J. Phys. Chem. 1994,98, 3247-3256
3247
ARTICLES Photophysics of the Lactone Form of Rhodamine 101 Jerzy Karpiuk,'v+ Zbigniew R. Grabowski,+ and Frans C. De SchryveP Institute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01 -224 Warsaw, Poland, and Department of Chemistry, Catholic University of Leuven, Celestijnenlaan ZOOF, 3001 Heverlee. Belgium Received: May 4, 1993; In Final Form: September 24, 1993"
The photophysics of the lactone form of the rhodamine 101 has been investigated in nonpolar and polar aprotic solvents. A single broad luminescence band in nonpolar or weakly polar solvents is observed from a chargetransfer (CT) state (with the dipole moment of =26 D) produced in the electron-transfer reaction in the excited state. In more polar solvents a second fluorescence band appears, revealing the dissociation of the C-0 lactone bond. The excited singlet state of the zwitterion (Z) is formed. Temperature-dependent measurements show that the zwitterion form is produced in all solvents; it is, however, quenched in less polar media. It has been found that the quantum efficiency of population of the Z form in the excited singlet state does not depend on the solvent and equals 0.23 f 0.03, which indicates purely intramolecular control of the branching into C T and zwitterion excited states. The excited state reaction in frozen solvents led to intense phosphorescence from a low lying (i, triplet i * state ) of the Z form. The ratio of fluorescence to phosphorescence intensities of the zwitterion (1:3) in rigid glass implies a supposition that the Z form is created in singlet and triplet states according to their spin statistical factors. The results of this work force us to verify the existing views on the spectroscopy of rhodamines and the role of the solvent in photophysics of these molecules. The observed intramolecular quenching of zwitterions has been ascribed to deactivation to a higher triplet, most probably of (n,?r*) nature. We suspect this mechanism to be responsible for the thermally activated nonradiative process in rhodamines. The lack of phosphorescence of rhodamine 101 in protic solvents is explained by an increase of energy of the 3(n,r*) state, presumably due to the hydrogen bonding. The data do not support the transition to the TICT state as a mechanism of thermally activated quenching of rhodamines.
Iatroduction The photophysics and photochemistry of the lactone form of the rhodamine B (LRB)l-3involvethreeof theimportant problems in current organicphysical chemistry, namely (i) the solvent effect on the photophysical processes of a molecule in the excited state, (ii) the surface crossing in passing from one electronic state to another, and (iii) the photoinduced electron-transfer reaction in the excited state. The electron transfer from the xanthene part of the molecule to the phthalide one, which seems to be an initializing step in the photophysics of this molecule, results in a charge-transfer (CT) excited state or produces the excited zwitterionic form (Z) of the molecule after bond cleavage between the C and 0 atoms in the lactonic ring, passing in this way ovto the potential energy hypersurface of the excited state of the zwitterion, Z*. Closely related to these questions is a problem more particularly connected with the class of rhodamine dyes, namely the radiationless deactivation pathway observed for zwitterion and/or cation forms of these dyes. This phenomenon was studied by several group.C1O The conclusions drawn from these investigations suggested that conformationalchanges of the amino groups in the excited state are responsible for the nonradiative decay. Drexhage4 postulated that torsional motion of the diethylamino group in rhodamine B is involved in the nonradiativeprocess. The synthesis of rhodamine 101 having amino groups rigidly linked to the xanthene skeleton, its very high luminescence quantum yield, and-in particular, the independence of this quantum yield on temperaturtstemed to be very strong arguments confirming ~~
t Polish Academy of Sciences.
t Catholic University of Leuven. *Abstract published in Advance ACS Abstracts, March 15, 1994.
0022-3654/94/2098-3247S04.50/0
that this model explains the nature of the nonradiative decay. Vogel et al.' investigated several rhodamines with different substitution patterns at the amino nitrogens and postulated the presence of a nonemissive TICT state being accessible from the S1state after crossing an activationbarrier in the torsional motion. This explanation found support in other studies.gJ0 The increase of temperature would in this model enable more efficient crossing of the barrier between the S1 excited state of the zwitterion and the TICT state, which should be accomplished by twisting of the dialkylamino groups. The occurrence of this dependence has also been reported to be linked to the degree of alkylation of the amino groups. One of the main structural arguments which supported the TICT hypothesis for rhodamine dyes was the lack of temperature dependence of the fluorescence quantum yields and lifetimes for the ethyl ester derivative of rhodamine 101. According to the TICT model, the rigidly fixed amino group cannot yield the TICT states, the twisting motion being impossible. This question is also of great importance for the photophysics of lactones in view of the observed intramolecular quenching of the zwitterions generated in the excited-state r e a ~ t i o n . ~ In order to clarify. the processes taking place in the excited states of lactone forms of rhodamines and, in particular, to explain the role that the solvent plays in the photophysics of these molecules, we have undertaken the investigation of the photophysics of the lactone form of the rhodamine 101 (LR101). The dye rhodamine 101 (R101) is well-known for its excellent luminescence properties and is frequently used as an activemedium in dye lasers. Because of its very high quantum yield (about 1.011J2) and the insensitivity of this yield to variations of the temperature, it has also been proposed as a standard for fluorescencemeas~rements.'~These applications relate, however, 0 1994 American Chemical Society
Karpiuk et al.
3248 The Journal of Physical Chemistry, Vol. 98, No. 13, 1994 to the dilute solutions of the salt (mostly perchlorate) of the dye, where it occurs in the form of the zwitterion (ZR101) which is only one of the possible forms of the dye. Molecules of rhodamine 101,similar to the other rhodamines possessing a carboxylphenyl group in their molecules, may participate in equilibria between different molecular forms, three main ones of which are the lactone, the zwitterion, and the cation.
that by Barigelletti16who has studied the absorption spectra of rhodamine 101 lactone in a mixture of propionitrile and butyronitrileas a function of temperature and reported equilibrium constants for lactone-zwitterion equilibria in this solvent mixture. Studies of the lactone have the additional advantage that no protons or counterions are present in solution.
Experimental Section
lactone
zwitterion
cation Equilibria between such forms of rhodamine B in the ground state have been known for a long time.I4 The existence of the dye in a definite molecular form depends first of all on solvent polarity and proticity. The zwitterion is favored by polar protic solvents whereas the lactone dominates in nonpolar and polar aprotic media. Cations occur under acidic conditions. Recent studies on the ground-state equilibria between lactone and zwitterion forms of rhodamine B suggested that even in ethanol at room temperature, although the equilibrium is strongly shifted toward the zwitterion, there are a considerable amount of molecules in lactone form.15 This fact additionally enhances the importance of studying the photophysical behavior of this form of the dye. The lactone molecule-due to interruption of the ?r chain in the xanthene moiety-represents from the point of view of optical transitions a completely different system compared to the zwitterion.'-3 Depending on the polarity of the solvent, the rhodamine B lactone displays either a single luminescence band at shorter wavelengths or a dual luminescence,one band of which (at longer waves) has been ascribed to originate from the zwitterionic form produced in the dissociationprocess in its excited state.192 This paper is part of the study of the photophysics and photochemistry of lactonic forms of the rhodamines. Here we describe the photophysics of the lactone of rhodamine 101. The results allowed us to verify critically the existing picture of the photophysics of rhodamines, primarily due to thediscovery of the intense phosphorescence of this dye and identification of the role of the triplet state. The obtained results will also enable a more critical assessment of the TICT hypothesis in the case of rhodamines and at the same time show that the picture of the processes taking place in the excited states of the rhodamine dyes is much more complicated than it might be anticipated on the basis of the studies performed on the colored (ionicor zwitterionic) forms. The photophysical propertiesof the lactonic form of rhodamine 101 have not been investigated until now. The only report known to us which has dealt with the lactone form of rhodamine 101 was
Rhodamine 101 has been converted to its lactone form by hot extraction of the dye with toluene from the saturated solution of rhodamine 101 salt (perchlorate, laser grade, Lambda Physik) in 0.01 M NaOH. After the toluene fraction had been enriched with the colorless lactone, the toluene layer was separated from the aqueous layer and the solvent was evaporated. The slightly violet powder product was dried under vacuum over P205. On dissolving the product in an aprotic solvent such as 1,2dimethoxyethane, no absorption in the visible is observed. In turn, dissolution in methanol results in an absorption identical to that of rhodamine 101 perchlorate salt. Spectroscopic grade solvents n-dibutyl ether (Merck), diethyl ether (Merck, for fluorescence), 1,Zdimethoxyethane (DME, Aldrich, for HPLC), tetrahydrofuran (Merck), pyridine (Janssen), methanol (Merck), and absolute ethanol (Merck) were used as received from freshly opened bottles. Butyronitrile (BN) and propionitrile (PN) (Merck, for synthesis) were distilled three times: over KMn04, over PzOS, and over CaH2. Methyltetrahydrofuran (MTHF) was preliminary dried and subsequently distilled twice over CaH2. Samples of purified DMF were kindly obtained from Dr. A. Kapturkiewicz. Absorption spectra have been measured with Zeiss Specord M40 or Perkin Elmer Lambda 6 spectrometers. Fluorescence spectra were recorded with a Jasny spectr~fluorimeterl~ or with a SPEX Fluorolog 212. Quantum yields were determined using quinine sulphate in 0.1 N H2SO4 as a luminescence standard (@ = 0.51). The fluorescence spectra have been corrected for the spectral response of the detection system. Decay measurements were performed using the single-photon timing (SFT)instrument based on a synchronouslypumped, cavity dumped, frequency doubled DCM (320 nm) or rhodamine 6G (295 nm) dye laser.18 The detection system was essentially the same as previously described except the monochromator applied to observe the fluorescence was the American Holographic DB 101 double-subtractive monochromator. The fluorescence was observed through a polarizer set at the magic angle. Repetition of the excitation pulses was 800 kHz. The response function of the detection system was measured by using reference compounds at the same wavelength as the measured fluores~ence.1~ The referencesused were 1,2-dimethyl-POPOPin methylcyclohexane ( T = 1.13 ns) when in the spectral range below 540 nm and erythrosine in water ( T = 95 ps) in the 550-610-nm range. The temperature of the samples was maintained within f l K. The samples were not degassed as we have not found any marked influence of the dissolved oxygen on the decay curves with the exception of the increase of the longer decay component after the sample in DME was degassed by the freezepumpthaw technique. Time-resolved spectra were recorded by means of a sampling technique using a nitrogen laser excitation (FWHM = 0.6 ns) and the boxcar equipped with a sampler of gate duration 220 f 50 ps. The instrument was described elsewhere.20 The time resolution was i l ns. Measurements were carried out in the concentration range of 1W-10-6 M to check a possible influence of concentration of the observed results. No dependence of the fluorescence quantum yields on the solute concentration in this range was found, and the concentration typically used in the measurements was that of 1 0 - 5 M.
Photophysics of the Lactone Form of Rhodamine 101 Fluorescencedecays were analyzed by the single-curvemethod using the DECAN deconvolution program with reference convolution based on the Marquardt algorithm.21
The Journal of Physical Chemistry, Vol. 98, No. 13, 1994 3249 i T H F
ReSult.9
Absorption Spectra. In contrast to the zwitterionic forms of the rhodamine dyes which absorb strongly in the visible and emit fluorescence very efficiently, lactone forms of these dyes-due to the interruption of the *-electronic system characteristic for the xanthene moiety-do not possess any chromophore able to absorb in the visible.4 The same situation can be expected for the lactone of rhodamine 101. Therefore any absorption in the 500-590-nm region would indicate a presence of the ionic (or zwitterionic) forms of the dye in the ground state. Two forms of the dye, the zwitterion and the cation, have their absorption maxima in methanol at 568 and 577 nm, respectively. The absorption spectrum of the cation has been measured after 1 drop of 0.1 N HzSO4 has been added to the cell containing a methanol solution of LR101. The same method has been used to measure the absorption spectrum of the cation which is formed in DME after a drop of H2S04has been added. The absorption maximum is observed at 577 nm. The distinction between cation and zwitterion is important because these two kinds of species are formed in two different ways. The presence of the zwitterion might be caused by a possible (temperature dependent) groundstate equilibrium between lactonicand zwitterionicforms,whereas the cation in the aprotic media can originate only from the protonation of the dissolved lactone molecules by traces of protic impurities,almost alwayspresent, especially in more polar solvents. The possible presence of zwitterions in the ground state in larger amounts would complicate the interpretation of the results. Therefore the absorption spectra in the visible are of particular importance with respect to the ground-state equilibria. The Occurrenceof such an equilibrium in butyronitrile propionitrile mixtures has been reported by Barigelletti.16 The investigated solutions were checked carefully for the absorption in the visible region. In low polarity solvents like toluene or dibutyl ether we found a residual absorbance with a maximum at 583 or 585 nm, respectively. This absorbance was ascribed to the presence of cations resulting from residual amounts of protic impurities. In DME or MTHF no absorption in the visible has been found within the accuracy of the absorption spectrophotometer (0,001 absorbance unit) at the lactone concentration [L] = 8.40 X 10-5 M, which corresponds to an absorbance of 1.04 at 317 nm (6317 = 12 400 M-1 cm-l). Using the value of the molecular extinction coefficientfor the zwitterion in EtOH, 8568 = 95000 M-I cm-1,22and the value of 0.001 absorbance units as an estimation of the upper limit of the concentration of zwitterions in the ground state, we obtain [Z] 5 1 X 10-8 M. The ground-state equilibrium constant K(L * Z)1 1.2X 10-4andtheresultingAGo>+5.15kcal-mol-*(which corresponds to 1800 cm-1) in 1,2-DME, at 293 K. This value of AGO in DME exceeds by about 2.86 kcal-mol-l (1000 cm-l) the value given by Barigelletti for a solution of the inner salt of RlOl in a BN + PN mixture.16 OurmeasurementsonLRlOl ina BN +PNmixture(4:5,v/v, as used in ref 16) indicate the presence of thecation form of RlOl rather than the zwitterion one. In contrast to Barigelletti, who reported the presence of an absorption band with a maximum at 568 nm and ascribed this band to the zwitterion, we haveobserved a (small) peak at 577 nm. Doubling the lactone concentration (measured in proportion to the intensity of the first absorption band of the lactone at 320 nm) resulted in only a 25% increase of the absorbance at 577 nm. This could be explained by saturation of the protic impurities protonating the lactone molecules. After 1 drop of 0.1 N H2SO4 was added to the cell containing 3 mL of the BN + PN solution of LR101, the solution became deep purple and the absorption spectrum changed
+
220
320
270
370
wavelength [nm] Figure 1. Absorption spectra of the lactone form of rhodamine 101 (LR101)in aproticsolventsat 293 K, continuouscurve,butylcther (BE); dashed curve, tetrahydrofuran (THF); dot curve, butyronitrilc (BN). completely. The maximumof the peakin thevisible (most intense) was at 577 nm, and the spectrum in the UV indicated a full transformation of the lactone into the colored form. On the basis of these findings we believe that the dominant source of the absorption in the visible in BN PN solution is the cation form of the molecule produced through protonation by protic impurities present in the solvent. Hinckley et al.15 have postulated that the necessary conditions for the stabilization of the zwitterion in a given solvent are (i) hydrogen bonding and (ii) sufficient dielectric/polarizability capabilities of the solvent. Our results suggest, however, that despite failing the fulfillment of the first requirement in aprotic solvents the observation of the L Z equilibrium in these media might be possible. Therefore, we do not exclude the presence of the zwitterion resulting from a possible equilibrium with the lactone form, especially on the basis of a small blue shift of the absorption maximum accompanied by a decrease of the band intensity with an increase of temperature, even if thermal expansion of the solvent is taken into account (576 nm at +20 OC, 574 nm at +SO OC, and 572 nm at +70 “C). The energy difference of the two species is anyway higher than that given in ref 16. Assuming the molar extinction coefficient for cationic species of RlOl to be equal to that of the zwitterion (Le., 95 000 M-I cm-I), we could estimate the total concentration of the ionic forms in a 2.8 X 10-5 M BN PN solution of LRlOl to be 3.7 X lo-’ M. This roughly corresponded to the value found by Barigelletti; we have shown, however, that it cannot be ascribed entirely to the zwitterion. This in turn further increases the estimate of the AGO value at 293 K for this system, as compared to that in ref 16. A priori assumption that only 25% of the absorption in the visible is to be ascribed to the zwitterion would yield the AGO = +3.29 kcal-mol-1 (1 150 cm-1 compared with 830 cm-1 given in ref 16). The real value could be even higher. The absorption spectra of the LRlOl in aprotic solvents of different polarities (Figure 1) are similar to those of LRB. The first absorption band exhibits slightly positive solvatochromism (shift by 500 cm-1 on going from dibutyl ether to butyronitrile). The band ends with a low intensity tail that extends to about 400 nm. On the basis of the analogy of the position of the first band in the absorption spectra of lactones of rhodamines with different substituents at amino groups with these of correspondingaromatic amines,23 we postulate that the transition in the first absorption band is a localized one within the “julolidine”part of the LRlOl molecule. The absorption spectra of the methanol or ethanol solutions of LRlOl indicate the presence of the zwitterions only. Fluorescence Spectra at Room Temperature. Fluorescence spectra of the solutions of LRlOl in different solvents at room temperature (Figure 2) are, as in the case of rhodamine B lactone,
+
+
Karpiuk et al.
3250 The Journal of Physical Chemistry, Vol. 98, No. 13, 1994 t,,
[x
10-3 cm-I]
22
f(E,
n)
Figure 3. Solvatochromic shifts of the maximum of lactone band of 400
500
600
700
wavelength [nm] Figure 2. Fluorescence spectra of rhodamine 101 lactone (LRlOl) in various solvents at 293 K. Abbreviations: BE, butyl ether; EE, ethyl ether; MTHF, methyltetrahydrofuran;DME, dimethoxyethane; BN, butyronitrile. strongly solvent dependent. In low polarity solvents like toluene, and dibutyl or diethyl ether, the spectrum consists of one broad band. In medium polarity solvents like chlorobenzene, butyl or ethyl acetate, and DME or THF, a second emission band appears. Double luminescence occurs also in dioxane. Increase in solvent polarity causes a gradual disappearance of the short-wavelength band, and the fluorescence spectrum seems to be composed of only one band which is accompanied by a strong increase in the observed quantum yield of the long-wavelength luminescence. This picture occurs in pyridine, butyronitrile, propionitrile, acetonitrile, and DMSO. In the following we will adopt the nomenclature used previously3 and refer to the higher energy band as the L (Lactone) band and to the lower energy band as the Z (Zwitterion) one since the Z band can, on the basis of its shape and spectral position, be ascribed to come from the excited zwitterion formed in the excited state. The excitation spectra of both luminescences as measured in THF are identical and match the absorption spectrum of the lactone. The L band is very sensitive to the solvent polarity, and its maximum shifts from 21 500 cm-1 in dibutyl ether to 19 160 cm-1 in THF. Also noticeable is the Stokes shift between absorption and fluorescence spectra, in the case of MTHF being almost 12 000 cm-1. The shift of the Z band as a function of the solvent is very moderate (maxima at 17 250 cm-1 and at 16 800 cm-1 in DME and DMSO, respectively). In order to estimate the dipole moment of the state emitting the L band (hereinafter referred to as the L& excited state), we have neglected the polarizability effects and assumed that the lifetime of the L& state is sufficiently larger than the solvent reorientation relaxation time. Then, using the model given by Liptay,24 the dipole moment was estimated from the slope of the linear plot of position of the maximum of L band, vmX, vs the function describing the solvent polarityf(a,n) = (a - l)/(2e + 1) - (n2- 1)/2(2n2 + 1) (Figure 3). Taking the value of 0.7 nm for the cavity radius, we obtain a value of 26 D for the difference between the excited and ground state dipole moment. Assuming the dipole moment for the ground state of the lactone molecule to be negligible, we obtain pcxcsi 26 D. Such a largevalue indicates full separation of charge (for 1 electron) over a distance of 5 A. This value is also similar to that found for the CT excited state of LRB.1 In analogy to the lactone of rhodamine B we ascribe the L fluorescence band to a highly polar excited state being formed through an electron transfer from the xanthene moiety to the phthalide one, which is not accompanied by the C-0 bond cleavage.
LRlOl as a function of the solvent polarity.
In contrast to the L band, the Z fluorescence does not exhibit such a solvent dependence, and therefore the dipole moment change on the transition corresponding to this band is rather small compared to the lactone form. Using thecharge distribution calculated by Bergamasco et for the xanthene system in the ionic form of rhodamine, the difference of the dipole moments in the excited state and in the ground state of the zwitterion can besafelyestimated toless than 10D. Assumingthedipolemoment for the ground state of the zwitterion of RlOl to be similar to that found for rhodamine B,Z6we can conclude that the dipole moment of the Z ; state is much smaller than that of the L&state. This in turn would indicate a significant difference in electron density distribution of both excited states and, consequently, their different electronic structures. Upon changing the Wavelength of the excitation light, we did not find any differences in the ratio of the intensities of both bands. The photophysical properties of LRlOl in various solvents are summarized in Table 1. Thequantum yield of the L fluorescence decreases with increase of the solvent polarity whereas the quantum yield of the Z fluorescence increases dramatically along with its decay time. This suggests the Occurrence of an efficient (intramolecular) quenching mechanism of this fluorescence in less polar solvents and stabilization of the zwitterion excited state with increasing solvent polarity. Fluorescence Decays at Room Temperature. The fluorescence decays also depend strongly on the solvent used and are emission wavelength dependent. Generally, the decays measured in the L band could be fitted as a monoexponential decay, yielding the lifetimes shown in Table 1, On the other hand, decays measured in the Z band were more complicated and generally had to be fitted by two- or three-exponential decays. Especially in cases where a double luminescence occurs, one has to assume a considerable overlap of both luminescences within the Z band. The temperature dependence of the fluorescence (see below) indicates that even in cases where apparently only one band (Z) is displayed there still can be two strongly overlapping bands, L and Z, with the lactone band being small, however detectable. In order to separate the contribution from the L band, we analyzed decays in the Z band by keeping the longest decay time constant and equal to the decay time recovered for the L band. The global analysis performed for the LRlOl decays in DME confirms the necessity of adding a second component ( 7 2 ) in the analysis of the decays of the zwitterion fluorescence after the contribution from the overlap with the L band has been separated. The nonexponentiality of the decays of Z fluorescence in solvents where the zwitterion excited states are quenched has been extensively investigated for LRB,3 so it seems to be an inherent property of the system. The decay curves measured at fluorescence maxima in the solvents where Z fluorescence is observed with relatively high quantum yield (BN, pyridine) could be fitted with a single decay time (similar to LRB in benzonitrile, when the
The Journal of Physical Chemistry, Vol. 98, NO. 13, 1994 3251
Photophysics of the Lactone Form of Rhodamine 101
TABLE 1: Lactone Form of Rhodamine 101. Photophysical Properties as a Function of the Solvent solvent dibutyl ether diethyl ether MTHF
'A
[cm-ll
v,,
;k [cm-']
31 750 31 750 31 550 31 550 31 550 30 950 31 250 31 250 31 050 17 650
;FZ
mu
21 500 20 830 19 690 19 160 19 420 19 050
[cm-ll
*L
0.023 0.034 0.033 0.006 0.004
