2302
J . Phys. Chem. 1989, 93, 2302-2306
exchanged zeolites B. In all cases a new bathochromically shifted band is superposed to the absorption on the unreduced silver zeolites (curves 1). This is less pronounced for the sample Ago.sNa11,,-A.
I 24d0 ; ,
.
~
~.
&do
, . '
,
. I : '
CM-'I
24dQ
~:
~
: ~: ' : I : 23dQ CM-1 24d0
' : ! : : ' : :. I 23d0
CM-1
Figure 14. C 0 2 adsorption at 25 "C on silver zeolite A of different reduction degrees performed with the same samples and the same conditions as described in Figure 11. After the remaining CO was pumped off at 150 O C for 1 h 15 Torr C 0 2 was added over 2 h followed by 1 h of evacuation. (1) Unreduced. (2) Reduction time 2 h. (3) Total reduction time 12 h. The composition of the three samples was as follows: (A) A ~ o , ~ N ~ (B) ~ ~A , ~~ -~A. , ,N ~ x ~(C) - AAg,dao,2-A. ,
is bathochromically shifted for the silver-exchanged zeolites. The frequency ratio is constant for all samples and has the same value as in the gas phase (zeolite, 1.029;35gas peak, 1.028760). We are now interested in finding out what happens in partly reduced silver zeolites with the stretching vibration vj of the weakly bound CO,. We therefore exposed the same samples as already described in section 5 to 15 Torr of C 0 2 for 1 h: (A) Ago,~Nall,l-A,(B) A&,3Na7,rA, (C) Agll,@ao,z-A. Before doing so the adsorbed C O was removed by evacuation at 150 OC for 1 h. The thus observed v3 bands after pumping off the C 0 2 by 1 h of evacuation at room temperature are shown in Figure 14. We refer these bands to weakly bound COS because after 1 h of evacuation at 150 OC they are completely removed. Curve 1 represents in all three cases, A, B, and C, the unreduced zeolites. Curve 2 was obtained after 2 h and curve 3 after 12 h of reduction, leading mainly to two effects on COz adsorption. The adsorption capacity is enlarged, especially for medium silver (60) Rothman, L. S.;Young, L. D. G . J . Quam. Spectrosc. Radial. Transfer 1981, 25, 505.
7. Conclusions Infrared studies on zeolites are usually limited to the required frequency range of the bands under study. Applying modem F U R instruments, it is now feasible to cover an enormous frequency range in a single experiment, thus opening new possibilities in exploiting complex systems. Our interest in the FTIR spectroscopy of metal-loaded zeolites is to understand interactions with gaseous reactants such as HzO, DzO, CO, C 0 2 , H2, and Dz. For this purpose we have constructed a high-vacuum cell for in situ transmission studies on self-supporting zeolite wafers of 15-20 pm thickness and attached it to a BOMEM DA3 FTIR instrument. The experimental results reported demonstrate that very high quality information is obtained in the spectral range investigated, namely 20-13 800 cm-'. We have been able to observe several new features such as Evans holes in the H20, D 2 0 stretching vibration region, overtone bands of H20, D,O, and CO,, sharp hydroxyl bands not reported up to now on zeolites A, changes in the far-IR region in presence of CO and upon reduction with H,. It was possible to distinguish between physisorbed, weakly bound, and strongly bound (chemisorbed) carbon dioxide as well as between slowly and fast adsorbing carbon monoxide. Specific spectroscopic and kinetic investigations on the thermal and the photochemical reactivity of gaseous molecules with metal-loaded zeolites and other substrates6I have now become more feasible. Acknowledgment. This work was supported by Grant No. 2.025-0.86 from the Swiss National Science Foundation and Grant N E F F 329 financed by the Schweizerisch Nationaler Energieforschungsfonds. We would like to thank H. Burgy for his contributions. Registry No. HzO, 7732-18-5;D20,7789-20-0; H2, 1333-74-0;D,, 7782-39-0; CO, 630-08-0; C02, 124-38-9. (61) Reller, A. Chimia 1988, 42, 87.
Picosecond Dynamics of Solvent Trapping following Electron Transfer in Transition-Metal Complexes T. Yabe, D. R. Anderson, L. K. Orman, Y. J. Chang, and J. B. Hopkins* Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803 (Received: May 16, 1988; In Final Form: October 3, 1988)
Picosecond Raman spectroscopy has been used to investigate electron transfer between ligands in the metal-to-ligand charge-transfer states of D3 transition-metal complexes. In all of the complexes studied, dynamic solvent effects on the electron-transfer process are shown to be minimal. The interligand electron coupling is apparently not strong enough to overcome the vibrational reorganization energy. As a result, delocalization of electron density over all three ligands cannot occur. An interpretation of the mechanisms responsible for producing this result is presented based on quantum mechanical electron-transfer theory.
Introduction
overthe past
10 years there has been intense interest in the photophysics of d6 polypyridine complexes.~-4 Many of these complexes exhibit strong metal-to-ligand chargetransfer (MLCT) absorptions in the 350-550-nm region of the spectrum. As such, this class of molecules has been strongly
* Author to whom correspondence
should be addressed.
0022-3654/89/2093-2302$01.50/0
pursued for potential use in solar energy conversion utilizing the water-splitting reaction. One of the issues which has been fiercely debated concerns the fate of the electron following photoinitiated (1) Crosby, G . A. J . Chem. Educ. 1983,60(10),
791. (2) Meyer, T.J . Pure Appl. Chem. 1986, 58(9), 1193. (3) DeArmond, M. K. Acc. Chem. Res. 1974, 7, 309. (4) Crosby, G . A,; Highland, R. G.;Truesdell, K. A. Coord. Chem. Reu.
1985, 64, 41.
0 1989 American Chemical Society
Electron Transfer in Transition-Metal Complexes
The Journal of Physical Chemistry, Vol. 93, No. 6,1989 2303
metal-to-ligand charge transfer. Ruthenium(I1) tris(2,2'-bipyridine) ([Ru(bpy)J2+), for example, has 0 3 symmetry. As a result, the transferred charge can be either delocalized over all three ligands or localized on a single ligand. It is thought that the electron localization process is primarily driven by perturbations such as vibrational distortions and solvent trapping. Metal complexes of this type can therefore be used as excellent models to test electron-transfer theory. Many experimental investigations have been performed to investigate the issue of electron localization or delocalization. The most convincing experiment to date was that of Woodruff et al.536 which demonstrated that the electron appeared to be localized on a nanosecond time scale. This result is based on transient resonance Raman spectroscopy. In this experiment, excited-state frequencies of the [ R ~ ( b p y ) ~ MLCT ]~' state were compared to ground-state frequencies of electrochemically generated Na+(bpy-). The close similarity of observed vibrational frequencies in the two cases leads directly to the conclusion that the excited MLCT state of [ R ~ ( b p y ) ~ is] ~representative + of the chargelocalized species (bpy-). Further resonance Raman experiments were performed6 on the series [O~(bpy),(X)~-,]~+ where n = 1, 2, 3 and X = Ph2PCH=CHPPh2. If the electron was delocalized in the MLCT state, the Raman spectrum of [ O ~ ( b p y ) , ' / ~ - ] ~ + should be radically different from the spectrum of [Os(bpy-)(X),I2+ where the electron is forced to be localized by the asymmetry of the molecule. The vibrational frequencies observed6 showed an invariance to ligand substitution demonstrating that the electron was indeed localized in all complexes including [ R ~ ( b p y ) ~ ] ~Additional +. experiments such as time-resolved fluorescence polarization7~*and resonance Raman on other metal c o m p l e x e ~ ~have * ' ~ confirmed this result. In order to develop a full understanding of the nature of the MLCT states, it is important to ask whether the apparent localized states are intrinsic to the molecule or induced by an inter- or intramolecular perturbation. This problem has been considered theoretically," resulting in models of both intrinsically localized12 and delocalizedI3configurations. Experimentally, there is evidence from time-resolved l~minescence'~-'~ and Raman experimentsL7 in low-temperature glasses that the [ R ~ ( b p y ) ~ MLCT ]~' states may be initially delocalized. In the glass, solvent motions can be slowed down to permit more convention1 methods of probing the MLCT state prior to solvent reorganization. If solvent trapping is the major pathway leading from a delocalized configuration, it should be possible to observe the interconversion process in the low-temperature glass. We have investigated the nature of the MLCT states in transition-metal polypyridine complexes using picosecond Raman spectroscopy. The apparatus is based on a powerful new high repetition rate picosecnd laser system pioneered in our laboratory. The high repetition rate allows detailed resonance Raman experiments to be performed while lowering the risks of unwanted multiphoton excitation processes which can be very troublesome when lower repetition rate laser systems are used. The initial results of our work were published in an earlier (5) Bradley, P. G.; Kress, N.; Hornberger, B A.; Dallinger, R. F.; Woodruff, W. H. J . Am. Chem. SOC.1981, 103, 7441. (6) Caspar, J. V.; Westmoreland, T.D.; Allen, G. H.; Bradley, P. G.; Meyer, T.J.; Woodruff, W. H. J . Am. Chem. SOC.1984, 106, 3492. (7) Carlin, C. M.; DeArmond, M. K. Chem. Phys. Lett. 1982,89(4), 297. (8) Myrick, M. L.; Blakely, R. L.; DeArmond, M. K. J . Am. Chem. SOC.
1987, 109, 2841. (9) McClanahan, S. F.; Dallinger, R. F.; Holler, F. J.; Kincaid, J. R. J . Am. Chem. SOC.1985, 107,4853. (10) Chung, Y.C.; Leventis, N.; Wagner, P. J. Leroi, G.E. Inorg. Chem. 1985, 24, 1965. (11) Kober, E. M.; Meyer T.J. Inorg. Chem. 1984, 23, 3877. (12) Braterman, P. S.; Heath, G. A,; Yellowlees, L. J. J . Chem. Soc., Dalron Trans. 1985, 1081. (13) Ferguson, J.; Herren, F.; Krausz, E. R.; Maeder, M.; Vrbancich, J. Coord. Chem. Rev. 1985, 64, 21. (14) Ferguson, J.; Krausz, E. R.; Maeder, M. J . Phys. Chem. 1985, 89, 1852. ( 1 5) Ferguson, J.; Krausz, E. Inorg. Chem. 1987, 26, 1383. (16) Ferguson, J.; Krausz, E. Chem. Phys. Lerr. 1982, 93(1), 21. (17) Krausz, E. Chem. Phys. Left. 1985, 116(6), 501.
RAMAN SHIFT (cm-1)
Figure 1. One-color transient Raman spectrum of [ R u ( b p ~ ) ~ with ]~+a laser energy of 10 pJ/pulse a t the sample. Sample concentration was 2
mM. Excited-state bands have been labeled with an asterisk. The spectra are identified from top to bottom as (a) glycerol solution below the glass point a t -15 OC, (b) glycerol solution above the glass point at 22 OC, and (c) water solution at 22 O C . communication.18 In that study the Raman spectrum which had previously been a ~ s i g n e dto~ ,the ~ localized [ R ~ ( b p y ) ~ config]~' uration was monitored at various pulse widths ranging from 30 to 150 ps. In aqueous solution at room temperature no dynamics could be observed. However, in viscous glycerol an apparent increase in the intensity of bands associated with the localized configuration was observed when the laser pulse width was varied from 30 to 150 ps. From these data it was not possible to distinguish between two possible interpretations: (1) a dynamic interconversion of a delocalized to localized configuration or (2) a localized configuration whose Raman intensity varied from a simple solvent shift of resonance enhancement due to nonequilibrium solvent dynamics. This paper explores the short-time solvent effects on electron transfer for a series of related complexes. In particular, [Ru( b ~ y ) ~ ][~R+~,( M e ~ - b p y ) ~and ] ~ +[ R , ~ ( b p y m ) ~have ] ~ + been investigated, where Me2-bpy is 4,4'-dimethyL2,2'-bipyridine and bpym is 2,2'-bipyrimidine. Electron trapping effects as they relate to solvent motions are probed by measuring the time-dependent concentration of the electron-localized configuration while varying the viscosity of the solution. A common solvent is used, and the viscosity is controlled through temperature. The MLCT states of the complexes discussed in this paper are generally regarded to consist of a mixture of low-lying excited electronic states. In order to prepare the same Boltzmann populated states through photoexcitation, the temperature should not be varied in a way that significantly alters the ground-state vibrational populations. This is important since photoexcitation carries the ground-state population distribution into the excited state. For this reason, a glycerol solvent was chosen since the glass point is only --25 OC below room temperature. W h i l e this work was in progress, B r u s et aI.l9 have published results of similar picosecond experiments on [Ru(bpy)J2+ in low-temperature glasses. It was found19 in this investigation that the electron-localized configuration appeared on a time scale of > 2v,, in which case K , ~ = 1 and k,, = v,K,. This is the high coupling limit which occurs when the electronic coupling Hab2between ligands is large in comparison to the reorganization free energy E,. In this limit the reaction coordinate follows an adiabatic trajectory. The second limit occurs when Ha: is small such that K , ~
