J. Phys. Chem. 1981, 85, 2050-2054
2050
the triplet-state decay at 77 K in the present study is most likely due to nonreproducible distribution of the heavy atoms relative to the phosphor upon freezing. This would explain the present data because of the various conformations that the rotating -CH2CH2X substituent can assume relative to the carbazole ring system. These would average out in the dynamic MS-RTP system.
MS-RTP has been observed via the internal heavy-atom effect for bromo- and chloronaphthalene, bromobiphenyl, and bromotriphenyl without the addition of external heavy a t o m ~ . ~ . ~The * ~ heavy J ~ atom bonded directly to these chromophores was sufficient for induction, through spinorbit coupling, of MS-RTP. Additional data on the 77 K lifetimes of the halogensubstituted carbazoles dissolved in hexane, given in Table I, lend support to this explanation of the effect of halogens on MS-RTP intensities and lifetimes. Single exponential decay curves were obtained for carbazole and N-(2cyanoethy1)carbazole over the time span studied. The decay curves for N-(2-iodoethyl)carbazole and N-(2chloroethy1)carbazole are similar, showing two components, one with a short 1.7-9 lifetime and another nonreproducible, longer lifetime. Multicomponent lifetimes have been commonly observed in 77 K studies of phosphors and external heavy atomsagIn the internal heavy-atom effect, a 77 K phosphorimetry study showed that variations in bromine’s position relative to the phosphor produced large variations in the radiative rate constanb among the various positional isomers studied.16 The nonreproducibility of
Summary Although few phosphorescent heterocyclic compounds have been induced to exhibit MS-RTP to date, there appears to be no intrinsic reason why they canot be made to do so under more favorable experimental conditions. Specifically,in the carbazole series reported here, the most important factors in the induction of room-temperature phosphorescence appear to be the relative solubility of the species in the micellar aggregate as influenced by the substituted species polarity and the efficiency of the micelle in organizing external heavy atoms and chromophores for effective spin-orbit coupling. By judicious choice of heavy atoms and surfactant, it may be possible to induce MS-RTP in other heterocyclic aromatic compounds.
(15) N. J. Turro and M. Aikawa, J. Am. Chem. SOC.,102,4866 (1980). (16) A. K. Chandra, N. J. Turro, A. L. Lyons, Jr., and P. Stone, J.Am. Chem. SOC.,100, 4964 (1978).
Acknowledgment. We thank Professor J. A. Hirsch for discussions and F. for of the come pounds studied.
Temperature Dependence of Electronic Energy Transfer to Cr( 111) Complexes. Evidence for the Rate-Determining Effect of the Preexponential Factor M. Maestri’ and D. Sandrini Istittito Chimico “G. Clamlcian” dell’Universit6 di Bologna, Bologna, Italy (Received: March 19, 1980; In Final Form: March 3, 1981)
The variations in the rate constants for the energy-transfer quenching of (3CT)Ru(phen)z(CN)zby a family of Cr(enIzXY+complexes have been examined. The activation energies associated with the energy-transfer processes have also been measured. It is shown that the low values of k , exhibited by all of the complexes cannot be attributed to activation energy terms. It is suggested that the reason for the low rate constanb lies in the low values of the preexponential factor, ke2, caused by poor overlap between donor and acceptor orbitals. Correlations between kenoand the nephelauxetic parameter p of the various complexes strongly support this hypothesis.
Introduction Excited-state quenching processes involving transitionmetal complexes are a subject of great intere~t.l-~In principle, such a quenching may occur by a t least three different mechanisms,2 i.e., (i) energy transfer, (ii) electron transfer, and (iii) “catalytic” deactivation caused by heavy-atom or paramagnetic effects. The actual nature of the quenching mechanism has been established in several cases on the basis of the observed sensitized emission of the metal complex (energy-transfer quenching) or the (1) F. Wilkinson, Pure Appl. Chem., 41, 661 (1975). (2) V. Balzani, L. Moggi, M. F. Manfrin, F. Bolletta, and G. S. Laurence, Coord. Chem. Rev., 15, 321 (1975).
(3) V. L. Ermolaev, E. G. Sveshnikova, and T. A. Shekhverdov, Russ. Chem. Rev. (Engl. Transl.),44, 26 (1975). (4) V. Balzani, F. Bolletta, M. T. Gandolfi, and M. Maestri, Top. Curr. Chem., 75, 1 (1978). (5) E. J. Marshall, N. A. Philipson, and M. J. Pilling, J. Chem. SOC., Faraday Trans. 2,72,830 (1976); E. J. Marshall and M. J. Pilling, ibid., 74, 579 (1978); J. B. Ersts and M. J. Pilling, ibid.,74, 1403 (1978). 0022-3654/81/2085-2050$01.25/0
formation of reduced or oxidized species of the reactants (electron-transfer quenching). Catalytic deactivation is a low-efficiency mechanism which cannot compete with energy or electron transfer. In the case of energy transfer, the rate constants have been found to depend critically on many parameters, e.g., the donor, the nature of the ligands, the stereochemistry and charge of the complex, and the solvent.1-3,5 For an energy-transfer quenching mechanism eq 1can k-d + A kG4 *D*.-Akklm P D****A G D + *A , kd
*D
(1)
kd
be used.6 The rate constant of the unimolecular step which converts the *D.. .A precursor into D-..*A successor complex in the encounter, according to the absolute reaction rate t h e ~ r ycan , ~ be written as eq 2 , where AG% (6) V. Balzani, F. Bolletta, and F. Scandola, J.Am. Chem. SOC.,102, 2152 (1980).
0 1981 American Chemical Society
The Journal of Physical Chemistry, Vol. 85,
Electronic Energy Transfer to Cr(II1) Complexes
ken = kenOe-AG’/RT = ( k T / h ) k e A S ’ / R e - f l / R T
(2)
the free energy of activation of this step, k e 2 is the preexponential factor, AH* is related to the experimental activation energy, AS*is related to the partition functions of the activated and initial states, and k , usually called the transmission coefficient,is introduced to take into account the possibility that the process is not adiabatic because of unfavorable electronic interactions. When the donoracceptor interaction is sufficiently weak, the internal and solvent degrees of freedom are unperturbed in the transition state, so that AS* can be assumed to be negligible,6 and AG*
c31$
(3)
In such an approximation, eq 2 reduces to
ken = k e 2 e - f l / R T= ( k T / h ) k e - @ J R T
(4)
When steady-state approximations are used, the forward kinetic scheme (eq 1) leads to eq 5 for the experimental
k, =
kd
1
+ eAGIRT+ (k-d/keno)f?AG’/RT
(5)
quenching constants, where AG is the standard free energy of the forward energy-transfer step. Equation 5 is a more nearly correct and more general formulation of the Sandr0s6~9 equation (eq 6). To a first approximation, AG is kd 4‘‘
= 1 + ,-[E(*DP)-E(*A,A)]/RT
(6)
given by the difference between the zero-zero spectroscopic energy of the donor and the acceptor (eq 7). As is cusAG = -Ew(*D,D)
+ Ew(*A,A)
(7)
tomary in chemical kinetics, the free energy of activation can be expressed as a function of the free-energy change. An empirical relation of this kind, which has been successfully used for energy-transfer processes, is given by eq 8.6910 In this equation AG*(O) is the so-called intrinsic
AG* = AG
+ [AG*(O)/ln 21 In [ l + exp(-AG In 2/AG*(0))] (8)
barrier, which is a measure of the distortion of the internal nuclear coordinates of the molecule and of the solvent arrangement around the molecule in going from the ground to the excited state.6 When AG*(O) N 0 (Le., when the excited states involved are not distorted or only slightly distorted) and k e 2 >> k d , eq 5 reduces to the Sandros equation (eq 6), and in such a case k , = l/zkd for F(*A,A) = P(*D,D). In several cases, energy-transfer processes involving transition-metal complexes exhibit rate constants lower than that expected on the basis of the Sandros equa(7) S. Glasstone, K. J. Laidler, H. Eying, “The Theory of Rate Processes”, McGraw-Hill, New York, 1941; K. J. Laidler, “The Chemical Kinetics of Excited States”, Oxford University Press, London, 1955. (8) K. Sandros, Acta Chem. Scand., 18, 2355 (1964). (9) A. A. Lamola in “Energy Transfer and Organic Photochemistry”, A. Lamola and J. J. Turro, Eds., Interscience, New York, 1969, p 17. (10) As previously noted” eq 8 was first derived by Marcus (J.Phys. Chem., 72,891 (1968)) for atom and proton transfer reactions on the basis of the BEBO model and then used by Agmon and Levine (Chem. Phys. Lett., 52, 197 (1977)) to discuss concerted reaction kinetics. We emphasized that we used eq 8 in a purely empirical way, regardless of ita theoretical derivation. (11) F. Scandola and V. Balzani, J. Am. Chem. SOC.,101,6140 (1979).
No. 14, 1981 2051
tion.1-3*5~6J2J3 In a recent paper from this laboratory,13it has been suggested that the rate-determining factor in the exothermic energy transfer from aromatic triplets to Cr(1II) complexes is a low value of the preexponential factor, k,,O, of the energy-transferrate constants, which was also shown to be related to the nephelauxetic parameter p. In this paper we report the results of the quenching of Ru(phen),(CN), phosphorescence by some Cr(II1) complexes. As for the quenching of aromatic triplets by Cr(1II) complexe~,’~J~ also in this case we have found that the rate constants are lower than those expected for exothermic energy transfer between undistorted excited states. In order to elucidate the role played by the various factors in determining the low values of the energy-transfer rate constants, we have measured not only the energy-transfer rates but also their temperature dependence. The results obtained are discussed in the light of the previous proposed treatment. Experimental Section Materials. The R ~ ( p h e n ) , ( C N )(phen ~ is 1,lOphenanthroline) was prepared by the procedure described by Demas et aLl4 for the analogous bipyridine complex. The absorption and the emission spectra and the phosphorescence lifetime were in agreement with those previously reported.15 Pure samples of t-[Cr(en)z(NCS)z]C104,16t-[Cr(en)2NCSC1]C104,’7t-[Cr(er~)~Cl~]Cl,’~ t-[Cr(en)20NOC1]C104,17 and t-[Cr(en)zF2]C11S were prepared and purified as indicated in the literature. All of the other chemicals were of reagent grade. Apparatus. Luminescence measurements were performed with a Perkin-Elmer MPF-3spectrofluorimeter, using R 446 and R 955 photomultiplier tubes. Data concerning the intensity of quenching measurements were corrected for the absorption of exciting (and emitted) light by the quenchers. Excitation and emission wavelengths were selected to minimize such corrections, which never amounted to more than a few percent. Low-temperature (77 K) emission spectra were performed by using the Perkin-Elmer phosphorescence accessory 010-0504 Dewar flask sample tube. Emission lifetimes were measured by means of an apparatus consistingof the following components: a Lambda Physik pulsed 1-MW N2 laser, a Bausch and Lomb highintensity monochromator, a R 446 photomultiplier tube, and a Tektronix R 7912 transient digitizer. The doublet-doublet absorption measurements were performed by means of an Applied Photophysics ruby laser apparatus with a pulse duration of -30 ns and an emission power of -100 mJ at 347 nm. The viscosity of the solvent was determined with an Ostwald viscosimeter calibrated with solvents of known viscosities. The reduction potentials of the ground-state complexes were determined by means of an AMEL Model 448 ap(12) F.Wilkinson and A. Farmilo, J . Chem. Soc., Faraday Trum. 2, 74, 2083 (1978). (13) V. Balzani, M. T. Indelli, M. Maestri, D. Sandrini, and F. Scandola, J . Phys. Chem., 84, 852 (1980). (14) J. N. Demas, T. F.Turner, and G . A. Crosby, Inorg. Chem., 8,674 (1969). (15) J. N. Demas, J. W. Addigton, Steven H. Peterson, and E. W. Harris, J . Phys. Chem., 81, 1039 (1977). (16) C. L. Rollinson and J . C. Bailar, Inorg. Synth., 2, 200 (1946). (17) W. W. Fee, J. N. Mac, B. Harrowfield, and W. G . Jackson, J. Chem. Soc. A , 2612 (1970). (18) G . Brauer, “Handbook of Preparative Inorganic Chemistry”,Vol. 2, 2nd ed., Academic Press, New York, 1965, p 1357. (19) J. W. Vaughn, 0.J. Stvan, Jr., and V. E. Magnuson, Inorg. Chem., 7, 736 (1968).
2052
The Journal of phvsical Chemistry, Vol. 85, No. 14, 1981
Maestri and Sandrini
TABLE I : Energy-Transfer Quenching of ('CT) Ru(phen),(CN), by Cr(en),XY+ Complexes
(1)t-Cr(en),(NCS),+ ( 2 ) t-Cr(en),NCSCl+ (3 ) t-Cr(en) ,Cl,+ ( 4 ) t-Cr(en),ONOCl+ ( 5 ) t-Cr(en),F,+
E(=Eg)," cm-
pb
13 720 14 020 14 160 14 430 1 5 180
0.71 2 0.727 0.734 0.748 0.790
AG,~ cm-'
k,,d M-' s - '
3520 3220 3080 2810 2060
5.30 X 4.13 X 3.56 X 1.92 X 0.30 X
AH*
7.30 X 5.48 X 4.64 X 2.39 X 0.36 X
10' 10' 10' 10'
lo8
,f
cm- '
keno,e s-' 10' 10' 10' 10' 10'
790 640 80
A G* ,g
cm- '
3920 3680 3560 3400 3040
Energy of the lowest excited doublet state from emission measurements. The nephelauxetic ratio, p , is calculated from the equation p = B,,, lex/Bion,taking Bion = 918 cm-I and the relationship E(2E,) = 21 Bcomplex(see ref 13). Calculated from eq 7, taking Eoo(gA,A)= E('Eg), and Eoo(*D,D)= EOCT) Ru(phen),(CN),= 17240 cm-I. Quenching constants measured at 20 "C. e Calculated from eq 5 and 8, with k , = 3.66 X l o 9 M-' s-', k-, = 4.31 X l o 9 s-' and A G * ( O ) - 0 . f Activation enthalpy values associated with the k , calculated by means of the equation AH' = E, - RT. g Free energies of activation obtained from the difference between t i e A G values expected for the observed rate constants on the basis of eq 6 and the actual A G values calculated from eq 7 (see Figure 3). The expected activation energy terms would be 1060 cm-' higher if the excited state of Ru(phen),(CN), is taken to be at 1 8 300 cm-' instead of at 17 240 cm-I.
paratus for cyclic voltammetry. Procedures. The experiments were performed in a DMF/water solution of HC1 N), 1:3 v/v. The solutions were thermostated ( 7 0.1 OC) a t the desired temperature. Deaeration was performed by the freezepump-thaw method.
Results In agreement with Demas et al.,15 Ru(phen),(CN), was found to emit phosphorescence in rigid matrix (77 K) with a maximum at 580 nm and in fluid solution with a maximum a t 620 nm. t-Cr(er~)~(NcS)~+, t-Cr(en)2NCSC1+,t - C r ( e ~ ~ ) ~ C t-l ~ + , Cr(en)20NOC1+,and t - c r ( e ~ ~ ) ~were F , + found to emit at 77 K with maxima at 729, 713, 706, 693, and 657 nm, respectively. Under our conditions only t-Cr(en),(NCS),+ 3 1I I was found to emit in fluid solution with a maximum at 729 I 3.2 3.4 3.6 nm. Absorption spectra of mixtures of Ru(phen),(CN), and Cr(II1) complexes were additive and did not change during the experiments. In fluid solution, at room temperature, the Cr(II1) complexes were found to quench the Figure 1. Plots of (0)In I vs. 11T and (0) In T vs. 1/ T for (%T) R~(phen),(CN)~ phosphorescence, and the Stern-Volmer Ru(phen),(CN), emission, in deaerated solution. quenching plots were linear a t least for the quencher concentrations used. The values of the quenching constants are reported in Table I. In the case of t-Cr(en),(NCS),+, the quenching of Ru(phen),(CN), phosphorescence was accompanied by the sensitized emission of the quencher. t-Cr(en),(NCS),+ exhibits a strong doublet-doublet .uP absorption-in the iisible region, with maximum at 540 -C nmS2O At this wavelength we have compared the (,E,) t-Cr(en)2(NCS)2+ concentrations present after the laser pulse in two different solutions having the same optical density a t the excitation wavelength. The first solution contained 7.5 X M t-Cr(en),(NCS),+ and the second one 6.0 X M Ru(phen),(CN), and 4.7 X M t-Cr18.54 I 3.40 3.50 3.60 3.10 (en),(NCS),+. In the latter solution lifetime measurements showed that 50% of (TT)Ru(phen),(CN), was quenched 1 x lo3,OK-' T by t-Cr(en)z(NCS)z+. The amount of (,E,)t-Cr(en),Figure 2. Plot of In k , vs. 1/ T for the energy-transfer quenching of (NCS)2+formed in the second solution, corrected for the (%T) Ru(phen),(CN), by t-Cr(en),(NCS),+. direct absorption of Cr(II1) and for the quenching of Ru(11), indicated that the doublet concentration was formed As shown in Figure 1, the intensity and the lifetime of with the same efficiency as in the direct excitation of tRu(phen),(CN), phosphorescence decrease in the same Cr(en)2(NCS)2+.This result shows that, even in the less way with increasing temperature in the range -5 to 50 OC. favorable case of & (Ru(phen),(CN),) = 1, the efficiency The apparent activation energy is -1.6 X lo3 cm-' mol-' of energy transfer from (3CT) Ru(phen),(CN), to (,E ) (-4.5 kcal mol-l). t-Cr(en)2(NCS)z+has to be as high as the 4Tzg 21$ The quenching constants of Ru(phen),(CN), emission efficiency of intersystem crossing for t-Cr(en),(NCS),+: by the Cr(II1) complexes are also temperature dependent which is known2' to be 0.8. (an example is shown in Figure 2). The experimental activation energies, E,, have been used to obtain the ac(20) M. Maestri and D. Sandrini, unpublished results. tivation enthalpies, AZP,reported in Table I. Under our (21) D. Sandrini, M. T. Candolfi, L. Moggi, and V. Balzani, J. Am. Chem. SOC.,100, 1463 (1978). conditions, the cyclic voltammetric experiments gave a I
I
a
I
--
I
The Journal of Physical Chemlstry, Vol. 85, No. 14, 1981 2053
Electronic Energy Transfer to Cr(II1) Complexes
reduction potential (E,) of 1.24 V vs. NHE for Ru( ~ h e n ) ~ ( C and N ) ~an irreversible reduction peak at 4.80, -1.02, and -1.22 V vs. NHE for t - c r ( e ~ ~ ) ~ ( N c S t-Cr)~+, (en)20NOC1+,and t - c r ( e ~ ~ ) ~respectively. F~+,
Discussion R u ( ~ h e n ) ~ ( C is N )known ~ to emit from a CT excited state which is formally a triplet. This excited state is likely 0 to be undistorted because it originates from a d x* metal CD to ligand charge transfer (MLCT) transition, which delocalizes the excited electron on the ligands and thus is not 8expected to change the ground-state geometry. At 77 K the maximum of the R ~ ( p h e n ) ~ ( Cemission N)~ was found to be a t 580 nm. The energy corresponding to the first 3040 crn-1; vibrational band, 17 240 cm-l, is taken as the energy of the zero-zero transition.22 7 The reduction potential of (3CT)R ~ ( p h e n ) ~ ( C Ncal)~, culated by adding to the formal potential of the ground -8 -4 -2 0 state the one electron potential corresponding to the exA G r10-3,cm-l cited-state energy,4is -0.89 V vs. NHE. The comparison with the (irreversible) reduction potentials of the Cr(II1) Figure 3. Plots of log k vs. A G for the energy-transfer quenching of (kT) Ru(phen)$CN), by &HI) complexes. The complexes are labeled complexes reported above suggests that an electronas in Table I. The curve represents the expected behavior according transfer quenching mechanism should be allowed only for to eq 6, with k, = 3.66 X lo8 h4-ls-I. t-Cr(en),(NCS),+. On the other hand, a comparison between the excited-state energy of R ~ ( p h e n ) ~ ( Cand N ) ~the in the exoergonic region can be due to an activation enenergy of the lowest excited state of Cr(II1) complexes thalpy term, A P , and/or a low value of the preexponential (Table I) shows that quenching via energy transfer is in factor, keno. On the assumption that the low values of k all cases thermodynamically allowed. That the quenching are only due to an activation enthalpy term, this shoud actually takes place via energy transfer is confrmed by the be of the order 3000-4000 cm-’ (Table I, Figure 3). The sensitized emission observed in the case of t-Cr(en)2experimental activation enthalpies, however, are much (NCS)2+,which is the only Cr(II1) complex among those smaller (80-790 cm-’, Table I). Such small values of the used in this study that exhibits luminescence in fluid soactivation enthalpies are fully consistent with the fact that lution, under our experimental conditions. A definitive the excited states involved are not distorted (see above). proof for an energy-transfer mechanism is given by the The only explanation for low k , values and for the very resulta of the laser flash spectroscopy measurementswhich small activation enthalpies remains a small value of the show that the efficiency of energy transfer from (3CT) preexponential factor caused by a poor electronic interR ~ ( p h e n ) ~ ( Cto N )(%) ~ t - C r ( e ~ ~ ) ~ ( N has c s )to ~ +be at least action (small k).25 When, as in our case, AG is large and as high as 80%. Thus, we can conclude that the predomnegative and AG* is very small, eq 5 reduces to eq 9. If inant quenching mechanism is energy transfer for the tc r ( e r ~ ) ~ ( N C S )For ~ + .the other quenchers used the enerkq = kdke>/(ka + ken’) (9) gy-transfer mechanism seems also likely as their reduction potentials are less favorable for an electron-transfer the preexponential factor, keno,is not larger than k 4 , lower quenching. than diffusion values of k , are expected. As the low values The Cr(II1) complexes have their 2E (not di~torted)~‘ of ke> have to be related to nonadiabatic reasons (k . Equation 9 in fact higher limiting values for the energy transfer to the 2E, reduces to k = k d for ke,,” >> k a and k , = ke>kd/k4 for state. keno
