FRET versus PET : ratiometric chemosensors assembled from naphthalimide dyes and crown ethers

Novel bi-chromophoric naphthalimide derivatives containing benzo-15-crown-5 and Nphenyl-aza-15-crown-5 receptor moieties BNI2 and BNI3 were designed and prepared. Significant Förster resonance energy transfer (FRET) from donor (D) amido-naphthalimide to acceptor (A) amino-naphthalimide chromophore as well as photoinduced electron transfer (PET) between the N-aryl receptor and amido-naphthalimide fragment were revealed by the steadystate and time resolved UV/Vis absorption and fluorescence spectroscopy. Upon the addition of alkaline-earth metal perchlorates to an acetonitrile solution of ligands, FRET mediated fluorescence enhancement was observed, which was a result of inhibition of PET competitive deactivation pathway. The studied compounds provide an opportunity to register two-channel fluorescence response using selective excitation of either of the photoactive units and, thus might be of interest as ratiometric probes.


Introduction
Förster resonance energy transfer is an unique process making possible to generate fluorescence signals sensitive to molecular conformation, association and separation in the 1-10 nm range. 1,2This mechanism has been widely used in medicinal diagnostics, optical imaging and molecular biology as a spectroscopic ruler to study structure of proteins and nucleic acids.In recent years, significant emphasis has been placed on the development of highly selective fluorescent FRET based chemosensors for metal cations because of their potential applications in biochemistry and environmental research.Among various photoinduced processes that are commonly involved in the signaling or response phenomena, the resonance energy transfer seems to be an optimal strategy for designing ratiometric probes. 3,4cording to ratiometric method, analyte concentration can be quantified by using the ratio of intensities of the well resolved fluorescence peaks with reasonable intensities at two different wavelengths for analyte free and analyte bound probe. 5Such self-calibration using two emission bands can eliminate the influence of indicator dye concentration, environmental conditions and instrumental efficiency.Furthermore, the pseudo-Stokes shifts of FRET based probes are larger than the Stokes shifts of either the donor or acceptor dyes; thus, the possible self-quenching as well as fluorescence detection errors due to back scattering effects from the excitation source will be efficiently avoided. 6phthalimide derivatives are a special class of environmentally sensitive fluorophores.
The fluorescence of 1,8-naphthalimides with electron donating groups at C-4 position of naphthalene ring has been of great interest for several decades in connection with an array of technical, medical and electronic use.Because of its intense fluorescence and good photostability, this type of compounds has found application in a number of areas including coloration of polymers, 7,8 laser active media, 9,10 fluorescent markers in biology, [11][12][13] anticancer agents and analgesics in medicine, 14 electroluminescent materials, [15][16][17] fluorescence switchers, 18-20 liquid crystal displays 21,22 and ion probes. 23,24 date some examples of naphthalimide based FRET probes have been reported in literature.6][27] A FRET-based ratiometric chemosensor for in vitro cellular fluorescence analyses of pH based on naphthalimide-coumarin system was reported by Zhou et al. 28 Selective ratiometric chemosensors for Cu 2+ and Zn 2+ were obtained using dansylamidenaphthalimide conjugates with variable polymethylene linker length between the chromophores. 29,30[33][34][35][36][37][38][39][40] Scheme 1. Structure of compounds MNI1-7 and BNI1-3 2][43] These compounds displayed pronounced enhancement of emission intensity by coordination with metal cations, which was a result of inhibition of PET between crown ether receptor conjugated with N-phenyl ring and fluorophore.Herein, we report the design, synthesis and investigation of cation-dependent behavior of FRET-based ratiometric sensors BNI2-3 by integrating amido-naphthalimide probes MNI5 and MNI7 as FRET donors and amino-naphthalimide MN1 as an FRET acceptor.In this case, the strategy for detection of metal ions is based on modulating of FRET process, and thus emission intensity of acceptor amino-naphthalimide fragment, by means of incorporation of competitive PET deactivation pathway.In order to receive more complete comparative picture for the influence of crown ether groups on the efficiency of FRET interaction in a bi-chromophoric system we prepared nonionophoric dyad compound BNI1.Naphthalimides MNI1, 44 MNI2, 44 MNI3, 44 MNI5 44 and MNI7 43 have been synthesized earlier and were also included in photophysical studies as reference compounds.

Experimental Section
Steady-state optical measurements.The absorption spectra were taken on a Varian-Cary 5G spectrophotometer.The fluorescence quantum yield measurements were performed using a Varian-Cary 5G spectrophotometer and a FluoroMax-3 spectrofluorimeter.Spectral measurements were carried out in air-saturated acetonitrile solutions (acetonitrile of spectrophotometric grade, water content <0.005%, Aldrich) at 20 ± 1 ºC; the concentrations of studied compounds were of about 0.5-2.0•10 - M. All measured fluorescence spectra were corrected for the nonuniformity of detector spectral sensitivity.Coumarin 481 in acetonitrile (ϕ fl = 0.08) 45 was used as reference for the fluorescence quantum yield measurements.The fluorescence quantum yields were calculated by the Eq. ( 1), 46 where I(t) is the fluorescence intensity, Gauss is the Gaussian profile of the excitation pulse, in which t 0 is the excitation pulse arrival delay, ∆t -the excitation pulse width, and A -the amplitude.The parameter τ is the lifetime of the excited state.The initial condition for the equation is I(-∞) = 0. Typically, the fit shows a χ 2 value (Pirson's criteria) better than 10 -4 and a correlation coefficient R > 0.999.The uncertainty of the lifetime was better than 1%.Routinely, the fluorescence accumulation time in our measurements did not exceed 90 s.
Transient Absorption Setup.The laser system and frequency-conversion apparatus employed to excite samples were the same as for time-resolved fluorescence measurements.
White light continuum (360-1000 nm) pulses generated in a 5 mm methanol cell were used as a probe.The variable delay time between excitation and probe pulses was obtained by using a delay line with 0.1 mm resolution.The solutions were placed in a 1 mm circulating cell.
Whitelight signal and reference spectra were recorded with a two-channel fiber spectrometer (Avantes Avaspec-2048-2).A home-written acquisition and experiment-control program in LabView made it possible to record transient spectra with an average error of less than 10 -4 times the optical density for all wavelengths.The temporal resolution of our setup was better than 60 fs.Temporal chirp of the probe pulse was corrected by a computer program with respect to a Lawrencian fit of a Kerr signal generated in a 0.2 mm glass plate used in a place of the sample.

Equilibrium constant determination. Complex formation of compounds BNI2 and
BNI3 with Mg 2+ and Ca 2+ in acetonitrile at 20 ± 1 °C was studied by spectrofluorometric titration. 47,48The ratio of dye to M 2+ was varied by adding aliquots of a solution of metal perchlorate * of known concentration to a solution of ligand BNI2 or BNI3 of known concentration.The fluorescence spectrum of each solution was recorded, and the stability constants of the complexes were determined using the SPECFIT/32 program (Spectrum Software Associates, West Marlborough, MA).The following equilibria were considered in the fitting (Eq.(3) and Eq. ( 4), L = BNI2 or BNI3; M 2+ = Mg 2+ or Ca 2+ ): In doing so, it was found that the experimental data corresponded to the theoretical ones if only the Eq. ( 3) was taken into account and the formation of the complexes with composition of 2 : 1 was not observed.
The equilibrium constants for protonation of ligand BNI3 was not determined by this method because of high stability (K > 10 7 M -1 ) of protonated form (BNI3)•H + .Anhydrous magnesium perchlorate was used as received.

Determination
(BNI3)•H + was done in the presence of 2 eq.HClO 4 in ligand solutions, which can be understood from the fact that the further addition of HClO 4 did not result in the fluorescence enhancement and complex formation had already been finished.

Design and synthesis of the compounds
Following the classical description of FRET model, a requirement for efficient energy transfer is that there should be a spectral overlap between the emission of the donor and absorbance of the acceptor dyes.It is well known that absorption and fluorescence characteristics of the 1,8-naphthalimides depend on the nature of the substituent at C-4 position of the 1,8naphthalimide ring involved in the charge transfer interaction with dicarboximide moiety.In the construction of dyad probes BNI2-3, amido-naphthalimide was chosen as an energy donor, because it has strong emission in the visible range centered at 440-460 nm, which covers a part of amino-naphthalimide's absorption (λ max = 410-430 nm). 44Fig. 1 shows the overlap between the absorption and fluorescence spectra of reference compounds MNI1 and MNI2 in acetonitrile, fulfilling a favorable condition for FRET.Another factor, which influences the FRET efficiency is the space separation of donor (D) and acceptor (A) units.Since the transfer rate drops rapidly with the increase of D-A distance, we used rather short and rigid phenyl spacer in BNI2-3.Furthermore, low conformational flexibility of phenyl group would also hinder a dyad molecule from adopting a conformation where both naphthalimide moieties are arranged as an internal aggregate stabilized by π-stacking interaction, in which the formation of non-emissive states could be suggested.
Crown ether groups were incorporated in the N-aryl fragment of more electron deficient amido-

Physical Chemistry Chemical Physics Accepted Manuscript
naphthalimide chromophore (in comparison with amino-naphthalimide), because in this case strong PET interaction is expected for both benzo-15-crown-5 and N-phenylaza-15-crown-5 ether receptors. 41,43heme 2. Synthetic route to compounds BNI1-3 The synthesis of bi-chromophoric naphthalimide derivatives BNI1-3 was carried out were subjected to reduction using tin (II) chloride in the presence of hydrochloric acid.The experimental details concerning the synthesis of target compounds can be found in Supplementary Information.

Photophysical properties of the compounds
Photophysical characteristics of BNI1-3 were measured in acetonitrile solution and the data are presented in Table 1.First of all, we studied the resonance energy transfer characteristics of non-crowned derivative BNI1.Absorption spectrum of BNI1 (Fig. 2a), as expected, showed the presence of two long wavelength bands corresponding to the absorption location of amidonaphthalimide donor (reference compound MNI2) and amino-naphthalimide acceptor (reference compound MNI1).Given a high level of additivity of BNI1 spectrum, there could be supposed a lack of any overlap of molecular orbitals between the individual fluorophores in the BNI1 ground state.

Physical Chemistry Chemical Physics Accepted Manuscript
Selective excitation of BNI1 using 340 nm light produced single band emission at 520 nm (Fig. 2b), which is characteristic of amino-naphthalimide fragment.† In contrast, under the same conditions, the equimolar mixture of fluorophores MNI1 and MNI2 demonstrated the emission at around 450 nm originating from the amido-naphthalimide MNI2.This result indicates that in bi-chromophoric system the excitation energy transfers effectively from donor to acceptor unit, whereas in the case of dilute solution of equimolar mixture FRET interaction is not observed.To get deeper insight into the nature of excited state deactivation pathways we measured the excited state lifetime of donor chromophore in the compound BNI1.In comparison with the † A weak shoulder at 450 nm in the fluorescence spectrum of BNI1 results from residual fluorescence of donor chromophore.The appearance of this shoulder can be explained by the relatively high fluorescence quantum yield of MNI2 (Table 1).

Physical Chemistry Chemical Physics Accepted Manuscript
single amidonaphthalimide MNI2 (τ = 10 ns), it was shorter by more than four orders of magnitude (τ D = 0.31 ps), ‡ implying the existence of a fast non-radiative process more likely to be the resonance energy transfer.The efficiency of the energy transfer (Φ ୖ ) in dyad compound BNI1 was calculated equal to 0.99997 (99.997 %) according to Eq. ( 5). 2 The pretty close value of Φ ୖ (99.95%) was obtained from calculations using Förster theory (for details see Supplementary Information).Such high value of Φ ୖ could be a result of a rather short distance between donor and acceptor chromophores (r = 12.0 Å as obtained from the optimized geometry of BNI1 (Fig. S1)), which is about 3.5 times shorter compared to critical Förster radius (R 0 = 41.8Å) for this system.The introduction of crown ether substituents in the naphthalimide dyad molecule BNI1 results in the reduction of the energy transfer efficiency.As it was shown in our previuos publications, the presence of electron releasing benzo-15-crown-5-, benzo-1,4-dioxane or Nphenyl-aza-15-crown-5 ether groups in the N-aryl fragment of amido-naphthalimides MNI3, MNI5 and MNI7 leads to dramatic decrease of emission intensity with respect to highly emissive non-crowned derivative MNI2 due to efficient photoinduced electron transfer between the naphthalimide chromophore and receptor moieties. 41,43,44Keeping this in mind, one could ‡ The data for MNI2 are obtained from the analysis of fluorescence kinetics, for compound BNI1 -from the analysis of transient absorption spectra (see Experimental Section).

Physical Chemistry Chemical Physics Accepted Manuscript
conclude that in the case of crown-containing dyad compounds BNI2 and BNI3, the deactivation of the donor chromophore excited state would proceed via both electron and energy transfer.
Additionally, radiative decay (fluorescence) and other possible non-radiative ways of relaxation (except PET and FRET) should be taken into consideration.The representative scheme showing all these energy degradation channels in dyads BNI2 and BNI3 is depicted in Fig. 3. Provided each photophysical process in Fig. 3 is characterized by the first order rate constant, the Φ ୖ value can be expressed as the ratio of FRET rate constant (k FRET ) to the sum of rate constants of all other processes mentioned above (Eq.( 6)).
In the Eq. ( 6), k R stands for the radiative rate constant of amido-naphthalimide chromophore, and k NR describes its non-radiative relaxation which is not related to energy or electron transfer.To estimate the sum (k R + k NR ) we used the value inversely proportional to the fluorescence lifetime of compound MNI2, where FRET and PET channels are not realized (Eq.( 7)).k FRET was calculated as difference of deactivation rate constants for compounds BNI1 and MNI2, supposing that the decrease of excited state lifetime on going from MNI2 to BNI1 is only a result of FRET interaction (Eq.( 8)).

Physical Chemistry Chemical Physics Accepted Manuscript
Simple analysis of Eq. ( 6) clearly shows that the FRET efficiency in a bi-chromophoric system can be modulated by changing the rate of phoinduced electron transfer.Thus, the increase of PET donor ability would reduce the amount of energy transferred to aminonapthalimide acceptor and thereby quench fluorescent output signal.For the evaluation of k PET in dyads BNI2 and BNI3, we used relaxation kinetics data of amido-naphthalimides MNI3 and MNI7 in which PET is the main deactivation pathway.As an example, Fig. 4a shows transient absorption spectra of MNI3 at different time delay between pump and probe pulses.It can be seen that the relaxation of singlet excited state (S 1 ) proceeds with a concomitant growth of two novel bands probably corresponding to ion-radical intermediates.An intense signal with maximum at 426 nm was assigned by us to benzodioxane cation-radical absorption as it drops into wavelength interval 400 -480 nm where the characteristic bands of either isomeric dimethoxybenzenes 50 or N,N-dimethylaniline cation-radicals 51 are located.From the analysis of kinetic data (Fig. 4b), PET rate constant (k PET ) for the compound MNI3 was found to be as high As the comparison of k PET and k FRET shows, PET and FRET are comparably competing processes.Using the data of time-resolved experiments and Eq. ( 6), we found that introduction of benzo-15-crown-5 ether receptor into the N-phenyl ring of BNI1 decreases energy transfer efficiency from 99.997% to 64%, whereas the presence of aza-15-crown-5 ether group possessing more strong PET donor ability in the compound BNI3 results in only 34% of excitation energy involved in FRET.The changes of Φ ୖ value in the range of naphthalimide dyads BNI1-BNI2-BNI3 was found to be in full agreement with the steady-state optical data showing the reduction of the fluorescence intensity and fluorescence quantum yield for the crown-containing compounds (Fig. 5, Table 1).

Page 11 of 17
Physical Chemistry Chemical Physics

Complex formation of naphthalimide dyads
We further examined the ability of crown-containing dyads BNI2 and BNI3 to switch their photophysical characteristics as a result of metal ion binding.For the complexation experiments we chose Mg 2+ and Ca 2+ , because these cations are known to form stable complexes with benzo-15-crown-5 and aza-15-crown-5 ethers in an acetonitrile solution, 41,43 which is very convenient when studying cation-induced optical effects.
The addition of magnesium and calcium perchlorates to the solution of ligands BNI2 and BNI3 in MeCN doesn't virtually change the position and intensity of the long wavelength absorption bands (compare abs max λ and ε λ values for the free ligands and corresponding complexes in Table 1).This observation indicates that charge transfer transitions in both naphthalimide units in molecules BNI2 and BNI3 are not affected by the coordination of metal ions with the crown ether receptors.The similar results were obtained for the monochromophoric naphthalimides MNI4-7 in our previous publications, 41,43 where the negligible changes in the absorption spectra were attributed to the lack of conjugation between the N-aryl group and naphthalimide chromophore resulted from the nearly orthogonal disposition of these fragments in space.Apparently, the same structural feature persists in dyads BNI2 and BNI3 explaining the similarity in spectral behavior.
In contrast to the absorption spectra, the addition of Mg 2+ and Ca 2+ led to the pronounced changes in the emission intensity.Considering the ability of PET process to be fully or partially blocked upon the complex formation, one should expect that the binding of metal cations by the BNI2 and BNI3 molecules would cause an increase in energy transfer efficiency giving rise to fluorescence enhancement of the acceptor amino-naphthalimide chromophore.As depicted in Fig. 6, the described situation was observed in the experiment.The fluorescence emission spectra of compounds BNI2 and BNI3 were recorded in the presence of graduated amounts of corresponding metal perchlorates and the titration data were applied for the calculation of stability constants (Table 1).
From the analysis of cation-induced changes in the emission spectra, it can be seen that optical responses for the compounds BNI3 and BNI2 are different and indeed, BNI3

Physical Chemistry Chemical Physics Accepted Manuscript
Another factor approving the residual PET in (BNI3)•Ca 2+ is that the emission spectrum of the Ca 2+ -saturated solution of BNI3 doesn't contain the short wavelength shoulder arose from the fluorescence of donor chromophore which, in contrast, emerges in the case of BNI1 (Fig. 2b) and (BNI2)•Mg 2+ (Fig. 6a).In order to confirm our assumption concerning partial inhibition of PET in the (BNI3)•Ca 2+ , we measured spectral characteristics of protonated form of ligand BNI3 (Table 1).
The addition of HClO 4 in acetonitrile solutions of BNI3 resulted in formation of highly fluorescent complexes (BNI3)•H + similar in φ fl value to BNI1.In this complex the lone electron pair of receptor's anilino nitrogen is fully engaged in coordination with the cation due to formation of N-H σ-bond, which breaks the conjugation, significantly lowers the potential energy of N-aryl group and thus completely quenches PET process.
Finally, we would like to clear how the studied bi-chromophoric compounds can be applied for the ratiometric detection of ions.The main idea of ratiometric measurements is based on self-calibration of a sensor system.This means that probe's response should contain not only the signal reporting on the analyte binding.There must be a different signal (or signals) that could allow to account or compensate all the effects that influence the fluorescence parameter(s) besides the analyte-probe interaction.In the case of crown-containing dyads BNI2 and BNI3, the excitation of aminonaphthalimide chromophore using the visible light (λ ex = 440 nm) leads to emission at 520 nm.Noteworthy, this signal does not depend on whether a cation is present in the crown ether moiety or not and hence, can be used for self-calibration.In contrast, the fluorescence output obtained as a result of donor chromophore excitation with the UV light (λ ex = 340 nm) is strongly cation-dependent.Thus, the ratio of the yellow-green emission intensities at 520 nm obtained at λ ex = 340 nm and λ ex = 440 nm (I 340 /I 440 ) was found to increase with the

Physical Chemistry Chemical Physics Accepted Manuscript
amount of Mg 2+ or Ca 2+ added in the solution for both compounds BNI2 and BNI3 (Fig. 6, the inserts).This provides an opportunity to calculate the concentration of a cation ([M n+ ]) without necessity to know exactly the concentration of a sensor according to Eq. ( 9), 5 where R is the ratio I 340 /I 440 for the studied sample with an unknown M n+ content, R max and R min correspond to the ratios I 340 /I 440 measured for the analyte bound and analyte free probe respectively, and K D is the complex dissociation constant.For instance, the ratio of emission intensities R for the 6.5 µM solution of BNI3 containing 2 equivalents of Ca 2+ is 0.54.Assuming that R min and R max values are as high as 0.15 and 0.92 (found from the titration curve (Fig. 6b)) and K D = 1 / K = 1 / 10 5.04 M (lg K = 5.04 for (BNI3)•Ca 2+ , Table 1), the equilibrium concentration of calcium cations in the solution is [Ca 2+ ] = 9.4•10 -6 M.This value is very close to the one (9.7•10 - M) obtained from the calculations of the solution composition using SPECFIT/32 program.

Conclusion
Novel crown-containing naphthalimide dyads BNI2 and BNI3 were synthesized based on convergent approach.The compounds were designed as ratiometric cation FRET chemosensors comprising PET amido-naphthalimide fluorophore linked with the aminonaphthalimide fragment.Steady-state and time-resolved optical studies revealed that the resonance energy transfer operating between the photoactive units competes fairly well with the photoinduced eletron transfer from the crown ether receptor and, thus can be switched on by the presence of metal ions.As a result, a fluorescence enhancement of the acceptor aminonaphthalimide chromophore occurs upon the complex formation with Ca 2+ and Mg 2+ , the more pronounced effect being observed in the case of aza-15-crown-5-containing derivative BNI3 where PET interaction was supposed to be stronger compared to BNI2.
In contrast to conventional "off-on" or "on-off" PET sensors with one fluorophore, our dyads open a way for ratiometric fluorescent detection of ions, because the FRET mediated "offon" signal output obtained at λ ex = 340 nm can be self-calibrated with respect to emission channel at λ ex = 440 nm entirely unresponsive to the presence of analyte.Thus, the presented results have shown that compounds BNI2 and BNI3 can be of interest for the development of fluorescent ratiometric chemosensors for various kinds of cationic analysis.Due to versatile structural modification on the crown ether moiety, selective probes for various metal cations can be prepared.Related research is currently underway in our laboratory.
fluorescence quantum yields of the studied solution and the standard compound, respectively; A i and A 0 are the absorptions of the studied solution and the standard, respectively; S i and S 0 are the areas underneath the curves of the fluorescence spectra of the studied solution and the standard, respectively; and n i and n 0 are the refraction indices of the solvents for the substance under study and the standard compound (n i = n 0 = 1.342, acetonitrile).
of fluorescence quantum yields of complexes.The fluorescence quantum yields of complexes (BNI2)•Mg 2+ and (BNI3)•Ca 2+ were determined using solutions of ligands BNI2-3 in CH 3 CN containing an excess of the corresponding metal perchlorate in order to obtain 90-95% of ligand bound with the cation.The required M(ClO 4 ) 2 excess was calculated from the known stability constants using the SPECFIT/32 program.The measurement of ϕ fl for * Calcium perchlorate for complexation studies was dried in vacuum (7-8 mm Hg) at 240 °C and kept anhydrous over P 2 O 5 in desiccator (Caution!Calcium perchlorate may explode when heating.It decomposes at 270 °C49 ).

Fig. 3 .
Fig. 3. Excited state relaxation pathways of donor and acceptor chromophores in crown-containing dyads BNI2 and BNI3.Plain and dashed arrows denote radiative and non-radiative processes respectively.Sign «NI» denotes naphthalimide moiety.

7 )Fig. 4 .
Fig. 4. Transient absorption spectra of MNI3 at different time delay between pump and probe pulses (a) and transient absorption time profile of MNI3 (b) in acetonitrile.a) 1 -Absorption band corresponding to cation-radical of benzodioxane fragment; 2 -absorption of singlet S 1 state of amido-naphthalimide chromophore; 3 -possible position of absorption signal of amido-naphthalimide anion-radical.b) 1 -Time profile at absorption maximum of amido-naphthalimide excited S 1 state (480 nm); 2 -Time profile at absorption maximum of benzodioxane cationradical (420 nm).

43 Page
demonstrates higher extent of fluorescence enhancement.The reason for the observed difference might be explained by the more efficient PET in the free ligand BNI3, which intensifies FRET switching contrast.It should, however, be said that the binding of Ca 2+ by BNI3 doesn't seem to quench PET interaction completely, because the fluorescence quantum yield of complex (BNI3)•Ca 2+ (ϕ fl = 0.36) is somewhat lower as compared to the one of dyad BNI1 (ϕ fl = 0.47) not containing crown ether moiety.Similar results were obtained for complex (MNI7)•Ca2+  , where sensor properties of monochromophoric naphthalimides MNI6 and MNI7 were studied.

Fig. 6 .
Fig. 6.The changes in the fluorescence spectrum of compound BNI2 (a) and BNI3 (b) in acetonitrile solution.Excitation wavelength λ ex = 340 nm.The inserts show the ratio of the fluorescence intensity at 520 nm measured using excitation light λ ex = 340 nm (I 340 ) to the fluorescence intensity at 520 nm measured using excitation light λ ex = 440 nm (I 440 ).a) Concentration of ligand BNI2 C L = 4.5•10 -6 M; b) concentration of ligand BNI3 C L = 6.5•10 -6 M.

Table 1 .
Photophysical properties and stability constants of mono-and bi-chromophoric naphthalimides and their complexes in acetonitrile at 20 ºC.