Photophysical characterization of a photochromic system: the case of Merocyanine 540

The fluorescence emission of merocyanine 540 (MC540) in 95% ethanol was studied under continuous irradiation. Fluorescence spectra from excited states of both normal (N) and photoisomeric (P) species are identical. The laser fluence dependence of the fluorescence intensity is interpreted on the basis of a photochromic system involving N and P ground states and first excited singlet states. Common flash photolysis equations are generalized in order to include a photoequilibrium between isomers. The emission data are used together with previous flash photolysis and optoacoustic results to obtain P fluorescence and photoisomerization quantum yields as 0.07 {plus minus} 0.02 and 0.20 {plus minus} 0.04, respectively, P absorption cross section at the maximum (560 nm) as 4.74 {times} 10{sup {minus}16} cm{sup 2} (125 {times} 10{sup 3} M{sup {minus}1}{center dot}cm{sup {minus}1}), and the energy difference between the ground states as 165 kJ{center dot}mol{sup {minus}1}.


Introduction
Merocyanine 540 (MC540) is an anionic polymethine dye with various applications in photobiology.1 Since 1974 it has been MC 540 used as a probe to monitor the state of membrane polarization through changes in its absorption or fluorescence properties.2 It was later employed to selectively stain and kill leukemic and immature hemopoietic cells upon irradiation.3 Finally its ability to sensitize the inactivation of encapsulated viruses was reported.4 In view of the multiple applications to photobiology, many studies were devoted to the interaction of MC540 with micelles,5 liposomes,6 and membranes.7 (1) For a recent review see: Sieber, F. Photochem. Photobiol. 1987Photobiol. , 46, 1035Photobiol. -1042 (2) (a) Ross, W. N.; Salzberg, B. M.; Cohen, L. B.; Davila, . V. Biophys. J. 1974, 14, 983. (b) Waggoner, A. J. Membr. Biol. 1976, 27, 317. (c) Waggoner, A. S.; Grinvald, A. Ann. N.Y. Acad. Sci. 1977, 303, 217. (d) Waggoner, A. S. Annu. Rev. Biophys. Bioeng. 1979, 8, 47. (e) Easton, T. G.; Valinsky, J. E.; Reich, E. Cell 1978, 13, 475. (3) (a) Valinsky, J. E.; Easton, T. G.; Reich, E. Cell 1978, 13, 487. (b) Meagher, R. C; Sieber, F.; Spivak, J. L; J. Cellular Physiol. 1983, 116, 118. (c) Sieber, F.; Spivak, J. L.; Sutcliffe, A. M. Proc. Natl. Acad. Sci. USA 1984, 81, 7584. (d) Sieber, F.;Sieber-Blum, M. Cancer Res. 1986, 46, 4892. (4) Sieber Membr. Biol. 1980, 54, 141 Regarding the photophysical properties, the first results were related to the fluorescence characteristics in solution and micelles.8 Recently a fluorescence and photoisomerization study as well as the singlet molecular oxygen sensitization properties was reported9 and, finally, a detailed comparison of the photophysical properties in ethanol and in liposomes was published.10 In alcohols MC540 has a fluorescence quantum yield of ca. 20% at room temperature and a singlet lifetime of 400 ps. It has a very low triplet and singlet oxygen sensitization quantum yields (less than 1% and 0.1%, respectively). Photoisomerization takes place from the first excited singlet state and is a very efficient process (ca. 50% yield). The absorption spectra of the stable ground state (N) and the ground-state photoisomer (P) strongly overlap.9,10 This later characteristic, together with the short singlet lifetime, makes the knowledge of the P photophysics an important point to understand the behavior of the dye upon irradiation, even when short laser pulses are used. The last statement particularly points out the difficulties in the use of complete conversion method in flash photolysis,11 which was performed in ref 9 and 10 to determine the absorption coefficient of P and further to calculate isomerization quantum yield. This method may produce misleading results if P, present during the laser pulse, photochemically reverts to N. In this case, the saturation limit would not reflect a complete conversion; rather N and P coexist in a photostationary state. In a previous work10 an upper limit was reported for the photoisomerization yield from flash photolysis measurements and a lower limit obtained from light-induced optoacoustics (LIOAS) results.
In this work, we perform steady-state fluorescence measurements, which have already proved to be a powerful tool for the characterization of 3,3'-diethyl oxadicarbocyanine iodide (DOD-CI)12 and we use the results together with flash photolysis and LIOAS early results10 to completely characterize N and P pho- Biophys. Res. Commun. 1983, 111, 768. (8) Dixit, N. S.; Mackay, R. A. J. Am. Chem. Soc. 1983, 105, 2928 (PAR 162) which was triggered internally at 1 kHz.
Fluorescence spectra were recorded by using the monochromator at different irradiation fluences or the fluorescence intensity was monitored as a function of excitation fluence by using the interference filter. In this arrangement, the laser power and the boxcar signal were plotted simultaneously in an , , ' recorder (Hewlett Packard). The laser power was varied by means of the electric current or by rotating the half-wavelength plate.

Results
The measurements were performed at 20 ± 2°C. The absorbance of the MC540 solution in ethanol 95% was always less than 0.05 in 1 cm at the excitation wavelength. Typically it was 0. 01 /cm. The substance was excited at 488, 514, 580, 590, and 595 nm.
The excitation fluence (f) was varied between the maximum possible and the minimum which produced a detectable fluorescence signal (typical variation range 3.5 orders of magnitude).
In all cases, the maximum power was included in the range where an appreciable conversion to the photoisomer could be attained.
The MC540 fluorescence emission spectrum is shown in Figure  1. No spectral distribution change could be observed when the excitation intensity was changed by a factor of 25 000, indicating that N and P have also identical fluorescence emission spectra. In view of this result, the fluorescence intensity (7f) in a particular wavelength range is always proportional to the whole emission.
The value of the boxcar output voltage (proportional to f), normalized to the excitation fluence, i.e., Iff, is plotted against the velocity of light absorption a = /¡ at different excitation  In these plots the low excitation a-independent range corresponds to N fluorescence and the high, also a-independent range corresponds to the photoequilibrium between P and N after the saturation is reached.

Discussion
The system can be kinetically analyzed on the basis of the four-level scheme depicted in  Considering that the excited-state concentrations are much less than the ground states and neglecting the depopulation of *N and *P by circulation (w « 1 /rfN; 1 /rfp), we arrive at the expression [P] _ß _ ( + 5 )< + k + w 0 (2) where B = / = eP/tN, and are photoisomerization quantum yields from 'N to P and 'P to N respectively, c0 is the total concentration, and w is the rate of change of concentration by circulation of the solution.
The total fluorescence intensity is with , fluorescence quantum yields. If we replace (2) in (3) and [P] + [N] = c0 is used, we obtain (4) where AT is a proportionality constant, G = B(p + f)/(Bp + 1), as = (k + /( (5/> + !)),/> = / , and/= / . Equation 4 is used to fit the data of Figure 2. The fitting parameters are also quoted in the graphs. Equation 4 contains three adjustable parameters K, as, and G. No useful information can be extracted from K, which depends on factors such as spectral bandwidth, photomultiplier sensitivity and applied voltage, and geometry. as is not meaningful either because w, the speed of sample renewal within the irradiated volume, is unknown due to turbulence in the flow. The useful information, regarding fluorescence and photoisomerization quantum yields of N and P, is contained in G. This parameter is obtained from the measurements at different wavelengths, and it is used in the form O'1 = (p/p+f) + (l/p+yjfl-1 A plot of G~' vs S'1 allows the calculation of p and /. Nevertheless, this is not a straightforward process and an iterative method had to be used, because B values are unknown. In order to know them, we need a further analysis of the flash photolysis data of ref 10 (6b) x is a factor that takes into account the back isomerization 'P to N and is defined as Equations 6a and 6b correspond to eq A5 and A3 of the Appendix and the definition of eq 6c is included in eq Al. From the quoted limits of we can establish that x ranges between 1 and 0.6 at lower limit. Flash photolysis results must be analyzed for ^0. In the Appendix, the common flash photolysis equations, obtained for complete conversion, are generalized in order to include the photoequilibrium. Absorbance difference spectra obtained either at low energy excitation ( ) or at saturation regime ( ") can be expressed in terms of x. The values of , Ae, and as a function of x, calculated by using eqs 6, are given in Table I.
To obtain p and / we have to use an iterative procedure. We begin with the B values obtained from the assumption of = 0 (x = 1) in order to obtain the first estimation of p and / from the slope and intercept of the representation of eq 5. The value of p gives a better estimate of x. From the specific value of x, the parameters quoted in Table I can be calculated by using eqs 6 and a corresponding set of B is obtained. From the plot of 1 /G vs 1/5 a new estimation of p allows the procedure to be reproduced until two successive iterations render no difference in the value of x. The final values are plotted in Figure 4, from which p = 0.49 ± 0.17 /= 0.39 ±0.13 (7) From Table I, taking the latter value of p as entry, we estimate x = 0.74 ± 0.08. This renders = 0.40 ± 0.04 from eq 6a. The photophysical parameters of the photochromic system P,N are quoted in Table II. The lifetime of 'P (P fluorescence lifetime) was calculated from the formula of Strickler and Berg13 and = 0.07. The value may be considered realistic on the basis of the good agreement obtained from the similar calculations for 'N. where a is related to the branching ratio defined by Rulliere.14 Equation 8 does not hold for MC540 according to the results of Table II. In fact, /(1 -) = 0.48 and /(1 -) = 0.22. This result indicates that the model cannot be applied to MC540; i.e., either 'N or 'P or both of them deactivate through direct internal conversion. Regarding the values of = 0.20 and (· = 0.07, we conclude that the main deactivation of 'P is internal conversion which can take place as a direct process or through a twisted state. A similar result was found for DODd 12.15,16 Finally, we can obtain quantitative conclusions regarding the activation barrier in the excited singlet energy surface which is responsible for the temperature dependence of and .10 We assume that all the processes competing with fluorescence take place from a twisted excited state (t). Under this condition T N = (fcN + *Nl)-' (9) where /cfN is the radiative rate constant of N and /cNt the one corresponding to the passage to the twisted state, kf1 is assumed to be indépendent of temperature and has a value of 3.95 X 10s s"1 calculated from = &fNTfN (10) and the values of Table II. If we replace (9) in (10), we obtain We can calculate the temperature dependence of kNl using eq 11 and the values of Table I in ref 10 for at temperatures lower than 300 K. In this way, 23 ± 1 kJ-mol"1 is obtained from an Arrhenius plot. This is the activation energy for the photoisomerization step. From the value of the photoisomerization rate constant kw = / = 0.40 X (430 ps)"1 = 9.30 X 108 s"1 at 20°C (see Table II), a preexponential factor of 1.2 X 1013 s"1 is calculated.
The experimental activation energy, obtained under the assumption of a complete nonradiative deactivation via a twisted state, is equal to the energy barrier between *N and "t". This model may not be realistic for 'N, as was discussed above, and direct internal conversion may take place. In the latter case, the value of 23 kJ-mol"1 is a lower limit for the energy barrier.

Appendix
Flash Photolysis Equations under Photoequilibrium. We shall discuss some results of ref 10 and we shall generalize some common flash photolysis equations and apply them to a photochromic system.
From the application of the complete conversion method in flash photolysis, the difference absorption coefficient (Ac = cP -cN) at 560 nm was calculated as Ac^= -30 X 103 VT'cm-1 assuming that P did not isomerize back to N (i.e., = 0). When this value was used to compute the isomerization quantum yield ( ), a value of 0.54 was obtained.
On the other hand, assuming that P cannot be higher in energy than 'N for an efficient photoisomerization, i.e., A£P as defined in Figure 3 is 210 kJ/mol (the energy gap between N and *N) and taking into account the result £ = 67 ± 6 kJ/mol from LIOAS, = 0.32 is obtained. The corresponding value for the difference absorption coefficient would be Ac560 = -50 X 103 M_1-cm_1. This limit would correspond to the maximum photoisomerization quantum yield from P to N ( ).
During the laser irradiation (pulse of 15 ns fwhm) we can integrate the kinetic equations derived from Figure 3 under the conditions of low absorbance and steady state for only 'N and 'P (see eq 1) and in the absence of flux (w = 0), neglecting the thermal decay k from P to N. The integration renders, for [P]f, the concentration of P at the end of the pulse of energy Eh cross section S, and wavelength [P]f = £ 6 + 6 CoU -e~bEl\ = ) = xc"(l (Al) with b -2.303(6 + 6 )^-ĉ 0 is the total concentration, and x is defined as the fraction of c0 which is present as P in the saturation-photostationary state ([P]" corresponds to £) » ¿r1). If = 0, x = 1 is the limit of the complete conversion. If ^0, then the measurement of AA" (the absorbance difference at saturation) at a certain analyzing wavelength, , is

Conclusions
The measurements performed in this work analyzed in conjunction with results from flash photolysis experiments provide a complete quantitative description of a photochromic system with very short lifetimes for excited states and complete overlap of ground-state absorption.
It must be pointed out that flash photolysis experiments need to be carefully interpreted in order to avoid erroneous results, particularly when high fluences are used in the excitation pulse.
Steady-state fluorescence is a very sensitive technique to study photochromic systems. It has the additional advantage that a maximum conversion limit can be reached with very low excited singlet state populations.  (16) Rentsch, S.; Dópel, E.; Petrov, V. Appl. Phys. B 1988, 46, 357. When AeK from eq A3 is replaced in eq A4, we obtain eq 6a of AA"(\a)Ia We now have the equations to obtain the values of Table I.

Equation A5
gives in terms of x. Equation A3, the definition of AeXo, and eq Al were used to calculate Ae(X), < ( ), and , respectively.