58
J . Phys. Chem. 1984, 88, 58-61
Temperature Dependence of the Fluorescence Lifetimes of Linear Alkanes: A Correlation with Photodecomposition Sergio Dellonte,* Lucia Flamigni, Francesco Barigelletti, Istituto di Fotochimica e Radiazioni d’Alta Energia del C.N.R., 40126 Bologna, Italy
Laszlo Wojnarovits,’ Institute of Isotopes of the Hungarian Academy of Sciences, H-1525 Budapest, Hungary
and Giorgio Orlandi Istituto Chimico “G.Ciamician” dell’liniversitci, Bologna and Istituto di Fotochimica e Radiazioni d’Alta Energia del C.N.R.,40126 Bologna, Italy (Received: March 15, 1983)
The temperature dependence of the fluorescence lifetimes was investigated for several n-alkanes, 2,3-dimethylbutane, and 2,3-dimethylhexane. The decay is composed of a temperature-independent and of an Arrhenius-type process. The latter is identified as internal conversion and is associated with the process that ultimately leads to molecular elimination reactions (mainly of H,). The temperature-independent decay is mainly attributed to intersystem crossing and is held responsible for radical intermediate formation by rupture of C-H or C-C bonds. The relative importance of the nonactivated decay increases with increasing branching in the molecules.
Introduction The saturated hydrocarbons show characteristic photophysical namely, a low fluorescence quantum yield (afI lo-,), broad and structureless fluorescence spectra with a large Stokes shift with respect to absorption, a short radiative lifetime (- 1 ns), a dependence of af on excitation wavelength and on temperature, and furthermore a photodecomposition yield close to unity. Upon closer examination the photophysical behavior of alkanes is related to their molecular structure. This correlation was first evidenced by Lipsky who studied and compared the fluorescence properties of a large series of alkanes including linear, cyclic, and branched one^.^,^ Recently we have studied the temperature dependence of the fluorescence lifetime of a series of cyclic alkanes, by means of N2 laser two-photon e x ~ i t a t i o n . ~The , ~ aim was to get information about the nature of the radiationless decay of the emitting state SI, specifically about the relative weight of internal conversion (ic) and intersystem crossing (isc). The decay rate constant was found to be composed of a temperature-activated part, assigned to ic in view of the large value of the preexponential factor (- 1OI2 s-l), and of a temperature-independent one, which was attributed because of its rate (10s-109 s-I) to isc. This study has now been extended to linear and branched alkanes, whose emission properties are in many respects different from those of the cyclic ones. A correlation is proposed between activated (ic) and unactivated (isc) decays and the different types of mechanism of photochemical bond breaking by comparison of the relevant yields. Discussing the present results with those obtained p r e v i o ~ s l y , we ~ . ~can rationalize the differences in the photophysical and photochemical behavior and relate them to the molecular structure. Experimental Section The alkanes (Fluka) used were passed twice through a 50-cm column of freshly activated silica gel and the purity was spectroscopically checked. A11 samples were sealed under vacuum in (1) Visiting scientist under agreement between the National Research Council of Italy and the Hungarian Academy of Sciences. (2) W. Rothman, F. Hirayama, and S. Lipsky, J . Chem. Phys., 58, 1300
1-cm Suprasil fluorescence cells. The excitation was obtained from a pulsed nitrogen laser (Lambda Physik, Gottingen) with a pulse duration of ca. 3.5 ns and a pulse power of 1 MW. A convenient focussing of the laser beam in the sample cell produced two-photon excitation. The emission was detected in the proximity of ,,A with a Hamamatsu R955 photomultiplier with a five dinode chain configuration to obtain a faster time response. Data were acquired with a Tektronix R79 12 transient digitizer equipped with a 7A19 vertical amplifier and interfaced to a Z80 based Cromemco microcomputer. Lifetimes were obtained by deconvolution analysis by averaging five measurements, the scattering being 18%. Low temperatures were maintained with a liquid nitrogen flow cryostat, whereas temperatures above 273 K were obtained by flowing water through the sample holder. Other details on the lifetime measurements were described Steady-state photolysis measureinents were also made to establish the chemical decomposition paths at 298 K of some of the liquid alkanes considered in the lifetime studies. The excitation was made by a bromine lamp emitting 7.6-eV photons and the products were analyzed gas chromatographically. The details of this technique are described in ref 8 and 9.
Results In Figure 1 the reciprocal lifetimes ( l / r ) are plotted as a function of 1/ R T for the alkanes investigated: n-octane, n-nonane, n-decane, n-dodecane, n-pentadecane, 2,3-dimethylbutane, and 2,3-dimethylhexane. The low temperature limit for each compound is determined by the freszing point. As it was noted for the cyclic alkanes: the trend is understood if the decay is composed of a temperature-independent and a thermally activated part according to eq 1. In Figure 1 the solid curves were calculated 1 / r = ko
-+ A exp(-E,/RT)
(1)
by the ko, A , and E, parameters (collected in Table I) obtained by a nonlinear iterative least-squares fitting procedure.
(1973). (3) S . Lipsky, “Chemical Spectroscopy and Photochemistry in the Vacuum-Ultraviolet”, C. Sandorfy, P. J. Ausloos, and M. B. Robin, E d , Reidel, Boston, 1974, p 495. (4) L. Flamigni, F. Barigelletti, S . Dellonte, and G. Orlandi, Chem. Phys. Left.,89, 13 (1982). (5) G. Orlandi, L. Flamigni, F. Barigelletti, and S. Dellonte, Radiat. Phys. Chem., 21, 113 (1983).
0022-3654/84/2088-0058$01.50/0
(6) F. Barigelletti, S.Dellonte, G. Mancini, and G. Orlandi, Chem. Phys. Lett., 65, 176 (1979). (7) G. Orlandi, S. Dellonte, L. Flamigni, and F. Barigelletti, J. Chem. Soc., Faraday Trans. 1, 78, 1465 (1982). ( 8 ) 1.Wojnarovits and G . Foldilk, Z f l M i f r . , 33a, 243 (1981). (9) 1.Wojnarovits, L. Kozlri, CS. Keszei, and G. Foldiik, J . Photochem.,
19, 79 (1982).
0 1984 American Chemical Society
The Journal of Physical Chemistr,y. Vol. 88, No. 1. 1984 59
Fluorescence Lifetimes of Linear Alkanes
TABLE I: Temperature-Independent Term h,, Activation Energy Ea, and Preexponential Factor A Obtained by Using Eq '1 k , , sW1
alkane
n-octane 2.9 x l o a 1.1x 10" 3.3 1.35 2.5 1.85 n-nonane 1.7 X l o a 2 . 2 x 10" 3.8 1.93 3.3 1.71 n-decane 1.5 X l o * 5.3 x 10" 4.4 2.37 4.2 1.77 7.6 x 107 6.2 X 10'' 4.6 3.65 n-dodecane 5.5 1.51 n-pentadecane 8.0 x 107 1.0 x 10lZ 5.2 4.80 7.3 1.52 2,3-dimethylbutane 5.2 X l o * 2.0 x 10" 3.6 0.99 6.1 6.2 2,3-dimethylhexane 6.1 X l o a 3.8 0.63 4.8 7.6 6.0 X 10" ' The fluorescence lifetimes and intensities are also reported. The accuracy of h , is about 2 15%, the estimated error in Ea is i 0 . 5 kcal/mol, and that of log A is k0.5. Taken from ref 2, excitation wavelength 1 6 5 nm.
15
10
-
-
TABLE 11: Quantum Yield of Activated and Nonactivated Decay and Quantum Yield of Radical and Molecular Decomposition in the Photolysis of Alkanes at 298 K
i
alkanes n-alkanes 2,3-dimethylbutane 2,3-dimethylhexane isopropylcyclohexane cis-1,3-dimethylcyclohexane trans-1,4-dimethylcyclohexane methylcyclohexane cis-decalin trans-decalin
C,-C,,
5'
01 1
7 6 I
3
2
4
I
l/RT, mole kcal-' Figure 1. Fluorescence rate parameters of alkanes as a function of temperature. The experimental points are shown together with the fitted curves: (1) 2,3-dimethylhexane;(2) 2,3-dimethylbutane;(3) n-octane; (4) n-nonane; ( 5 ) n-decane; ( 6 ) n-dodecane; ( 7 ) n-pentadecane.
The values of E,, 70 = 1/ k,, and 7298 of the n-alkanes slightly increase as the number of carbon atoms ( N c ) increases. Since there is also a systematic increase in the @f,298 values,2 the ratio @f,z98 K / ~ z 9 8K, corresponding to the radiative rate constant, is approximately the same for all the n-alkanes, that is, 1.7 X lo6 s-l as shown in Table I. This constancy, already reported in the literature,2J0Jiallows one to estimate the lifetimes for those n-alkanes for which only the fluorescence intensity, measured close to the absorption onset, is known. With regard to the fluorescence characteristics, 2,3-dimethylbutane and 2,3-dimethylhexane are off the line of the n-alkanes. Here the fluorescence yields are as large as for longer n-alkanes, but the lifetimes are lower resulting in a @f,298 K/7298 K value higher by a factor of -4. The room temperature lifetimes we have measured agree within *7% with the lifetimes determined by Henry and Helman,', Ware and Lyke," and Katsumura et al.lz There are no published data of directly measured lifetimes of the 2,3-dimethylalkanes, although Choi and Lipsky,13according to their quenching experiments with 0.43 ns for 2,3-dimethylperfluorodecalin, predicted 7298 butane. Although it is less than half the value we measured, considering the accuracy of lifetime calculation from the quenching data, the disagreement does not seem dramatic. For a few of the compounds studied previously45and considered in the following discussion (isopropylcyclohexane, cis- and trans-decalin), the quantum yields of the main types of photochemical decomposition, which were not available in the literature,
-
-
(10) M. S. Henry and W. P. Helman, J. Chem. Phys., 56, 5734 (1972). (11) W. R. Ware and R. L. Lyke, Chem. Phys. Lett., 24, 195 (1974). (12) Y. Katsumura, Y. Yoshida, S. Tagawa, and Y. Tabata, Rudiut. Phys. Chem., 21, 103 (1983). (13) H. T. Choi and S. Lipsky, J . Phys. Chem., 85, 4089 (1981).
*nonact
@act
@radical
@rnolecul
0.3 0.55 0.40 0.55a 0.40'
0.7 0.45 0.60 0.45 0.60
0.V d d
0.8
0.60 0.45e
0.40 0.55
0.35a
0.65
0.50e
0.50
0.30'
0.70
0.30e 0.70 0.7-0.8 0.3-0.2 la,b 0.4-0.6 0.6-0.4 a Reference 4. Practically no activated process in these alkanes. References 8, 16, and 17. There are not photolytic measurements, but from radiolytic dataz3 and also from photolytic measurements with other highly branched alkanes,lh*22 @radical 0.3-0.6. e Reference 9. la$
-
have been measured. In isopropylcyclohexaneat 298 K, the 7.6-eV photolysis mostly takes place through the rupture of the C-C bond between the two tertiary carbon atoms, yielding unimolecular elimination products (cyclohexene and propane or cyclohexane and propene) with @ 0.2 or radical intermediates (cyclohexyl and propyl) with @ 0.5. The other decomposition ways involve atomic and molecular elimination of hydrogen. The overall quantum yield of the radical-forming reaction was estimated to be 0.6, considering that the remaining @(H2+H),which is -0.3, is mainly due to H, production and therefore @(H) I0.1. The photochemistry of cis- and trans-decalin is very complicated and also, the two isomers exhibit different reactivities according to the present data. As in other similar compounds, the primary processes involve Hz and H elimination (+(H2) > +(H) and @(H2+H) 0.3 in cis-decalin and 0.6 in the trans isomer) and C-C bond rupture, preferentially occurring between the two tertiary carbon atoms. The first step of the latter reaction is expected to be biradical formation,14 probably in the triplet state. Thus, the overall quantum yield of radical formation in the primary photoreaction is estimated to be 0.8 in cis- and 0.5 in trans-decalin. These data are reported in Table I1 together with those pertinent to the other compounds and found in the literature.
--
-
Discussion Photophysical Processes. In the preceding section it has been shown that the decay of the S1excited singlet state, in the linear alkanes investigated, is composed of a temperature-independent term of the order of lo8s-' and of a thermally activated component with a frequency factor of the order of 101'-10'2 s-I and an activation energy of 3-6 kcal/mol. Since the radiative process contributes negligibly to the SI decay (af