. Barrett, J. J. Morris and P. J. Taylor, J. Chem. Soc., Perkin T rans. 2, 1989. 1355. Paper 9/04520A 4766 Phys. Chem. Chem. Phys., 1999, 1, 4759»4766