J . Phys. Chem. 1990, 94, 7 128-7 132
7128
scatter and the estimated uncertainty of our results are improved relative to previous work.
Acknowledgment. This work was supported by the U S . Department of Energy, Office of Basic Energy Sciences, Division
of Chemical Sciences. We thank Albert Y. Chang and Michael D. DiRosa for assistance in the laboratory and David F. Davidson and James A. Miller for helpful discussions. Registry No. H. 12385-1 3-6; 02.7782-44-7.
Picosecond Raman Investigation of Interligand Electron Transfer in Ruthentum( II) Complexes T. Yabe, L. K. Orman, D. R. Anderson, Soo-Chang Yu, Xiaobing Xu, and J. B. Hopkins* Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803 (Received: January 16, 1990; In Final Form: May 9, 1990)
Interligand electron transfer has been investigated in the excited MLCT electronic states of mixed ligand ruthenium(I1) polypyridine complexes containing bipyridine, bipyrimidine, and carboxybipyridine. Two-color picosecond Raman spectroscopy has been used to unambiguously establish the vibrational spectrum corresponding to the ligand-localized MLCT states. With these data direct measurements of electron-transfer dynamics are obtained. It is found that interligand electron transfer is complete on a time scale significantly faster than the 30-ps experimental time resolution.
Introduction
Experimental Section
The photochemistry of Ru( 11) complexes has been extensively studied and is still of wide interest’ since the complexes represent a potential source of electrons that can be pumped into subsequent reactions. The excited triplet MLCT states have been characterized as having the photoexcited electron localized on one ligand.2-4 In recent years, there has been great interest in understanding the rate of interligand electron transfer (ILET) following MLCT excitation. By characterization of the rates of ILET, it should be possible to determine the mechanisms leading to electron localization on the excited-state potential surface. In addition, since these molecules have a fixed distance between the electron donor and acceptor ligands, they provide excellent models to test the effects of electronic coupling and energy gap on rates of electron transfer (ET). There have been many recent attempts to measure interligand ET rates, using various time-resolved techniques. Dual emitting states have been observed in complexes of RbS and Ir.6 For Ru(I1) the emission spectrum in mixed-ligand complexes is found to be nearly identical from each chromophore. As a result, time-resolved measurements of electron transfer are virtually impossible. In addition, the electronic coupling in Ru(l1) complexes is thought to be much higher than for other metals. Due to this fact, E T rates are expected to be higher in complexes containing Ru(II).’ It has recently become possible to directly measure ILET by utilization of time-resolved Raman techniques.s-’O With this method, detailed information pertaining to ligand-specific molecular structure can be obtained by means of vibrational frequencies. This paper discusses the use of picosecond Raman spectroscopy to measure ILET in mixed-ligand complexes of 2,2’-bipyridine (bpy) with 2,2’-bipyrimidine (bpym) or 4,4’-dicarboxy-2,2’-bipyridine (carboxy-bpy). Two-color Raman spectroscopy is used to increase the confidence of the spectroscopic assignments. In addition, this method significantly reduces the spectral congestion by removing ground-state bands from the spectrum. Conclusions regarding ILET dynamics are therefore much more straightforward. These results provide an excellent opportunity to investigate the relative importance of electronic coupling and nuclear tunneling to electron-transfer processes.
The experimental apparatus has been previously described in detail.8-10 The picosecond laser is based on a high repetition rate (2 kHz) regenerative laserl1*I2that amplifies chirped pulses from a Quantronix 416 mode-locked laser. Following temporal compression, the laser pulses have a 30-ps pulse width and 1.25 mJ of energy at 1.064 pm. Harmonic generation in BBO (/3-barium borate) converts the fundamental laser frequency to the second and third harmonics used in this experiment. One-color excitation takes place in a rotating cell to ensure that no two pulses interrogate the same region of sample. In the two-color experiments both a free flowing jet and flowing quartz cell were used. Triply distilled water was used as the solvent in all aspects of this work. A conventional scanning double monochromator is used as the dispersive apparatus. Detection consists of a photomultiplier tube and gated integration. The two-color experiment has been previously described in detail.8-’0 Chopped laser excitation (100 Hz) is used to rapidly obtain raw two-color and one-color background spectra in a single scan of the monochromator. The transient spectrum is computed
‘Author
10
whom correspondence should be addressed.
~
~~
~
~
~~
~
~~
~
(1) (a) Nocera, D. G.;Winkler, J. R.; Y m m , K. M.; Bordignon, E.; Gray, H. B. J . Am. Chem. SOC.1984, 106, 5145. (b) Krueger, J . S.;Mayer, J . E.; Mallouk, T. E. J . Am. Chem. SOC.1988, 100, 8232. (2) (a) Bradley, P. B.; Hornberger, B. A,; Dallinger, R. F.; Woodruff, W. H. J . Am. Chem. SOC.1981, 103, 7441. (b) Casper, J. V.; Westmoreland, T. D.; Allen, G. H.; Bradley, P. B.; Meyer, T. J.; Woodruff W. H. J . Am. Chem. SOC.1984, 106, 3492. (3) Carlin, C. M.; DeArmond, M. K. Chem. Phys. Lett. 1982,89(4), 297. (4) Carroll, P. J.; Brus, L. E. J . Am. Chem. SOC.1987, 109, 7613. (5) Halper, W.; DeArmond, M. K. J . Lumin. 1972, 5, 225. (6) (a) Watts, R. J.; Brown, M. J.; Griffith, B. G.; Harrington, J. S. J . Am. Chem. Soc. 1975, 97, 6029. (b) Watts, R. J.; Griffith, B. G.; Harrington, J. S.J . Am. Chem. SOC.1976, 98,674. (c) Watts, R. J.; White, T. P.; Griffith, B. G. J . Am. Chem. SOC.1975, 97, 6914. (7) DeArmond, M. K.; Carlin, C. M. Coord. Chem. Reu. 1981, 36, 325. (8) Orman, L. K.; Hopkins, J. 8. Chem. Phys. Lett. 1988, 149, 375. (9) Orman, L. K.; Chang, Y. J.; Anderson, D. R.; Yabe, T.; Xu, Xiaobing; Yu, Soo-Chang; Hopkins, J. B. J . Chem. Phys. 1989, 90, 1469. (IO) Chang, Y. J.; Xu, Xiaobing; Yabe, T.; Yu, Soo-Chang; Anderson, D. R.; Orman, L. K.; Hopkins, J. B. J . Phys. Chem., in press. (11) Chang, Y. J . ; Veas, C.; Hopkins, J . B. Appl. Phys. Lett. 1986, 49, 1758. (12) Chang, Y . J.; Anderson, D. R.; Hopkins, J . B. International Conference on Lasers; 1986; McMillan, R. W.; Ed.; STS Press: McLean, VA, 1987; p 169.
0022-3654/90/2094-I 1 28%02.50/0 0 1990 American Chemical Society
The Journal of Physical Chemistry, Vol. 94, No. 18, 1990 7129
Electron Transfer in Ru(I1) Complexes
1200
1300
1400
RAMAN
1500 FREQUENCY
1600
1700
Figure 1. One-color picosecond Raman spectrum obtained by excitation at 354.7 nm for Ru(bpym)32+.The two spectra are produced under different laser energy densities. The relative difference in laser fluence is 4 X IO’. Dashed lines are drawn at the frequencies corresponding to bands assigned to the ground electronic state. Asterisks denote excited-state bands. Spectra were obtained in a rotating quartz cell. (A) Ground-state spectrum obtained under low-power excitation. (9) Excited-state and ground-state spectra obtained under high-power excitation; 25 pJ/pulse in a 0.1-mm diameter beam waist.
by subtracting the one-color background spectrum from the raw two-color spectrum. Several scans of this type are typically added to increase the signal-to-noise ratio. Compounds were synthesized as chloride salts by using previously published method^.'^*'^ Purification was achieved by repeated recrystallization and separation on Sephadex LH20. Ru(bpy),C12 was purchased from Aldrich Chemical Co. and used as received. Product identification was verified by comparison to published ultraviolet and visible absorption spectra. The effects of undesired photochemical modification of sample were checked in two ways. First, the adsorption spectra of the compounds were compared before and after laser interrogation. No indication of sample degradation was observed. Second, the Raman spectra were checked to ensure that they were invariant to illumination time over a period of many hours. By comparison, a complete Raman spectrum typically could be obtained in 15 min.
Spectroscopic Analysis of BPYM Complexes To extract ILET dynamics from the vibrational spectrum, assignments must be established for both the electronic state and ligand parentage of the observed vibrational bands. Figure 1 shows the one-color Raman spectrum of [ R ~ ( b p y m ) ~at ] ~354.7 + nm. Excitation at this wavelength produces the excited MLCT state. In addition 354.7 nm is also in resonance with an electronic absorption of the excited MLCT state. As a result, one-color Raman spectra obtained at this wavelength will exhibit vibrational bands from both the ground and excited electronic states. The electronic-state parentage of the vibrational bands is determined by observing the laser power dependence of the spectrum. The low laser power spectrum is shown in Figure IA, and the high laser power spectrum in Figure 1 B. Bands assigned to the ground state are connected by dashed lines. Bands that appear to be enhanced in the high-power spectrum are assigned to the excited MLCT state, and these bands are indicated by asterisks. It is worth noting that the strong ground-state doublet at 1553 and (13) (a) Hunziker, M.; Ludi, A. J . Am. Chem. SOC.1977, 99,7370. (b) Allen, G. H.; White, R. P.; Rillema, D. P.; Meyer, T. J. J . Am. Chem. SOC. 1984, 106, 2613. (14) (a) Elliot. X. M.; Hershenhart, E. J. J . Am. Chem. Soc 1982, 104, 7519. (b) Sprintschnic,G.; Sprintschnic, H. W.; Korsch, P. P.; Whitten, D. G. J . Am. Chem. SOC.1977, 99,4947.
1200
1300
1400
RAMAN
1500
1600
1700
FREQUENCY
Figure 2. One-color picosecond Raman spectrum obtained at 354.7 nm. Concentration is 1 mM. Laser fluence is 25 pJ/pulse in a 0.1-mm diameter beam waist. Dashed lines are drawn at the frequencies corresponding to excited-state bpy bands. Solid lines indicate ground-state bpy bands. The filled in peaks are excited-state bpym bands. Asterisks denote excited-state bpy bands. Arrows are explained in the text. These arrows indicate how the ratio of ground-state to excited-state bpym Raman bands change throughout the series of compounds. Spectra were obtained in a rotating quartz cell. (A) Ru(bpym)32+;(B) Ru(bpym)2-
(bpy)2t; (C) R u ( b p y ” w V ; (D)Ru(bpy)32t. 1581 cm-’ becomes unresolved in Figure 1B due to the growth of an excited-state band at 1560 cm-l. The assignments are confirmed in the two-color spectrum discussed later where ground-state bands are removed by a spectrum differencing technique. These results agree qualitatively with earlierIs one-color transient Raman results on [Ru(bpym)J2+ with the exception that there are many more resolved features in the present experiment. The one-color Raman spectra for the series of mixed ligand [R~(bpym)~-,,(bpy),,]~+ are shown in Figure 2, where n = 0-3 in spectra A-D, respectively. The mixed-ligand complexes in spectra B and C show a mixture of ground-state bands from each ligand. The ground-state bands assigned to the bpy ligand are easy to identify in the spectra because the frequencies are very similar to those found in [ R ~ ( b p y ) ~ ]which ~ + , has been fully characteri ~ e d . These ~ , ~ ~bands ~ are connected by solid lines in Figure 2. This method of assigning the ligand parentage of Raman bands depends on the invariance of the vibrational frequencies of a particular ligand to chemical substitution at another coordination site of the complex. For the complexes discussed in this paper most frequencies remain constant within the resolution of the apparatus as summarized in Tables I and 11. As a result, this method provides extremely high confidence in the assignments. The filled in bands in Figure 2 are assigned to the excited MLCT state localized on the bpym ligand. These bands are identical with those identified in the [Ru(bpym),12+ spectra shown in Figure 1. The excited-state bands due to bpy appear to be absent in the mixed-ligand spectra shown in Figure 2B,C. Dashed lines indicate the expected position of the bpy excited-state bands. In the [ R ~ ( b p y ) ~spectrum ]~+ shown in Figure 2D these excited-state bands are the dominant features observed. However, they are clearly absent in the mixed-ligand spectra shown in Figure 2B,C. For completeness, it should be noted that the bpy band at 1491 cm-I was once assigned2., as an excited-state band but later reassigned* as a ground-state band. Confirmation of excited-state band assignments is demonstrated in the two-color spectrum shown in Figure 3. In this figure the spectra are obtained under conditions where the excited MLCT (15) Chung, Y. C.; Leventis, N.; Wagner, P. J.; Leroi, G. E. fnorg. Chem. 1985, 24, 1966.
7130 The Journal of Physical Chemistry, Vol. 94, No. 18, 1990
Yabe et al.
TABLE Io
669 669 bPYm* 669 1014 1013 bPYm IO13 1039 1040 bPYm* I040 1 I79 1 I77 bPYm* 1 I79 209 bPY * 247 1247 bPY" 1252 283 bPY* 316 321 1317 bPY 1 320b I322 bwm* 337 1333 bb;m 1338 357 1356 bPYm* 1356 413 1409 bpym 1412 426 bPY* 473 1469 bPYm I474 I4746 bPYy 1480 bPY 494 1491 1489 bPY 1 493b bPYm* 1486 1502 bPY * 524 1526 bPYm* 1524 1546 bPY * 1560 bPY 556) (1561) bPYm 1553 1 560b bPYm * 1581 bPYm 1607 1605 1605 bPY Assignments and Stokes frequencies for the ground and excited states. Bands in parentheses are believed to be partially overlapped. In this case the assignment is given to the band with the largest contribution to the peak intensity. First column gives ligand parentage. Excited-state bands are identified by asterisks. Vibrational frequencies were obtained by using a calibration procedure that included the attenuated Rayleigh line in the Raman spectrum. Frequencies are in units of cm-' and are believed to be accurate to f1.5 cm-I. Frequencies determined from the pure two-color spectrum shown in Figure 3C. TABLE I P
assgnt bPY * carboxy-bpy* carboxy-bpy carbox y- bpy * bPY* carboxy-bpy carboxy- bpy * bPY carboxy-bpy* carboxy- bpy bPY* carboxy-bpy ca rboxy-bpy * carboxy-bpy bPY * bPY bPY * bPY * carboxy-bpy bPY bPY carboxy-bpy
[Ru(carboxybPY),12+ 1216 1216 1283 1298 1310 1364 I373 1433 1449 1480
I546 1618
1200
1300
1400
RAMAN
1500
1600
FREQUENCY
Figure 3. Comparison of one-color and two-color picosecond Raman spectra of Ru(bpy),(bpym). Concentration is 2 mM. Dashed lines are drawn at some of the frequencies of excited-state bpym bands. Asterisks also indicate excited-state bpym bands. Time delay between pump and probe pulses is 0 ps. Spectra were obtained in a flowing quartz cell. (A) Total raw two-color spectrum obtained at 354.7 nm with 532-nm excitation. (B) One-color spectrum obtained at 354.7 nm. (C) Pure twocolor spectrum obtained by subtracting spectrum (B) from spectrum (A).
pure transient spectrum in frame C is calculated as the difference spectrum A-B. In Figure 3C ground-state bands have completely subtracted out. This transient spectrum contains bands due entirely to the excited MLCT state. The bands are denoted by asterisks in the spectrum. The assignments obtained from the two-color data agree completely with those resulting from the one-color spectra discussed in Figure 1. Several new bands obscured in the one-color data are now apparent in Figure 3C. In particular, the excited-state band at 1560 cm-I, which was overlapped by ground-state bands in the one-color spectrum, is now [Ru(carboxyclearly visible. The existence of this excited-state band is re~ P Y ) ( ~ P Y ) ~[ IR~u+( ~ P Y ) ~ I ~ ' sponsible for the apparent frequency shift of the cluster of unI209 resolved bands near 1553 cm-' in Figure 2A-C. As the 1214 ground-state bpym bands become less intense in Figure 2B,C, the 1272 excited-state bpym band at 1560 cm-' begins to dominate. 1278 The results of ILET on the relative population of electrons 1282 localized on the bpy and bpym ligand are shown in Figure 4. The top frame is a pure two-color transient spectrum of a 50% mixture 1307 of [ R ~ ( b p y ) ~and ] ~ +[ R ~ ( b p y m ) ~ ] ~The + . bottom frame is the 1316 1317 1363 pure two-color transient spectrum of the mixed-ligand complex 1375 [R~(bpym)(bpy)~]~+. The filled-in peaks correspond to the excited I426 MLCT state localized on the bpym ligand. The data in this figure confirm the conclusion that no bands are observed in the mix1447 ed-ligand complex due to the excited MLCT state localized on the bpy ligand. In fact, the limited spectral congestion provided 1480 by the two-color technique allows us to state that the fraction of (1491) I489 the excited state localized on the bpy ligand must be less than I502 1% compared to that localized on the bpym ligand in the first 30 1546 1542 ps. Time-delay spectra were obtained out to 20 ns by optically (1561) 1560 delaying the pump and probe lasers. No further time evolution (1 605) 1605 of the excited state was observed.
,,Assignments and Stokes frequencies for the ground and excited states. Bands in parentheses are believed to be partially overlapped. In this case the assignment is given to the band with the largest contribution to the peak intensity. First column gives ligand parentage. Excited-state bands are identified by asterisks. Vibrational frequencies were obtained by using a calibration procedure that included the attentuated Rayleigh line in the Raman spectrum. Frequencies are in units of cm-I and are believed to be accurate to f l . 5 cm-I. state is produced by excitation at 532 nm and probed by a lowfluence laser pulse at 354.7 nm. The raw two-color spectrum is presented in frame A, and the one-color spectrum in frame B. The
Spectroscopic Analysis of Carboxy Complexes The vibrational analysis of the spectra corresponding to the carboxy-bpy complexes is similar to that discussed above. For this series of complexes the spectra are quite congested. However, the analysis can be simplified by using two-color spectroscopy. Figure 5 shows the raw two-color, one-color, and pure transient spectra of [R~(carboxy-bpy),]~+ in frames A-C, respectively. Transient bands assigned to the excited MLCT state localized on the carboxy-bpy ligand are indicated by asterisks in Figure 5C. Ground-state bands appear as negative-going peaks due to a combination of population bleaching and a change in optical
The Journal of Physical Chemistry, Vol. 94, No. 18, 1990 7131
Electron Transfer in Ru(l1) Complexes
J(JL
L
n
A
*
bPY
bPYm
EEE
-
,
,
,
,
,
,
,
,
1250
1200
RAMAN
,
,
,
,
,
,
,
1300
,
,
,
,
1350
FREQUENCY
1200
1300
1400
1600
1600
170
RAMAN FREQUENCY
Figure 4. Pure two-color picosecond Raman spectra of Ru(bpy),-
(bpym)obtained by excitation at 532 nm and probing at 354.7 nm; concentration is 2 mM. Both spectra have been corrected for one-color background signal by using the procedure described in the text. Filled in bands correspond to the excited state of bpym. Time delay between pump and probe pulses is 0 ps. Spectra were obtained in a free flowing jet. (A) Expanded spectrum of 5050 mixture of Ru(bpy),2+and Ru(bpym)32+. (B) Expanded spectrum of Ru(bpy)2(bpym)2+with a time delay between the pump and probe of 0 ps.
Figure 6. Pure two-color picosecond Raman spectra. Dashed lines are drawn at some of the frequencies of excited-state carboxy-bpy bands. The filled-in peaks correspond to excited-statecarboxy-bpy bands. Solid lines denote ground-state bands of carboxy-bpy. Arrows indicate the positions of a ground and excited carboxy-bpy band, which are discussed in the text. Pump is 532 nm, and probe wavelength is 354.7 nm. Concentration is 2 mM. Time delay between pump and probe pulses is 0 ps. Spectra were obtained in a flowing quartz cell. (A) [Ru(carboxybpyh12+; (B) [Ru(carbox~-bp~)(bpy)~I~+; (0[ R U ( ~ P Y M ~ + .
nected by solid lines. It is apparent from the data in Figure 6 that bpy excited state bands are not observed in the spectrum of the mixed-ligand complex. It should be noted that the largest bpy excited-state band at 1282 cm-I is near to but clearly shifted in frequency from the carboxy-bpy excited-state band at 1274 cm-l. The bpy excited-state bands at 1426, 1480, and 1502 cm-l are absent from the spectrum. At 1546 cm-I there is an apparent "peak" in the pure two-color spectrum (Figure 6B), which might be mistakenly assigned as a bpy excited-state band. This "peak" is not an excited-state band. It appears as a result of the two adjacent ground-state bands (indicated by solid lines) that are undergoing depletion. The signal-to-noise in these spectra allows a limit of less than 3% to be placed on the excited-state population that could be localized on the bpy ligand after the first 30 ps. Time-delay spectra were obtained out to 20 ns with no further time evolution of the excited state observed. 1200
1300
1400 1500 RAMAN FREQUENCY
1600
1700
Figure 5. Comparison of one-color and two-color picosecond Raman spectra of [Ru(carboxy-bpy)J2+.Concentration is 2 mM. Dashed lines are drawn at some of the frequenciesof excited-statecarboxy-bpybands. Asterisks also indicate excited-state carboxy-bpy bands. Solid lines denote ground-state bands. Time delay between pump and probe pulses is 0 ps. Spectra were obtained in a free-flowing jet. (A) Total raw two-color spectrum obtained at 354.7 nm with 532-nm excitation. (B) One-color spectrum obtained at 354.7 nm. (C) Pure two-color spectrum obtained by subtracting spectrum (B) from spectrum (A).
density of the solvent. When the optical density of the solvent increases due to the transient absorption of the excited MLCT state, the effective optical path length of the probe laser beam in the solution decreases. This causes a decrease in the Raman scattering of the ground-state bands, which results in the appearance of a bleach in the Raman difference spectrum.*-I0 Pure transient spectra for the mixed-ligand complexes [Ru(~arboxy-bpy)~-,(bpy),]~+are shown in Figure 6 , where n = 0, 2, 3, in frames A-C, respectively. The filled-in peaks are the excited-state bands assigned to carboxy-bpy from Figure 5. Dashed lines indicate the expected position of the bpy excited-state bands. Ground-state bands from bpy and carboxy-bpy are con-
Rates of Interligand Electron Transfer One of the problems with extracting interligand electron-transfer rates from the Raman data presented in this paper is that no dynamics are actually observed-the growth of one transient at the same rate as the decay of another is not observed. It might be argued that Raman scattering from the higher energy ligand is not observed because the resonance Raman enhancement for this ligand is reduced in the mixed-ligand complexes. Although this is unlikely, the possibility cannot be entirely ruled out. However, there is one piece of experimental data that suggests that this is not the case. Consider the lower energy ligand, which is thought to be the final acceptor of MLCT excitation (bpym or carboxy-bpy). In both the bpym and the carboxy-bpy series of complexes, there is a very dramatic change in the ratio of the intensities of the ground- to excited-state Raman bands. In particular, consider the bpym series of complexes shown in Figure 2 and compare the bpym ground- and excited-state bands at 1338 and 1356 cm-I, respectively. These two bands are marked by arrows in Figure 2A-C. As more bpy is substituted into the complex (Figure 2B,C), the ground-state bpym band decreases in intensity by a factor of 3 compared to the excited-state band. In other words, as the concentration of bpym in the complex is reduced, the ground-state bpym band decreases in intensity much faster than the excited state bpym band. This result is interpreted
1132
The Journal of Physical Chemistry, Vol. 94, No. 18, 1990
P
1200
1300 1400 1600 R U N FREQUENCY
B
b
L
1600
1700
Figure 7. Total two-color picosecond Raman spectra that show both ground- and excited-state bands. Arrows are indicated to mark the position of a carboxy-bpy ground- and excited-state band for comparison in frames A and B. The excited-state band is additionally denoted by an asterisk. The intensities of these bands is explained in the text. Pump is 532 nm, and probe wavelength is 354.7 nm. Concentration is 2 mM. Time delay between pump and probe pulses is 0 ps. Spectra were obtained in a flowing quartz cell. ( A ) [Ru(carboxy-bpy),]*+; (B) [Ru(carboxy-bpy)(bpyM2+;(C) [Ru(bpy131
'+.
to indicate that there is another route to excitation of the bpym ligand, namely through the bpy ligand. If this is true, the intensity of excited-state bpym will remain invariant to ligand substitution since in every case-[Ru(bpym)J2+, [ R ~ ( b p y m ) ~ ( b p y ) ]and ~+, [Ru(bpym)(bpy)J2+-one bpym excited state is produced regardless of which ligand is initially excited. This same result is also found for the carboxy-bpy series of complexes shown in Figure 6. Two unobscured bands are marked by arrows that represent an excited-state (positive) and ground-state (negative) feature. Figure 7 illustrates the same point for the raw two-color spectrum before subtraction where ground- and excited-state bands are both observed in the spectrum as positive peaks. Figures 6 and 7 illustrate that as more bpy is substituted into the complex, the ratio of intensity for excited-state to ground-state carboxy-bpy bands increases. Here again, as in the bpym-substituted complexes, it appears that there is another route to excitation of the lowest lying ligand. Apparently, initial excitation of the bpy ligand
Yabe et al. is followed by ILET to the lowest lying ligand. Excitation appears to take place through the bpy ligand, and yet no excited-state bpy ligand is detected in the picosecond transient Raman spectrum. I t should also be pointed out that in each mixed-ligand complex, the intensity of the Raman spectrum representing the lowest energy excited MLCT state remains invariant in time. The transient Raman results presented in this paper therefore suggest that fast ILET is occuring and is largely complete during the time window sampled by the laser pulse width. This conclusion suggests a limiting ILET rate that is significantly faster than L3 X 1Olo s-l.
Conclusions The picosecond Raman spectra of mixed-ligand complexes of bpy with bpym and carboxy-bpy indicate that the major fraction of the electrons in the excited MLCT state are localized on the lower energy ligand on a 30-ps or faster time scale. The signal-to-noise characteristic of these data allow a limit of 5 1% and 5 3 % to be placed on the population that remains on the higher energy bpy ligand in the bpym and carboxy-bpy complexes, respectively. The energy gaps that separate the dissimilar ligands on the electron-transfer reaction coordinate can be estimated from ground-state electrochemistry. This is a result of the fact that the electrochemically reduced complex is thought to mimic excited-state behavior with the added electron residing on the ligand.I6 For the mixed-ligand complexes discussed in this paper the energy gap^^'-^* are 3240 and 3080 cm-l between the higher energy ligand (bpy) and the lower energy ligand (bpym and carboxy-bpy), respectively. It has been suggestedI9 that in mixed-ligand complexes it might be possible to utilize the chemical energy stored in the higher energy ligand in the MLCT excited state. The present results indicate that electron transfer is largely complete on a 30-ps or faster time scale for rather modest energy gaps. These results suggest that the chemical potential of the higher energy (less reducing) ligand in mixed-ligand complexes cannot be efficiently utilized in a chemical reaction. Acknowledgment. The experimental assistance of Yong Chang is acknowledged in the initial stages of this research. (16) Motten, A. G.; Hanck, K.; DeArmond, M. K. Chem. Phys. Left. 1981, 79, 541. (17) Rillema, D. P.; Allen, G.; Meyer, T. J.; Conrad, D. Inorg. Chem. 1983, 22, 1617. (18) Note: this reference quotes the reduction potential for 2COOC2H,-bpy and not ZCOOH-bpy: Elliott, C. M.; Hershenhart, E. J. J . Am. Chem. SOC.1982, 104, 7519. (19) Kumar, C. V.; Barton, J. K.; Could, I. R.; Turro, N . J.; Houten, J. V. Inorg. Chem. 1988, 27, 648.
