J. Phys. Chem. 1984,88, 1976-1981
1976
observations. From the fit we get an overall decay rate of about 50 ns, in agreement with similar results on neat naphthalene M BMN) by Braun et al.zo Since a l crystals (doped with N 0.5 and y is of the order unity:' the same holds for the product a l y ;thus, the hopping time in the plane turns out to be of the order of 1 ps. The crystal examined was doped with minute amounts of impurities and we paid special attention to the characteristics of the excitation decay. The theoretical extension to ternary crystals, and thus a study of the total concentration range of the ratio CIOHB/CIODB, calls for additional methods (analytical and simulation procedures) beyond the scope of the present paper (see also ref 6 for interesting results). We also cannot exclude the (20) A. Braun, H. Pfisterer, and D. Schmid, J . Lumin., 17,
presence of coherence (such as direct trap excitation from the bottom of the k = 0 band22),which was inferred by evidence from measurements of heavily mixed crystals; from the present work we find, however, that an anisotropic three-dimensional randomwalk model offers a good explanation of the observed decay behavior of lightly doped crystals without invoking coherence.
Acknowledgment. P.A. is thankful for the support by NATO grant SA5205RG295/82. The support of the Deutsche Forschungsgemeinschaft and of the Fonds der Chemischen Industrie is acknowledged by A.B. This research was supported by N I H Grant N O R 01 N S 08116-14 to R.K. The computations were carried out at the computer centers of the University of Crete, of ETH Zurich, and of the University of Michigan. Registry No. Naphthalene, 91-20-3.
15 (1978).
(21) P. Argyrakis, D. Hooper, and R. Kopelman, J . Phys. Chem., 87,1467 (22) J. Hoshen and J. Jortner, J . Chem. Phys., 56, 5550 (1972).
(1983).
Picosecond Laser Spectroscopy of Dual Excited Electronic States of 4-( 9-Anthry1)-N,N-dimethylaniline Tadashi Okada,* Masafumi Kawai, Takaaki Ikemachi, Noboru Mataga,* Department of Chemistry, Faculty of Engineering Science, Osaka University, Toyonaka, Osaka 560, Japan
Yoshiteru Sakata, Soichi Misumi, The Institute of Scientific and Industrial Research, Osaka University, Suita, Osaka 565, Japan
and Shigeo Shionoya The Institute for Solid State Physics, The University of Tokyo, Roppongi, Minato, Tokyo 106, Japan (Received: September 16, 1983)
-
Quantitative measurements of S, SI absorption spectra of 4-(9-anthryl)-N,N-dimethylaniline (ADMA) in various solvents have been carried out by means of picosecond laser spectroscopy at room temperature. The results have been approximately explained by assuming the existence of closely lying dual excited states with different electronic structure, and the equilibrium constant between them has been obtained as a function of solvent polarity. Dipole moments of these two states have been estimated to be 12.5and 20.5 D, respectively. In low-viscosity polar solvents, the equilibrium between the two states is attained immediately after picosecond pulsed (-30 ps) excitation. In 2-propanol solution, however, both a rapid process and a much slower process of the interconversion between the two states have been observed, the mechanisms of which have been examined. Importance of the wide-band picosecond time-resolved absorption spectroscopy for investigating the change of electronic structure in the excited state, of which the studies by means of time-resolved fluorescence spectroscopy is rather difficult, has been demonstrated clearly.
Introduction The dual fluorescence phenomena observed for compounds with electron donor and acceptor groups separated by a single bond have been a subject of lively investigations by means of stationary as well as time-resolved measurements. Especially intensive investigations including picosecond time-resolved fluorescence measurements have been made for p-(dimethy1amino)benzonitrile (DMABN),'-*while 4-(9-anthryl)-N,N-dimethylaniline(ADMA) is also a typical compound showing a similar phenomenon and has been investigated to elucidate the relevant mechanisms. 1,3-8 (1) Grabowski, Z. R.; Rotkiewicz, K.; Siemiarczuk, A.; Cowley, D. I.; Baumann, W. Nouu. J . Chim. 1979, 3, 443 and references cited therein. (2) (a) Huppert, D.; Rand, S. D.; Rentzepis, P. M.; Barbara, P. E.; Struve,
W. S.; Grabowski, 2.R. J. Chem. Phys. 1981, 75, 5714. (b) Wang, Y.; Eisenthal, K. B. Ibid. 1982, 77, 6076. (3) Okada, T.; Fujita, T.; Kubota, M.; Masaki, S.; Mataga, N.; Ide, R.; Sakata, Y . ; Misumi, S. Chem. Phys. L e f f .1972, 14, 563. (4) Chandross, E. A. "The Exciplex"; Gordon, M., Ware, W. R., Eds.; Academic Press: New York, 1975; p 187.
0022-3654/84/2088-1976$01.50/0
The dual fluorescence spectra of DMABN are considered to be emitted from a strongly polar and a less polar excited state respectively, the former state emitting at a longer wavelength region in polar solvents. According to Grabowski et al., the a-systems of donor and acceptor groups are perpendicular to each other in the excited state with higher polarity! On the other hand, an alternative model based on the electrooptical emission and absorption measurements on ADMA has been proposed by Baumann et a1.' According to their results, the spectra of ADMA can be understood in terms of a solvent-induced polarizability effect without assuming any conformational change in the excited state even in polar solvents. (5) Okada, T.; Fujita, T.; Mataga, N. Z . Phys. Chem. (Wiesbaden) 1976, 101, 1 5 . (6) Siemiarczuk, A,; Grabowski, Z. R.; Krawczynski, A,; Asher, M.; Ottolenghi, M. Chem. Phys. Lett. 1977, 51, 315. (7) Baumann, W.; Petzke, F.; Loosen, K. D. Z . Naturforsch., A 1979, 34A,
1070. (8) Siemiarczuk, A.; Koput, J.; Pohrille, A. Z . Naturforsch. A 1982, 37A, 598.
0 1984 American Chemical Society
Picosecond Laser Spectroscopy of ADMA
The Journal of Physical Chemistry, Vol. 88,No. IO, 1984 1977
In the following, we demonstrate the solvent-induced change of the electronic structure in the excited ADMA on the basis of the results obtained by quantitative analysis of picosecond time-resolved transient absorption spectra in polar solvents.
Experimental Section Methods and Materials. Time-resolved absorbance spectra in the picosecond time region were measured with a passively mode-locked ruby laser system. The details of picosecond laser spectrometer were described in a previous paper.g In order to obtain quantitative values of the relative absorption intensity in the excited state of ADMA in various solvents, measurements were carried out repeatedly with great care under the condition that the absorbance at exciting wavelength (347 nm) in every solution was adjusted to the same value (usually absorbance 1.2, and [ADMA] = 2.4-3.1 X M depending upon the nature of solvent). The transient absorbance in each solution was measured as a function of the intensity of the exciting laser pulse. The obtained relationships were used to calibrate the observed absorbance of the excited state of ADMA. A mode-locked YAG laser system was used for the measurement of time-resolved fluorescence spectra. The fluorescence of ADMA excited by a third harmonic (355 nm) was sampled with a CS2 optical Kerr shutter and detected by an image orthicon television camera connected to an image intensifier. The details of the system for this time-resolved fluorescence measurement were described elsewhere.I0 The preparation of ADMA was reported elsewhere." Spectrograde ethyl acetate, pyridine, and acetonitrile were used without further purification. Spectrograde hexane, ethyl ether, isopentyl acetate, and 2-propanol were distilled before use. GR grade isobutyl acetate and butyl formate and chromatographic reagent methyl acetate were carefully distilled before use. Quartz cells with 1-cm optical path were used for the measurements. All solutions were deaerated by freeze-pump-thaw cycles. Measurements were carried out at room temperature (26 f 1 "C). Absorption and fluorescence spectra were carefully measured before and after laser excitation, and it was confirmed for all solutions that no detectable photochemical decomposition occurred. Absorption Spectra of the Excited State of ADMA. It is well-known that absorption spectra of excited molecules are more or less distorted by the induced fluorescence emission caused by the probing light pulse in general. When the fluorescence transition probability is small, an accurate absorption spectrum may be obtained by subtracting from the observed spectrum the fluorescence intensity obtained by excitation without probe pulse. However, the probe light would be amplified significantly by the induced fluorescence emission when the transition probability is not so small. In the case of ADMA, its laser action has been found in nonpolar and slightly polar sdvents.12 In accordance with this result, we have confirmed that ADMA shows a strong induced fluorescence emission in some solvents, when excited by picosecond laser for the transient absorption measurement. The effect of the induced fluorescence upon the observed transient absorption may be given by the following equations. The increase dZ of the light intensity Z incident the medium of width dx containing excited molecules of concentration c can be written as
w
0
z
-
dZ = AC dx
+ BZC dx - ~ Z Cdx
(1)
where A and B are Einstein's coefficients and t is the extinction (9) Okada, T.; Migita, M.; Mataga, N.; Sakata, Y.; Misumi, S. J . Am. Chem. SOC.1981, 103, 4715. (10) Saito, H.; Kuribayashi, S.; Shionoya, S . Jpn. J . Appl. Phys. 1976,15, 947. Tanaka, Y.; Masumoto, Y.; Tanaka, S.; Shionoya, S. "Picosecond Phenomena"; Shank, C. V., Shapiro, S. L., Eds.; Springer-Verlag: West Berlin, 1978, p 256. (1 1) Migita, M.; Okada, T.; Mataga, N.; Sakata, Y.; Misumi, S.;Nakashima, N.; Yoshihara, K. Bull. Chem. SOC.Jpn. 1981, 54, 3304. (12) Nakashima, N.; Mataga, N.; Yamanaka, C.; Ide, R.; Misumi, S . Chem. Phys. Lett. 1973,18, 386.
4:
m
P
0 v)
m 4
c 4
5
500
I
I
600
700
I
800 A,",.
Figure 1. Transient absorbance spectra of ADMA in various solvents at 200 ps. Solvent: (1) hexane (e = 1.9), (2) ethyl ether (e = 4.3), (3) isobutyl acetate (e = 5.3), (4) ethyl acetate ( E = 6.0), (5) methyl acetate ( E = 6.7), (6) acetonitrile (e = 37.5).
-
coefficient of S, SI absorption. The total light intensity measured by an integrating detector like photodiode array may be written as follows when uniform distribution of the excited molecules is assumed.
Z = Zo exp(B - E)CX
A + --(exp(B B-€
- t)cx - 1)
(2)
The second term on the right-hand side of eq 2 represents the emission intensity in the absence of probe light (Zo = 0). Therefore, by subtracting this emission intensity from the observed transient absorption, one can obtain only the (B - t) spectrum instead of SI absorption spectrum, in the wavelength region where the S, the strong fluorescence is observed. As shown in Figure 1, a negative absorbance spectrum was observed around 450-500 nm in hexane as well as in ethyl ether solution. The value of B seems to be larger than the value of c in the fluorescent state of ADMA in this wavelength region where, owing to the amplification of the probe light pulse, it is difficult to obtain correct transient absorption spectra. Because ADMA shows fluorescence around 400-550 nm in nonpolar solvents and around 500-700 nm in such strongly polar solvents as acetonitrile, the transient absorption spectra measured in the region 720-870 nm were used for the estimation of the relative concentrations of the excited state.
-
Results and Discussion Transient Absorption Spectra of ADMA in Various Solvents. Absorption spectra of the excited state of ADMA measured at 200 ps after excitation are shown in Figure 1. Since no change of the spectral shape has been observed at delay times longer than few tens of picoseconds in all solutions used here except 2-propanol, the reorientational relaxation of polar solvent molecules around excited ADMA, which is responsible for the formation of the excited state with higher dipole, seems to reach their thermal equilibrium quickly. Actually, we have confirmed already by means of a picosecond streak camera that, in acetone and acetonitrile solution of ADMA, the rise time of the long-wavelength
1978 The Journal of Physical Chemistry, Vol. 88, No. 10, 1984
Okada et al.
4-
3-
2-
Y
1-
-I I
.
I
.
500
I
I
I
.
.
I
.
0-
.
800
700
600
Figure 2. Transient absorption spectra of 9-PA in hexane at 200 ps (A) and 9-PA-DMT in acetonitrile at 2 ns (B) ([9-PA] = 2 X lo4 M, [DMT] = 2 X 10-1M).
fluorescence is shorter than 20 ps.IL In the case of 2-propanol solution, time-resolved studies have been made as it is shown in the latter part of this paper. In Figure 1 as well as other figures for transient absorption spectra, no data are shown around 700 nm. This is due to the use of a filter to cut the scattered light pulse of the fundamental. The S, SI absorption spectrum of 9-phenylanthracene (9-PA) in hexane and transient absorption spectra obtained by laser photolysis of a ternary solution of 9-PA and N,N-dimethyl-ptoluidine (DMT) in acetonitrile are shown in Figure 2. It has been well established that the laser photolysis of aromatic hydrocarbon-N,N-dialkyl aromatic amine system in acetonitrile solution results in a rather efficient production of hydrocarbon anions and amine cation radicals which vanish by recombination in the hundred nanosecond to microsecond time regions.I3-l6 Actually, the absorption band around 450-500 nm is very similar to that in the spectrum of N,N-dialkylaniline cation radical,I7 and those in the wavelength region longer than 550 nm are similar to those in the spectra of anthracene anion radicals.ls Therefore, the absorption spectra in Figure 2A,B may be used as references to compare with transient spectra of ADMA in nonpolar and polar solvents, respectively. The absorption peak of excited ADMA in hexane is blue-shifted, and its spectral shape in longer wavelength region is somewhat different compared to those of 9-PA in hexane solution, respectively, which may be due to the charge-transfer interaction between anthracene and dimethylaniline moieties. In strongly polar solvents, transient absorption spectra are rather similar to that of Figure 2B, indicating the formation of an ion-pair-like state due to very strong charge transfer. The absorption spectra in slightly polar solvents seem to change gradually to the ionic spectra with increase of solvent polarity. Kinetics in the Excited State ofADMA. The reaction scheme for ADMA in the excited state may be written as
-
-1
I
I
0.35
Q40
Figure 3. Plot of In K vs. f,. Solvent: (1) ethyl ether, (2) isopentyl acetate, (3) isobutyl acetate, (4) ethyl acetate, ( 5 ) butyl formate, ( 6 ) methyl acetate, (7) pyridine, (8) 2-propanol.
equilibrium constant K between X and Y when the condition, k l , k2 >> k3 k4, IC5 k6, holds may be related to the Gibbs free energy difference AG between two states in solution.
+
+
RT In K = -AG
AG = AG,,,
hv
A
hv'
where X and Y represent the two excited states of ADMA and Y denotes the state with higher dipole moment. A is the ground state, and kl-k5 represent rate constants of various processes. The ~
(13) Taniguchi, Y.;Nishina, Y.; Mataga, N. Bull. Chem. SOC.Jpn. 1972, 45, 764.
(14)Taniguchi, Y.;Mataga, N. Chem. Phys. Lett. 1972, 13, 596. (15) Masuhara, H.; Hino, T.; Mataga, N. J . Phys. Chem. 1975, 79,994. (16) Hino, T.; Akazawa, H.; Masuhara, H.; Mataga, N. J . Phys. Chem. 1976, 80, 33. (17)Shida, T.;Hamill, W. H. J. Chem. Phys. 1966, 44, 2369. (18) Shida, T.; Iwata, S. J . Am. Chem. SOC.1973, 95, 3473.
+ AG,o~v
(4)
where AG,,, denotes the free energy difference in the gas hase, AGsolvis the difference of the free energy due to solvatiod, and px and H~ are the dipole moments of X and Y, respectivkly. a is the cavity radius in the reaction field theory of Onsager, and f, is the polarity parameter of the solvent with dielectric constant 6 . From eq 4 and 5, we obtain
P
The equilibrium constant K was estimated by using the results of quantitative measurements of the transient absorbance spectra. For the analysis of the spectra, it has been assumed that the absorbance spectra in hexane and in acetonitrile can be assigned to X and Y states, respectively, as a first-order approximation. This is in accordance with the fact that the observed absorbance spectrum in any solvent with intermediate polarity can be reproduced approximately by the superposition of the absorbance spectrum in hexane and that in acetonitrile. For the actual determination of the K values in various solvents, the following procedure was adopted. The absorbance spectrum in a solvent s (A,(h))is represented by
A,(V = a,Ax(V A
I
0.45 f t
+ b,Ay(M
(7)
where Ax(X) and AY(X) are the absorbance spectra of X and Y states and a, and b, are their fractions, respectively. The equilibrium constant in this solvent is given by using these fractions as
K = b,/a,
(8)
Ax(A) and Ay(X), which can reproduce the observed spectra in all solutions most satisfactorily, were determined by a self-consistent procedure, and a, and b, were obtained by the least-squares method. In Figure 3, In K is plotted as a function off,. A linear relation between In K andf, was obtained for all solutions examined here except 2-propanol and pyridine solutions. From this linear relation, (wY2 - px2) was determined to be 253 D2 taking a = 5 A and AGp
The Journal of Physical Chemistry, Vol. 88, No. 10, 1984 1979
Picosecond Laser Spectroscopy of ADMA
'FA
22L
t-
tT
Figure 4. Solvent effects on the fluorescence band maxima of ADMA. Observed values (0)are taken from ref 5 . See text.
0DS
I
I
I
I
I
I
2oops
d
Figure 5. Time-resolved absorbance spectra of ADMA in 2-propanol at room temperature.
was estimated to be 3800 cm-'. Using these values of (hy2- px2) and AGgas,we can estimate the equilibrium constants in hexane in hexane and and acetonitrile. K was obtained to be 1.5 X 5.5 X IOz in acetonitrile, which are consistent with the assumption that X and Y states are predominant in hexane and acetonitrile, respectively. In the case of 2-propanol solution, the reorientational relaxation process of the solvent does not seem to have finished at the delay time of 200 ps after excitation, owing to its slow bulk relaxation. The effective dielectric constant of 2-propanol at 200 ps can be estimated from the K value in Figure 3 to be about 5.5, which is very close to the dielectric constant, 4.35, measured at the wavelength X = 10.1 cm (about a 3 GHz wave) in l - p r ~ a n o l . ' ~ Under the condition that the rapid equilibrium between two excited states is established before their decay, the decay functions of X and Y may be given according to eq 1 as follows.
(9)
where X=
kx + Kky 1+K
kx = k3
+ kq, k y = k5 + k6
(11)
On the basis of these decay functions, the fluorescence quantum yield and the radiative rate constant are given by eq 12 and 13, respectively.
(19) Garg, S. K.; Smyth, C . P.J . Phys. Chem. 1965, 69, 1294.
+
k4 k6K = ___ 1+K
If we assume that the radiative rate constants k, and k6 are independent upon solvent property, the optimum values of k4 and k, can be estimated by means of eq 13, using observed quantum yield^,^ lifetime,5and K values estimated in this work. Estimated values were k4 i= 1.7 X IO8 s-l and k6 i= 2 X IO7 s-I. The apparent radiative transition probability kf = (PFX determined in the previous work5 can be recalculated by using k4, k6, and K to be kf = 1.7 X lo8 s-' in hexane, 1.2 X IO8 s-l in diethyl ether, and 7.3 X lo7 s-l in isobutyl acetate. These values are rather close respectively to the apparent kfvalues 1.43 X lo8 s-l, 1.5 X lo8 s-l, and 6.9 X lo7 s-I, which indicates that the above assumption of solvent-independent k4 and k6 values holds approximately. The above results suggest that the fluorescence spectra of ADMA in nonpolar and slightly polar solvents may be assigned mainly to the X state. Therefore, we can estimate the dipole moment of ADMA in the X state by using the solvent effects upon fluorescence band maxima in nonpolar and slightly polar solvents indicated by the solid line in Figure 4. Taking the cavity radius as 5 A, we have evaluated hx to be 12.5 D. Using gy2- px2 = 263 DZobtained from the slope of Figure 3, we have estimated the dipole moment in the Y state to be hy = 20.5 D. The broken line in Figure 4 represents the solvent shift of Y state fluorescence calculated by assuming pY = 20.5 D and a = 5 A. These values of the dipole moment are 10-20% larger than the values determined in heptane as well as dioxane solutions from electrooptical emission measurements by Baumann et aL7 Time-Resolved Absorption and FluorescenceSpectra of ADMA in 2-Propanol. The equilibrium between X and Y states of ADMA is established very rapidly in such polar solvents as indicated, for example, in Figure 1. However, in polar solvents with high viscosity the change of the absorption spectra indicating the processes toward the equilibrium between X and Y states may be detected. Actually, it has been observed that the transient absorption spectrum of excited ADMA in 2-propanol shows drastic change in the time range of about 100 ps after excitation. The absorption spectra at 0 and 200 ps are compared in Figure 5 . These spectra can be reproduced satisfactorily by superimposing the spectra of X and Y states with use of eq 5 in the same manner as in the analysis of solvent dependence of the transient absorption at 200 ps. The absorption spectra observed at 0 ps, Ao, and 200 ps, AZo0,have been given by A0
2
0.47Ax
+ 0.53Ay
A200 2 0.28Ax
+ 0.72Ay
(14)
It should be noted that about 50% of the excited ADMA molecules are already in the Y state immediately after excitation. The absorption spectrum of the ground-state ADMA shows broadening and trailing in the low-wavenumber side with increasing solvent polarity. This result has been interpreted as due to the existence of some conformational isomers formed by a weak intramolecular charge-transfer interaction between two chromophores induced by the interaction with polar solvent m o l e c ~ l e s . ~ The dipole moment of the ground-state ADMA determined by the electrooptical absorption measurement exciting at the lowwavenumber absorption shoulder in benzotrifluoride ( e = 9.2) was C m (about 5 D).7 The fact reported to be (16.2 f 6.5) X that the excited ADMA does not interact with N,N-dimethylaniline even in case it is used as solvent seems to suggest strongly the existence of the charge-transfer interaction between two chromophores both in the ground state and in the excited F-C state. The rapid formation of the Y state within the resolution time of our picosecond laser photolysis system may be explained as follows. Some fraction of the ADMA molecules interacting strongly with 2-propanol seem to have weak intramolecular charge-transfer character in the ground state, and its electronic structure will become almost the same as that of the Y state immediately after the excitation (within a picosecond) due to such a small change of solvent orientation as rotation of the OH group. It may be true that the most stable geometrical structure of the Y state is the one twisted around the single bond between the
Okada et al.
1980 The Journal of Physical Chemistry, Vol. 88, No. 10, 1984 3
-
2
r
m
0
2 1
I
400
I
500
I
I
600 nm
Figure 6. Time-resolved fluorescence spectra of ADMA in 2-propanol. Delay time from the excited pulse: (1) 7 ps, (2) 126 ps, (3) 273 ps.
anthryl and anilino groups.' However, it may not be necessary to take a fully twisted structure for showing the ionic absorption spectra assigned to the Y state in the above discussion. Concerning the specific interaction in the ground state between ADMA and dipolar molecules, which seems to be responsible for the very rapid appearance of Y state absorption spectra, the fluorescence spectra as well as fluorescence decay times were measured previously in mixed solvents of methylcyclohexane and N,N-dimethylformamide (DMF).",S We have confirmed the wavelength dependence of the fluorescence decay time and suggested the existence of various kinds of ADMA-DMF associates formed in the ground staten5 Time-resolved fluorescence spectra of ADMA in 2-propanol measured at room temperature are shown in Figure 6. The peak at 530 nm detected at 7 ps after excitation is due to the scattered light of the second harmonic of the YAG laser, and its bandwidth shows the spectral resolution of the detection system. The spectra measured at various delay times do not show any structure which indicates two emission bands from dual excited states, but seems to show the time-dependent red shift of a broad charge-transfer fluorescence due to a simple solvation process. In Figure 7 , normalized fluorescence intensities observed at various wavelengths are plotted as a function of time. The fluorescence intensity at 442 nm rises within the resolution time of the apparatus and decreases with the decay time of several tens of picoseconds. The rise time as well as the decay time of the fluorescence increases with increase of the wavelength where the observation is made, and at 560 nm the rise time of about 110 ps was obtained as shown in Figure 7 . Similar results were obtained previously on the same system by using the picosecond streak camera." Although structureless spectra were observed and the monotonous relationship between the rise time and the wavelength of observation was obtained by the measurement of time-resolved fluorescence spectra, the observed spectra should be explained as composed of the fluorescence bands emitted mainly from the states giving S I spectra due to less polar states (X) or those either the S, due to strongly polar states (Y). The existence of the "rapid" process in ADMA is somewhat different from the charge-separation processes in the case of 9,9'-bianthryl"sZ0 as well as the intramolecular heteroexcimer systems of the type P - ( C H ~ ) ~ N C ~ H ~ - ( C H 1-pyrenyl) ~),-( and -(9-anthryl) (n = 1 and 2) in polar In these systems, there is no charge separation induced directly by exci-
-
(20) Nakashima, N.; Murakawa, M.; Mataga, N. Bull. Chem. SOC.Jpn. 1976, 49, 854. (21) Hirata, Y.;Kanda, Y.; Mataga, N. J . Phys. Chem. 1983, 87, 1659. (22) Masaki, S.;Okada, T.; Mataga, N.; Sakata, Y.; Misumi, S. Bull. Chem. SOC.Jpn. 1976, 49, 1211. (23) Mataga, N.; Okada, T.; Masuhara, H.; Nakashima, N.; Sakata, Y.; Misumi, S. J Lumin. 1976, 12/13, 159. (24) Okada, T.; Saito, T.; Mataga, N.; Sakata, Y.;Misumi, S.Bull. Chem. SOC.Jpn. 1977, 50, 33 1.
1
,
\
I
-50
I
I
0
50
I 100
I
I
150 t l p s
Figure 7. Time dependences of fluorescence intensities detected at various wavelengths: (1) 442 nm, (2) 465 nm,(3) 515 nm, (4) 560 nm, (5) response function of the apparatus.
tation, but it will be initiated by thermal reorientational fluctuations of polar solvent molecules as well as solute chromophore groups. In the case of 9,9'-bianthryl, the "broken symmetrywz0 in polar solvents occurs only in the excited state because the molecule has no dipole moment both in the ground state and X state (locally excited state of anthracene moiety). Therefore, no rapid formation of a charge-transfer state is observed and the decay time of the fluorescence from the locally excited state and the rise time of the charge-transfer fluorescence are approximately the same (ca. 80 ps) in 2-propanol solution." Analogous results were obtained also in the case of the above intramolecular heteroexcimer systems with n = 1 and 2." The results in the present work have been explained assuming a "two-staten model. However, there is a possibility of the existence of "multiple statesn of excited ADMA with different degrees of solvation and twisting and, accordingly, with a different degree of charge transfer. There would be rapid interconversions between these states, and a statistical average of them would be observed. For example, in the case of 2-propanol solution, it was necessary to assume the existence of the charge-transfer state which can be observed immediately after excitation and shows an absorption spectrum similar to that of the Y state, although its solvation and intramolecular twisting degrees may be different from those in the relaxed Y state. This result suggests that various such states might exist. It should be pointed out here that the existence of various kinds of intermolecular heteroexcimers and solvated ion pairs under fairly rapid interconversion between each other in polar solvents has been proposed on the basis of the results of picosecond laser-induced photocurrent rise time and heteroexcimer fluorescence decay time measurements.21 This circumstance seems to be somewhat analogous to the present case of the excited ADMA system. The time-resolved fluorescence behavior of the present ADMA system does not seem to contradict such an assumption of the existence of many states interconverting with each other. Nevertheless, the most important electronic structures in the excited state of ADMA seem to be represented approximately by those of X and Y states showing respective absorption spectra. This result demonstrates clearly the importance of the wide-band picosecond transient absorption spectroscopy for the elucidation of the solvent-induced electronic structure change in the excited composite molecules. Concluding Remarks The studies of transient absorption spectra of excited ADMA in various solvents provide important information about the mechanisms of dynamical behavior in its fluorescent state. The results described in the present paper may be summarized as
J. Phys. Chem. 1984, 88, 1981-1987 follows. (1) The equilibrium constant between the two kinds of excited states in thermal equilibrium and their dipole moments in various solvents have been determined by means of picosecond quantitative transient spectroscopy. (2) Rapid and slow formations of the Y state have been observed in 2-propanol solution. The rapid formation of the Y state may be caused by the excitation of ADMA molecules with weak intramolecular charge-transfer character in the ground state due to the interaction with 2propanol. A slight reorientation of the solvent in close contact with ADMA will be sufficient, for its electronic structure becomes practically the same as that of the Y state and will be followed by a more extensive reorganization of the surrounding solvent molecules as well as a geometrical structural change of ADMA. (3) In the slow formation process in 2-propanol solution, any specific strong interaction between ADMA and polar solvent molecules may not exist at the moment of excitation, and the reorientation of solvent molecules and the twisting motion of
1981
excited ADMA will be coupled with each other, leading to the extensive charge separation. Although the Y state formation mechanism in other polar solvents examined in this work may be analogous to the case of 2-propanol solution, not only the “rapid” process but also the “slow” process in those solvents may be much faster than in 2-propanol, owing to their low viscosity. (4) The importance of the wide-band picosecond transient absorption spectroscopy for the elucidation of the solvent-induced electronic structure change in the excited composite molecules has been clearly demonstrated.
Acknowledgment. N.M. and T.O. acknowledge the support by the Toray Science Foundation and Mitsubishi Foundation, and N.M. also acknowledges the support by a Grant-in-Aid for Special Project Research on Photobiology from the Japanese Ministry of Education, Science, and Culture. Registry No. ADMA, 38474-09-8.
Theoretical Studies of the Gas-Phase Proton Affinities of Molecules Containing Phosphorus-Carbon Multiple Bonds Lawrence L. Lohr,* Department of Chemistry, University of Michigan, Ann Arbor, Michigan 481 09
H. Bernhard Schlege1,t Department of Chemistry, Wayne State University, Detroit, Michigan 48202
and K. Morokuma Institute for Molecular Science, Myodaiji, Okazaki, 444, Japan (Received: October 17, 1983)
Theoretical values for the gas-phase proton affinities (PA’s) of phosphaethyne, phosphaethene, and phosphabenzene have been obtained from ab initio SCF calculations employing analytic gradient techniques for geometry optimization. The sensitivity of the results to the choice of basis set and to the inclusion of correlation is discussed. The PA’s for P-site and C-site protonation are found to be nearly equal in both phosphaethene and phosphabenzeneat the split-valenceSCF level; however, polarization functions and electron correlation both act to favor P-site protonation over C-site protonation, resulting in an isomerization energy of approximately 50 kJ mol-’. By contrast we find C-site protonation to be strongly favored over P-site protonation for phosphaethyne even at the split-valence SCF level. Comparisons are made to computed and observed PA’s both for saturated C-P systems and for saturated and unsaturated C-N systems. The computed order of basicities, H3CPH2> C5H5P> PH3 > H2CPH > HCP (C-site), differs from that for the corresponding N systems, namely, C5HSN> H3CNH2> HICNH = NH3 > HCN, not only in the lower position of C5H5P,as is well-known, but also in the lower position of H2CPH.
Perhaps even more interesting, deuterium labeling experiments’ have indicated that P-site protonation occurs for phosphabenzene, while C-site protonation occurs for arsabenzene.2 These results have prompted us to carry out a theoretical investigation of the gas-phase PA’s and site preferences not only for phosphabenzene (C5H5P)but also for phosphaethene (H,CPH) and phosphaethyne (HCP), all of which contain phosphorus-carbon multiple bonds. Previous theoretical ~ t u d i e s of ~ - the ~ PA’s of phosphorus containing molecules have largely focused upon phosphine and alkyl phosphines, although values for the phosphorus-site (P-site) PA +Fellow of the Alfred P. Sloan Foundation 1981-3.
0022-365418412088- 1981$0 1.50/0
(1) Hodges, R. V.; Beauchamp, J. L.; Ashe, A. J., 111; Chan, W.-T. to be submitted for publication. (2) For reviews of group 5 heterobenzenes see (a) Ashe, A. J., I11 Acc. Chem. Res. 1978, 1 1 , 153. (b) Top. Current Chem. 1982, 105, 125. (3) Whiteside, R. A.; Binkley, J. S.; Krishnan, R.; DeFrees, D. J.; Schlegel, H. B.; Pople, J. A. “Carnegie-Mellon Quantum Chemistry Archive”, Department of Chemistry, Carnegie-Mellon University: Pittsburgh, PA, 1980. (4) Marynick, D. S.; Scanlon, K.; Eades, R. A.; Dixon, D. A. J . Phys. Chem. 1981,85, 3364. (5) Smith, S . F.; Chandrasekhar, J.; Jmgensen, W. L. J . Phys. Chem. 1982,86, 3308. ( 6 ) Clark, D. T.; Scanlan, I. W. J. Chem. SOC.,Faraday Trans. 2 1974, 70, 1222. (7) Thompson, C. J. Chem. SOC.,Chem. Commun. 1977, 322.
0 1984 American Chemical Society