J. Phys. Chem. 1988,92, 3429-3437 were considered but were found to be inadequate. This does not imply that the foregoing proposal must be unique but rather that a good alternative was not found. For instance, if reactive intermediates other than OH., such as H 0 2 . radical (OH- + H202 H02. H2047),are taken into account the rate law no longer agrees with the empirical one.
-
+
Concluding Remarks
From the data set considered here it is evident that the H202 oxidation of L-cysteine in the presence of structurally ordered FeTD and FeTL enantiomeric systems involves decomposition of a Cysteinate-FeII’T complex that rapidly forms at the “exposed” active sites (Figure 1). A negligible sterical discrimination between FeTL-L-cysteine and FeTD-L-cysteine diastereomers results from the conformational mobility of the axially bound substrate, and this is reflected in the absence of stereoselective effects in the reaction. The same complex forms too slowly in the “buried” centers to participate in the oxidation process, because the polymeric matrix partially hinders the accessibility of the central metal ion to the entering substrate molecule. In the absence of (47) Walling, C. Acc. Chem. Res. 1975,8, 125. Weinstein, J.; Bielski, B. H. J. J. A m . Chem. SOC.1979, 101, 58.
3429
hydrogen peroxide, the polypeptide matrix stabilizes the Fe(1II)T-SXys structure but does not control chiral discrimination between the diastereomeric complexes, even when the buried centers are involved. Conformational energy calculations support these conclusions. At variance with the results obtained with L-catecholamines,12 they show both the absence of a sterically preferred pathway for the closest approach of L-cysteine to the central metal ion and the steric feasibility for the formation of a number of almost equally populated diastereomeric adducts. Very poor discriminating effects would have been hence observed even if the structural features of the polymer-shielded active sites could remain unperturbed by the entering cysteine molecule. These findings highlight the difficulty so far associated with modeling a general pattern for stereoselective reactions between chiral species.
Acknowledgment. We thank P. De Santis, G. A. Lappin, and G. Nemethy for helpful discussions. Collaboration of E. Cataldo and G. Musci in some experiments is also acknowledged. This work was supported by MPI (Rome) and the Italian Research Council (CNR). Registry No. [Fe(tetpy)(OH),]*, 6 14 12-01-9; H202,7722-84- 1; sodium poly(L-glutamate), 26247-79-0; sodium poly(D-glutamate), 3081 1-79-1; L-cysteine, 52-90-4; cystine, 56-89-3.
Vibrational Relaxation and Intersystem Crossing in N2(a‘ng) William J. Marinel&,* Byron D. Green, Margrethe A. DeFaccio, Physical Sciences Inc., Research Park, PO Box 3100, Andover, Massachusetts 01810
and William A. M. Blumberg Air Force Geophysics Laboratory, Hanscom AFB, Massachusetts 01 731 (Received: August 10, 1987; In Final Form: December 17, 1987)
We have observed both electronic quenching and very fast level-specific vibrational deactivation of N,(allI,) by N2. Rate coefficients for these processes have been determined, and the role of the collisional coupling of the N2(a’II,) state to the N2(a”Z[) state in the relaxation of the a’II, state has been assessed.
Introduction
Lyman-BirgeHopfield (LBH) band emission due to transitions from N,(alII,) to N2(X1Zg+)ground state is a prominent feature in the ultraviolet spectrum of the quiescent as well as the aurorally disturbed atmosphere. Observations of LBH emission’” from high altitude auroras are consistent with the predictions of models that include Franck-Condon type excitation of ground-state molecular N 2 to the N2(a) state by energetic electrons followed by radiative r e l a ~ a t i o n . ~However, the band shape of LBH emissions from penetrating lower altitude auroras and from the dayglow and nightglow where N,(a)-state vibrational distributions are affected (1) Gentieu, E. P.; Feldman, P. D.; Meier, R. R. Geophys. Res. Lett. 1979, 6, 325. (2) Park, H.; Feldman, P. D.; Fastie, W. G . Geophys. Res. Left. 1977, 4 , 41. (3) Huffman, R. E.; LeBlanc, F. J.; Larrabee, J. C.; Paulsen, D. E. J. Geophys. Res. 1980, 85, 2201. (4) Takacs, P. Z.; Feldman, P. D. J. Geophys. Res. 1977, 82, 5011. (5) Rottman, G. J; Feldman, P. D.; Moos, H. W. J . Geophys. Res. 1973, 78, 258. (6) Paresce, F.; Lumpton, M.; Holberg, J. J. Geophys. Res. 1972, 77,4773. (7) Ajello, J. M.; Shemansky, D. E. J. Geophys. Res. 1985,90,9845-9861.
0022-3654/88/2092-3429$01 S O / O
by electronic quenching, vibrational relaxation, and intersystem collision-induced cascade are not u n d e r s t o ~ d . ~ ~ ~ ~ ~ Collisional processes involving N2(a) are complicated by the existence of two other states with comparable term energies: the a’*& and the wIA,. This trio of electronic states gives rise to a system of nested vibronic levels that are radiatively as well as collisionally coupled. The relative energies of these states is displayed by using a ladder diagram in Figure 1. The allowed radiative transitions w1AU a l n , and a ’ n , a’l& comprise the McFarlane infrared systemg of emissions in the 3-8.5-pm wavelength range. Transitions from the wlAu state to the a’n, state dominate the McFarlane bands9 The transition probabilities for these bands have not been experimentally determined. No emission resulting from a’ a state transitions has been observed,
-
-
-
(8) Torr, M. R.; Torr, D. G.; Eun, J. W. J. Geophys. Res. 1985, 90, 4427-4433. (9) McFarlane, R. A. IEEE J. Quantum Electron. 1966, 2, 229-232. (10) Herzberg, G.; Herzberg, L. Nature (London) 1948, 161, 283. (11) Douglas, A. E.; Herzberg, G. Can. J. Phys. 1951, 29, 294-300. (12) Lofthus, A. Can. J. Phys. 1956, 34, 780-789. (1 3) Vanderslice, J. T.; Tilford, S.G.; Wilkinson, P. G. Astrophys. J . 1965, 141, 395-426.
0 1988 American Chemical Society
Marinelli et al.
3430 The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 DIFFUSION PUMP isELOW
-4
,’,”!:!,”;
SHROUD
-5
3-
-
76
13
74
C
72~
-
L
’-
70t-‘ I
I
-1
-0
1
-I
I
-0
I
I
O-
o“2,
o‘ng
w‘Au
Figure 1. Energy level diagram for N, singlets.
presumably as the consequence of the long radiative lifetime of the a’ state and strong collisional couplings of a’-state levels to the a’n, or other states. LBH emission is observed from vibrational levels 0-6 of the a’n, state to the ground state. Levels higher than 6 are weakly predissociated. The Ogawa-Tanaka-Wilkinson-Mulliken band emissions from the a”ZU(u=O)-X1.Z +(u”=3-8) transitions are also ~ b s e r v e d . ’ ~The ~ ’ ~transition wfA,,-XIZp+ has been shown to be pressure dependent.I6 N o predissociations of this state have been observed. The collisional and radiative relaxation of the coupled singlet states has been studied in a series of papers by Golde and Thr~sh,”-~O Freund,,’ and van Veen et The early work by Golde and Thrush used a discharge flow reactor to create active nitrogen in Ar at relatively high pressures (1-6 Torr). Their observations of a’-X and a-X emission showed that N,(a,u=O) could be excited by N2(a’,u’=O) through the process N2(a’,u’=0)
+ Ar
POROUS TUBE
lATOR
I
-2
I.-’.’-‘-
n
E L E C T R O N GUN
-
N,(a,u’=O) + Ar AE = -1212 cm-’ (endothermic) (1)
This conclusion was supported by the observed variation in a’/a emission intensities with pressure. Electronic quenching of N2(a,u=O) by Ar and N, was measured to be quite slow, kAr3 3 x IO-’, cm3 molecule-’ s-’ and k N 22 1.7 X IO-’, cm3 molecule-’ s-’~*O Quenching by CO, was observed to be quite fast (kco2 = 6.5 X IO-“ cm3 molecule-’ s-I) and be strongly dependent on vibrational level. Golde and Thrush observed quenching of high vibrational levels of N,(a) by N, to be strongly dependent on both vibrational level and temperature. Apparent rate coefficients ranging from 1.6 X IO-], to 2.7 X IO-” cm3 molecule-’ s-l were observed for levels 0-6. These rates were comparable to Ar quenching, suggesting that no resonances are involved and little energy is transferred in the quenching process. This would seem to imply that the relaxation of N,(a,u’) is primarily vibrational, possibly proceeding via intersystem crossing with the a’ and w states. Model calculations by Freund,’ also indicate that intersystem radiative cascade may play an important role in the apparent lifetimes of the singlet states. His calculations show.that the a’ state is efficiently populated by cascade and that cascade processes affect the apparent a-state radiative lifetimes. Such a cascade could result in nonexponential population decays and an apparent (14) Ogawa, M.; Tanaka, Y . J . Chem. Phys. 1959, 30, 1354-1355. (15) Ogawa, M.; Tanaka, Y . J . Chem. Phys. 1964, 32, 754-758. (16) Tanaka. Y . ; Ogawa, M.; Jursa, A. S . J . Chem. Phys. 1964, 40. 3690-3700. (17) Golde, M. F.; Thrush, B. A. Chem. Phys. Lett. 1971, 8, 3 7 5 . (18) Golde, M. F.; Thrush, B. A. Proc. R. SOC.London A 1972, 330, 79-95. (19) Golde, M. F.: Thrush, B. A. Proc. R. SOC.London, A 1972, 330, 109-120. (20) Golde, M. F. Chem. Phys. Lett. 1975, 31, 348. (21) Freund, R.S. J . Chem. Phys. 1972, 56, 4344. (22) van Veen, N.; Brewer, P.; Das, P.: Bersohn, R. J . Chem. Phys. 1982, 77, 4326.
Figure 2. LABCEDE apparatus.
decrease in the a-state lifetimes with vibrational level, in contrast with those inferred from absorption oscillator strengths. The a’ radiative decay may calculations also demonstrate that a account for up to 20% of the total radiative losses for high vibrational levels of the a state, with the LBH bands accounting for the remaining losses. More recently van Veen et aL2, used direct two-photon laser excitation of N,(a,u=O,l) to measure relaxation rate coefficients for N2. They observed regimes of fast exponential (0.05-0.2 Torr), nonexponential(O.2-1.0 Torr), and slower exponential (1-10 Torr) decays. These regimes were ascribed respectively to simple relaxation of the a state, partially coupled relaxation of the a and a’ states, and fully coupled relaxation of the two manifolds. In the low-pressure regime, relaxation rate coefficients of 2.1 X lo-” and 2.0 X IO-” cm3 molecule-’ were measured for vibrational levels 0 and 1, respectively. These rate coefficients are the sum of electronic and, in the case of u’ = 1, vibrational relaxation. In the high-pressure regimes a decay rate coefficient of 2.3 X cm3 molecule-’ s-l is obtained for both u = 0 and 1. This rate coefficient is probably a measure of the deactivation of the a’ state, either through coupling with the a state or via relaxation to another manifold. In the measurements reported here, the collisional relaxation of N,(s) by N 2 has been investigated to characterize electronic quenching, vibrational relaxation, and the effects of collisional and radiative coupling of the a state to other significant singlet states of N,. These measurements provide a global picture of N,(a) relaxation both to provide insight into the upper atmospheric LBH band observations and to provide guidance for state-to-state laser-based investigations of the singlet systems of h2.
-
Experimental Section
These experiments were performed using the LABCEDE facility at the Air Force Geophysics Laboratory. The LABCEDE chamber is a cylindrical vacuum tank, 3.4 m long and 1 m in diameter, in which atmospheric species are irradiated by an electron beam. A schematic diagram of the experimental apparatus is shown in Figure 2. The gas to be irradiated enters the reaction chamber at one end through a large porous-tube array, flows under essentially plug-flow conditions along the longitudinal axis of the chamber, and flows out through a 32-in. diffusion lamp backed by a Roots blower/fore pump combination (effective pumping speed = 2.6 X IO4 L s-I). This flow pattern limits the residence time of gaseous species in the electron beam volume to a few milliseconds at most so that quenching by electron-beamcreated species is negligible. The electron beam propagates transverse to the axis of the vacuum tank, a little over a meter from the upstream end of the tank. Fluorescence from the electron-irradiated N2 was observed from the upstream end of the tank, perpendicular to the electron beam axis, through BaF, windows. The emission is spectrally resolved by using a McPherson 0.3-m monochromator equipped with a 2400 lines/” grating blazed at 150 nm. A partially solar-blind photomultiplier tube (HTV R1220) was used on the monochromator. In the absence of atmospheric attenuation this detection system has good response between 120 and 310 nm. The entire
Relaxation and Intersystem Crossing in Nz(alII,) monochromator and light path to the vacuum chamber was purged with Nl, but the effective system wavelength response was limited by residual 0, absorption to between 175 and 310 nm. The output of the photomultiplier is converted to voltage in a fast-current amplifier (Ithaco Model 121 1) and then synchronously detected by using a lock-in amplifier (Princeton Applied Research Model 124). A computerized data-acquisition system (Compaq/Data Translation DT2801A) was used to record the spectra from the lock-in amplifier. The data were subsequently transferred to a mainframe computer (Prime 450) for further analysis. The relative wavelength response of the detection system was determined by using two techniques. In the region from 175 to 220 nm, the intensities of LBH bands originating from a common upper level of the a’n, state were compared with the measured fluorescence branching ratios of S h e m a n ~ k y . , ~The ratios of measured to predicted intensities established the response function. In the region from 195 to 310 nm the response was determined by using a calibrated deuterium arc lamp (Optronics Laboratories). Additional calibration experiments were performed in which emission from CO(A-X) and NO(A-X) bands in the region from 175 to 290 nm was observed, in situ, from electron excitation of CO and NO. Fluorescence branching ratios for CO(A-X) from K r ~ p e n i e ,and ~ NO(A-X) from Piper and Cowles2s were used to correct observed intensities and obtain relative responses. All calibrations agreed to within 20% where they overlapped with the exception of the calibration using the deuterium lamp. From 200 to 230 nm the D, lamp deviated from the other calibration by as much as 50%. This is potentially due to scattered light. Hence the molecular emissions were used in this wavelength region. The response of the detection system in the 175-205-nm region varied by as much as 15% from day to day due to variations in purge efficiency. The response over this region was determined separately for each day’s experiments from the LBH branching ratios observed at low pressure. A 50% duty cycle pulsed electron beam with a frequency of 195 Hz was used to excite the N2 (with a high (99.9999%) purity grade). Average beam currents of approximately 5.5-8.5 mA at an energy of 4.5 keV were employed. The beam current was held constant to f 0 . 2 mA within a given series of experiments. A weak axial magnetic field (7 G) was used to confine scattered secondary electrons and enhance excitation efficiency. Observations of N2(C-B) emission were used to establish that these excitation conditions were well outside the realm of nonlinear beam-plasma interactions. The N2(C) state has an electron excitation crosssection function that is similar in shape and magnitude to the N,(a) state. Both states sample the same electron energy distribution.26 The C-state radiative lifetime is sufficiently short (37 ns),’ that collisional quenching is not important over the pressure range of these experiments. Moreover, there is no experimental evidence for a significant radiative or collisional feed of the C state from other electronic states. Hence, the population of the C state represents a balance between direct electron excitation and pure radiative relaxation. The spectrally resolved fluorescence from the C-B (0-0) transition was monitored for a variety of experimental conditions. The C-state populations were observed to increase linearly with electron current at a constant pressure. Additionally, the C-state populations varied linearly with N, pressure, for pressures between 0.1 and 10 mTorr. These results are presented in Figure 3. The cascade contribution from higher states such as the E state appear not to be significant under LABCEDE operating conditions. No (E-A) zmission (Herman-Kaplan bands) is observed. On the basis of X3 scaling, these bands would be more than 100 times more intense than the E-C transitions, so that C-state radiative feed is small. The observed C-state distribution is consistent with a Franck-Condon excitation (23) Shemansky, D. E. J . Chem. Phys. 1969, 51, 5487. (24) Krupenie, P. H. Natl. Stand. Ref: Data Ser. (US., Natl. Bur. Stand.) 1966, 5. ( 2 5 ) Piper, L. G.; Cowles, L. M. J . Chem. Phys. 1986, 85, 2419. (26) Cartwright, D. C.; Chutjian, A,; Trajmar, S.; Williams, W. Phys. Rea A 1977, 16, 1013. (27) Lofthus, A,; Krupenie, P. H . J . Phys. Chem. Ref. Data 1977,6, 113.
The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 3431
propagation. Because diffusion of Nz(a) ( 7 = s) molecules out of this field could be an important loss process, data were acquired with different beam geometries. The data presented in this article were taken with the electron beam defocussed so that it was larger spatially than the field of view. The results obtained in this manner did not differ from those obtained with a tightly focussed beam, demonstrating that diffusion was not a problem: diffusion into and out of the field of view was balanced to first order. We extracted relative vibrational populations for the bound N,(a) levels 0-6 and the C-state levels 2-4 using a linear least-squares spectral fitting code. In addition, we were also able to extract a relative population for the N2(cq/) state using emission from the cq/-a,(rl band. This emission occurs in the midst of the (C-B) Av = 2 sequence. This linear least-squares spectral fitting code uses known line positions and intensities to calculate emission profiles (basis functions) for each vibrational level of a given electronic state. The experimental spectrum is fit by using these basis functions in a least-squares sense to minimize the error in the predicted intensities. The output of this fit is a set of vibrational populations for each electronic state that can be made absolute if an exact calibration factor is determined. For our experiments absolute calibrations are not required; however, we do obtain a form of absolute calibration by setting the measured N2(a,u) populations in a ratio with the N2(C,u=2) population determined at the same pressure. Since both states are excited by the same energy electrons, this ratio essentially normalizes the a-state populations for variations in electron-beam current from pressure to pressure. More significantly, the C state is not quenched at any of the pressures we employ. Hence, the ratio of N2(a,u) to N2(C,2)is independent of pressure in the absence of N2(a) vibrational re-
Marinelli et al.
3432 The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 12
,
1
v)
N2(C3nU-E3nglAV* 2
I O
>
k
12
J
I
4
1
I
3
2v’
I
0.8
W
L 0.6 W
N 4
0.4 0
0.2
0.0
179
184
189
194
199
204
209
214
219
WAVELENGTH (nm) I O
291
293
295
I 297
j
299
,-..: I
WAVELENGTH (nm)
E-
Figure 5. Spectral fit to N,(C-B) Au = 2 sequence band with cql-a (0-1) band included. The cql-a band falls in the region between the C-B (2-0) and (3-1) bands and is not distinctly observable at this resolution. Spectral simulation of the Au = -2 sequence clearly indicates the presence of the cql-a (0-1) band.
0 8
Ln
z
bW -
5
06
n w
2
0 4
E *
0 2
4
0 1.0mTorr 0 72mTorr
hv\r\AhA-
00 179
184
189
194
199
204
209
A ,
214
219
WAVELENGTH (nml
Figure 4. N2(a) emission spectra at (a) 1.0 and (b) 72 mTorr. The data are shown as the light lines and the fits to the data are shown as the heavy
lines. laxation or feed. It is through an analysis using this ratio that we extract kinetic information.
Results and Analysis A typical spectrum of the LBH bands recorded at 1 mTorr is shown in Figure 4a along with the least-squares spectral fit to the data. The least-squares fit used the tabulated spectroscopic parameters from Huber and HerzbergZ8and the modified transition probabilities of S h e m a n ~ k y .These ~ ~ transition probabilities were scaled to match the 80-ps lifetimes for u = 0 measured by Holland29and by Pilling et aL30 in accordance with the recommendations of Ajello and Shemansky.’ The transition probabilities used are listed in Table I. In the 180-220-nm spectral region there are several bands originating from each vibrational level except u = 0, and the agreement between the fit and the data is excellent. In addition to the LBH features present, there are also several other emission features that cannot be identified as atomic lines. These features are more evident in the spectrum recorded at 72 mTorr and shown in Figure 4b. These features appear to be a progression from a‘,v=0 to low vibrational levels of the ground state (Ogawa-Tanaka-Wilkinson-Mulliken bands). However, the spectral fitting code does not reproduce the exact line positions when the spectral parameters from Huber and Herzberg**are used but instead produces features slightly irregularly displaced from those observed experimentally. The observation of multiple bands in this progression prompt us to believe the a’-X assignment. The differences between the data and the fit are possibly due to perturbations in the rotational structure of the a’ state that result in the displacement of peak rotational features to higher J values and an abnormally large Q branch.31 These perturbations have (28) Huber, K. P.; Herzberg, G. Van Nostrand: New York, 1979. (29) Holland, R. F. J . Chem. Phys. 1968, 51, 3940. (30) Pilling, M. J.; Bass, A. M.; Braun, W. J . Quant. Spectrosc. Radial. Transfer 1971, 11, 1593-1604. (31) Tilford, S. G.;Wilkinson, P. G.;Vanderslice, J. T. Astrophys. J . 1965, 141, 421-442.
I
I 1
0
2
0
z 5
6
VIBRATIONAL LEVEL
Figure 6. N2(a,u)/N2(C,2)population ratios at 1.0 and 7 2 mTorr ob-
tained from spectral fitting. not been included in our spectral fitting code. To account for these features, we have generated individual basis functions for each of the emission bands and fit the spectrum using these basis functions. The transitions a’-X(u’=O-+u’’=6-9) are tentatively assigned. A typical spectrum of the N2(C-B) Au = 2 sequence bands in the near-UV region around 298 nm and the least-squares fit to those bands are shown in Figure 5 . A basis function for the N2 cl-a (0-1) band, which falls between the C-B (2-0) and 3-1 bands, is required to achieve a satisfactory fit. As shown in Figure 2, the N2(C,u=O) population scales linearly with pressure. The nonzero intercept reflects detector dark-current levels. A similar plot of the N2(cq/,u=0) population also shows a linear dependence on N2 pressure. The ratio of N,(C) u = 2, 3, and 4 populations are constant with pressure and consistent with Franck-Condon type excitation by electrons and radiative relaxation of the excited states. This is further evidence that there are no additional radiative or collisional processes populating the C state. The relative vibrational populations of N2(a,u)/N2(C,u=2) that were determined from the spectral fits are given in Table 1I for each pressure studied. Averages are presented when multiple experiments were performed a t the same pressure. Agreement was generally better than 20%. Plots of the N,(a)/N2(C,2)
The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 3433
Relaxation and Intersystem Crossing in N2(a111,)
TABLE I: N2(a111,-X1E,+) Einstein Coefficients, Add, (SI), and Transition Frequency (cm-l) Ad",, S-'
transition frequency, cm-' u I'
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
sum
u'= 0
u'= 1
u'= 2
u'= 3
u'= 4
u'= 5
u'= 6
732 68 950 2 327 66 621 3 405 64319 3 050 62 047 1880 59 803 840 57 588 283 55 402 72 53 244 14 51 116 1 49016 0 46 946 0 44 905 0 42 892 0 40910 0 38 956 0 37 032 0 35 137 0 33 273 0 31 437 0 29 632
2 116 70617 3 183 68 287 1181 65 986 7 63713 1072 61 469 1987 59 254 1650 57 068 847 54911 303 52 782 78 50 683 14 48612 1 46 571 0 44 559 0 42 576 0 40 622 0 38 698 0 36 804 0 34 939 0 33 104 0 31 299
3 360 72 255 1723 69 925 54 67 624 1556 65 352 1090 63 108 5 60 893 709 58 707 1538 56 549 1307 54421 654 52 321 220 50251 54 48 209 9 46 197 1 44214 0 42261 0 40 337 0 38 442 0 36 577 0 34 742 0 32 937
3 865 73 866 234 71 536 1305 69 235 1063 66 962 58 64719 1209 62 504 709 60317 5 58 160 758 56 032 1316 53 932 983 51 862 440 49 820 132 47 808 29 45 825 3 43 872 0 41 948 0 40053 0 38 188 0 36 353 0 34 548
3616 75 449 127 73 119 1781 70818 7 68 545 1176 66 302 480 64 087 230 61 900 1060 59 743 318 57615 107 55515 890 53 445 1120 51 403 683 49 391 260 47 408 67 45 455 12 43 531 1 41 636 0 39771 0 37 936 0 36 131
2 929 77 005 1018 74 675 92 1 72 373 616 70 101 901 67 857 136 65 642 1034 63 456 74 61 298 540 59 170 785 57 071 43 5 5 000 349 52 959 974 50 947 867 48 964 420 47010 130 45 086 29 43 192 3 41 327 0 39 492 0 37 686
2 129 78 533 1972 76 203 96 73 901 1392 71 629 43 69 385 1000 67 170 187 64 984 547 62 826 600 60 698 36 58 598 750 56 528 378 54487 29 52 475 634 50 492 914 48 538 587 46614 225 44719 56 42 855 9 41 020 1 39 214
12 609
12 445
12285
12114
11 941
11 776
11 592
TABLE II: N7(a.v)/N7(C.2) Pooulation Ratios'
vibrational level press., mTorr 0.25 1.00 2.00 3.00 4.00 6.00 8.00 10.00 12.50 15.00 20.00 30.00 40.00 50.00 60.00 72.00
0 1.18 3.11 4.05 4.92 6.19 6.97 7.93 7.80 9.61 8.57 9.05 9.20 7.97 7.79 6.85 6.30
1
2
3
4
6.38 7.30 6.58 6.96 7.94 8.09 7.80 7.90 8.16 6.53 6.62 5.61 4.67 4.21 3.50 3.17
8.59 9.01 6.91 7.02 7.25 6.94 6.43 6.12 6.17 5.40 4.91 4.06 3.48 3.20 2.75 2.45
7.35 6.42 4.75 4.26 3.92 3.66 3.07 2.87 2.59 2.22 1.86 1.45 1.15 1.10 0.94 0.86
6.03 4.92 3.14 2.65 2.40 1.87 1.48 1.36 1.17 0.95 0.76 0.53 0.44 0.34 0.26 0.25
5 3.88 3.39 2.15 1.76 1.60 1.26 0.93 0.88 0.71 0.57 0.45 0.31 0.25 0.20 0.16 0.15
6 2.48 2.15 1.28 1.13 1.02 0.82 0.66 0.87 0.58 0.36 0.42 0.28 0.25 0.16 0.14 0.14
'All ratios should be multiplied by lo3 to obtain actual values.
vibrational population ratios at 1 and 72 mTorr are shown in Figure 6. At 1.O mTorr less than one collision occurs per a-state radiative lifetime on average. The observed vibrational populations agree quite well with those expected from Franck-Condon type excitation of the a and a' manifolds and radiative relaxation of the two states. The vibrational distribution in the a state as observed at 72 mTorr is substantially more relaxed than that at 1 mTorr. The populations of levels 5 and 6 have decreased by
nearly a factor of 10, while the-population of v = 0 has increased nearly twofold. In addition to the apparent vibrational relaxation, the total population of N2(a) decreases with pressure, indicating that electronic deactivation is also occurring at a slower rate. An average electronic deactivation rate coefficient for the a state can be obtained by considering how the total Nz(a) population, in a ratio with the N2(C,0=2) population, varies with pressure. The relevant reactions to be considered are
+ + ki"
N 2 + e-
N2(a)
N2
e-
N2(C,2) + e-
krC1
N2(CJ)
+ e-
N2(?) + N 2
kQ
kfc2
N,
N,(a)
N,(B)
(4) (5)
+ hvc
(6) Both N2(a) and N2(C) are in steady state, and their populations can be written as [N2(a)1 = k?[N21/(k,a + kQ[N21)
(7)
[N2(C,2)] = k f C ~ 2 [ N Z ] / k ~ ~ 2 In these expressions kf and k, represent the formation and radiative rate coefficients, k , is the average quenching rate coefficient, and [N2(a)] is the total N2(a) population obtained by summing over
3434 The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 TABLE III: N,(allI,-a"Z,)
If
0 1
2 3 4 5 6 7
8 9 10 11
12 13 14 15 16 17
18 19
sum ( a )
Einstein Coefficients, Add' (s-I), and Transition Frequency (cm-') Aotc,,,s-' transition frequency, cm-l
0 99.74 1211 1.54 -294 118.10 -1776 206.90 -3235 164.40 -467 1 92.23 -6083 42.33 -7473 17.11 -8840 6.50 -10184 2.31 -1 1506 0.83 -12806 0.31 -14084 0.14 -15341 0.05 -16575 0.05 -17789 0.00 -18981 0.00 -20151 0.00 -21301 0.00 -22430 0.00 -23538 99.74 I.'=
Marinelli et ai.
I.'=
1
614.20 2878 33.74 1372 0.08 -1 10 143.30 -1569 422.70 -3004 467.10 -4417 330.40 -5807 186.70 -7173 89.95 -8518 39.14 -9840 16.27 -1 1140 6.37 -12418 2.35 -13674 0.99 - 14909 0.99 -16122 0.17 -17314 0.07 -18485 0.02 -19635 0.00 -20764 0.00 -21872 647.9
u'= 2
351.60 4516 832.40 3010 2.30 1528 0.01 69 106.60 -1366 527.00 -2778 786.00 -4168 697.80 -5535 472.20 -6879 261.70 -8201 129.30 -9501 58.62 -10779 26.76 -12036 10.92 -13270 10.92 -14484 I .94 -15676 0.71 -1 6847 0.30 -17996 0.14 -19125 0.06 -20233 1186.
L"= 3 48.44 6127 84.17 462 1 804.80 3138 6.22 1679 0.13 244 59.40 -1 168 499.00 -2557 1003.00 -3924 1100.00 -5269 859.70 -6591 557.20 -7891 313.10 -9 169 157.90 -10425 79.35 -1 1660 79.35 -12873 15.62 -14065 7.66 -15236 3.19 -16386 1.25 -17515 0.49 -18623 943.7
I."= 4
o'= 5
ut= 6
1.30 7710 161.80 6204 1325.00 4721 644.50 3263 35.66 1827 0.18 414 25.59 -974 397.40 -2341 1073.00 -3685 1424.00 -5008 1308.00 -6308 946.90 -7586 605.10 -8842 339.20 -10077 339.20 -1 I290 91.83 -12482 41.63 -1 3653 19.65 -14803 9.79 -15931 4.76 -1 7040 2169
0.01 9266 5.37 7759 33.20 6277 17 12.00 4818 439.90 3382 76.51 1970 0.03 580 8.42 -786 275.30 -2130 988.60 -3452 1618.00 -4752 1716.00 -6030 1423.00 -7287 101.20 -8521 626.40 -9734 354.60 -10926 203.70 -12097 102.60 -13247 48.98 -14376 22.32 -1 5484 2567
0.00 10793 0.02 9287 12.46 7805 553.90 6346 1996.00 4910 249.80 3498 114.90 2108 0.18 74 1 1.91 -602 163.10 -1924 821.10 -3224 1625.00 -4502 2019.00 -5759 1901.00 -6993 1465.00 -6993 1008.00 -9399 622.80 -10569 358.50 -11719 203.80 -1 2848 110.40 -13956 2927
all vibrational levels. As has been demonstrated, losses due to diffusion are negligible. Combining expressions 7 and 8 yields the result
sum ( a 3 0.00 1.54 118.10 350.20 693.70 I 146.00 1683.00 23 I 1 .OO 3019.00 37 38 .OO 445 1 .oo 4667.00 4234.00 2432.00 2310.00 1472.00 876.60 484.30 264.00 138.00
I
I
10
20
4 0
(9)
-
where 6 = kfc,z/kf"
(10)
There is thus a linear relationship between the ratio [N,(C,F 2)]/[N,(a)] and total N, pressure. The quenching rate coefficient is simply determined by multiplying the ratio of the slope to the intercept of the line by k,". A plot of [N2(C,2)]/[N2(a)]versus N, number density is shown in Figure 7. The data do indeed form a straight line with slope (9.55 f 0.44) X cm3 molecule-' and intercept (1.54 f 0.05) X (statistical error bars only). The radiative rate used in expression 10 is necessarily an average over all of the populated vibrational levels. The a-X transition probabilities vary slowly with vibrational level, and hence averaging these values is not a significant source of error. However, radiation from the a-a' transition must also be considered as part of k:, and this component will vary rapidly with u due to the roughly frequency-cubed scaling of the low-frequency infrared transitions. We have calculated the transition probabilities for the a-a' transition using the known line positions and Franck-Condon f a ~ t o r s ~and ' * ~the ~ transition moments of Yeager and M c K o ~ . These ~ ~ are sum(32) Yeager, D. L.; McKoy, V. J . Chem. Phys. 1977, 67, 2473
4
-0 r 2 -
30
a , r;;l
N.
N
a
20
10 0
[N2](
10'4M0LECULE ~ m - ~ )
Figure 7. Stern-Volmer plot for N,(a) electronic relaxation.
marized in Table 111. When both transitions are considered, the averaged transition probability, ( k : ) , becomes 13 835 725 s-', a 15% increase over the a-X value. This results in a value for kQ of (8.9 0.7) X cm3 molecule-' s-l. A similar analysis for vibrational level 6 only, which can have no feed from higher
*
*
The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 3435
Relaxation and Intersystem Crossing in N2(a1ng) levels, yields a larger quenching rate coefficient of (8.5 f 1.8) x lo-” cm3 molecule-’ s d . The data interpreted in this manner support the existence of a fast vibrational deactivation/intersystem crossing process competing with a slow electronic deactivation as the primary pathways for collisional relaxation of N2(a) molecules. The absence of any real curvature in the plot of Figure 7 might indicate that radiative and/or collisional feed from either N2(a’) or N2(w) is not a significant source of excitation for N2(a). This prompted us to proceed with a vibrational deactivation analysis of N2(a) assuming that the a’ and w manifolds do not play an important role in a-state vibrational relaxation. This process allows us to place bounds on the deactivation rates. The vibronic relaxation of N2(a) can be treated as a single quantum vibrational relaxation within the manifold that competes with a deactivation process that removes vibrational quanta entirely from the manifold. At the lowest pressure employed (0.25 mTorr), N2(a) only relaxes radiatively. The steady-state rate of formation of a given level can be equated to its radiative loss, hence R, = N,OA,
401
I5I
30
25t
(11)
where R, is the rate of formation of level u, Nuois the ratio of N2(a,u)/N2(C,2) in the collisionless limit, and A, is the total emission transition probability for level u. When normalized by the C state, R, is independent of pressure. At higher pressures collisional feed from higher vibrational levels provides an additional source, and collisional quenching becomes an additional loss mechanism. Hence expression 11 becomes Equating expressions 12 and 13 gives
where kEUis the electronic quenching rate coefficient for level u, ku,L,lis the vibrational quenching rate coefficient for level u, and k,+l,, is the vibrational quenching coefficient for the level u + 1, which feeds level u. An equation such as (1 3) can be written for each vibration level and summed for all levels above the level of interest to extract a vibrational quenching rate coefficient. The resulting expression is given by N,O m=u--u (No+,’ - N”+m)Au+m-+ Nu m=l NJ,
c
Expression 14 relates a simple sum of experimentally measured vibrational populations and transition probabilities to the N 2 pressure in a linear manner. The model assumes a knowledge of kEufor each level, for which we substitute our averaged value obtained from the Stern-Volmer analysis presented previously. With this assumption we can obtain a vibrational quenching rate coefficient for each bound vibration level. The predissociation of levels above u = 6 provides a convenient upper limit for the sums of expression 14. Plots of the left-hand side of expression 15 versus N2 number density are shown in Figure 8 for vibrational levels 5 and 6. These levels are least effected by cascade from higher levels. For each of these levels as well as levels 0-4, the data clearly define a straight line. The intercepts for all the lines are generally larger than 1 by more than the expected experimental error. The intercepts are as follows: 1.91 f 0.18, u = 0; 1.88 f 0.12, u = 1; 2.20 & 0.15, u = 2; 2.67 0.34, u = 3; 1.35 d~ 0.32, u = 4; 1.07 0.21, u = 5; 1.32 f 0.25, u = 6. The errors quoted are l a limits obtained solely from the fit to the data. This could possibly indicate the presence of an additional source term, perhaps from the a’ state, or is perhaps due to propagation of experimental errors through the analysis. Both explanations would predict larger deviations from a unit intercept for low vibrational levels: as a result of near resonances in the a’-a vibrational levels for low c (see Figure l ) , and to the extended sums required for the calculation of rate coefficients for the lower levels.
*
45t
[N2](
d4MOLECULES ~ m - ~ )
Figure 8. Vibrational relaxation analysis decay plots for vibrational levels 5 and 6 .
TABLE IV: N2(a,v+v-1) Rate Coefficients U 10”k,,,-,, cm3 molecule-I 2.7 f 0.8 4.0 f 1.0 10.7 f 1.1 26.6 f 1.5 23.0 f 1.2 11.5 i 1.3
SKI
The rate coefficients derived from the analysis are given in Table IV. The largest single source of error in these rate coefficients is the uncertainty in the radiative transition probabilities, especially for the high vibrational levels where coupling should be small. Of course the values are at least uncertain by the magnitude of the electronic quenching rate coefficient, which may not be constant with vibrational level. These uncertainties are included in the overall estimate of experimental accuracy.
Discussion On the basis of our analysis of the fluorescence from electron-irradiated N 2 in the 0.2-70-mTorr pressure range, we are able to (a) quantify the relative production efficiencies into several electronic states as well as the vibrational distributions within those states and (b) measure effective vibrational quenching rates for N2(a,u’=O-6). We will discuss these findings in relation to upper atmospheric fluorescence observations. We will also consider in this section the implications of this quenching on atmospheric emission signatures in the UV-vis and infrared regions. At our lowest pressures, losses are dominated by radiative decay and will accurately reflect the production rates. Under these conditions, the steady-state populations of the N2(C) and N2(a) states are given by eq 7 and 8. Using k,C = (37 ns)-’ and k: = (80 ~ s ) - (ref ’ 27 and 29, respectively), we obtain kfa/kp = 3.1, where kfa includes the sum over all observed vibrational levels (3-6. Inclusion of production into levels above L;’= 6, which predissociate, would raise this value by 1395, to 3.6. Use of a slower radiative rate23would increase this ratio further. C a r t ~ r i g h t ~ ~ uses a value of 2.1 for this ratio in his model of the auroral upper atmosphere. Green and c o - ~ o r k e r s ~calculate ~ . ~ ’ a value of 2.64 for primary excitation only. We have observed emission only from (33) Cartwright, D. C. J. Geopfiys. Res. 1978, 83, 517-531. (34) Piper, L. G.; DeFaccio, M. A.; Rawlins, W. T. J . Pfiys. Cfiem. 1987, 91, 3883. (35) Cartwright, D. C. Phys. Reu. A 1977, 16, 1041. (36) Peterson, L. R.; Sawada, T.; Bass, J. N.; Green, A. E. S. Comp. Pfiys. Commun. 1973, 5, 239-262. (37) Jackman, C. H.; Garvey, R. H.; Green, A. E. S. J . Geophys. Res. 1977, 82, 5081.
3436 The Journal of Physical Chemistry, Vol. 92, No. 12, 1988 the N,(a’,u’=O) level, even though Cartwright’s predictions indicate that over 99% of the a’-state excitation is into higher vibrational levels at altitudes representative of the low pressures employed in our experiments. Cartwright shows excitation peaking at L.‘ = 7 (from a Franck-Condon type process); however, levels above this may be predissociative. Our data indicate that the a’ and a states are strongly coupled, a) at low pressures and colradiatively and collisionally (a’ lisionally (a’ a) at higher pressures. Consequently, the sum of the a and a’ formation rates (relative to the C state) may be more appropriate to compare with Cartwright’s excitation cross sections. Cartwright’s Table 133yields ( k ; + k?’)/kfC = 3.42, compared to our corrected value of 3.6. Hence, our results are in agreement with Cartwright’s values. Our calculations of the radiative lifetimes of the a’ state indicate that there is a strong dependence on vibrational level. Under collisionless conditions radiation from the lowest level of the a’ state is dominated by emission to the ground state. For increasingly higher vibrational levels, a’-a emission initially competes with and eventually dominates emission due to the a’-x transition. This implies that under collision-free conditions the primary deexcitation mechanism for these states is radiative relaxation of the a’ state to form the a state followed by rapid emission of the a-X LBH bands. However, the lower vibrational levels of the a’ state are dominantly collisionally coupled to the a state at all pressures encountered in our experiments on the basis of rate coefficients measured by van Veen et al.** and Piper.34 They measured a coupling rate coefficient for a’(0) to a(O), a process that is 1212 cm3 molecule-l s-I. This rate cm-’ endoergic, to be 2.3 X should reflect a reasonable lower bound for the more resonant and exoergic transfer rates for higher levels of the a’ state. The clear absence of emission from a’-state levels greater than u’ = 0 supports this contention that the collisional rate dominates the emission rate for a’ levels greater than zero. This statement appears to be true even when the lower limit to the coupling coefficient discussed above is used in model calculations. Our measurement of the a-state total emission intensity is in good agreement with Cartwright’s predictions; however, the details of the intra- and intersystem relaxation processes must be determined by more specific laser excitation measurements. In addition to transfer from the a’ state, cascade excitation of the alH, state from higher energy electronic states must be considered. Emission from the cq/l&+ u’ = 0 state into L.’= 1 of the a l H g state is observed in the midst of the Nz(C-B) Au = 2 sequence at 296.7 nm (see Figure 5). This state has been observed to have the largest optical emission cross section of all the Nz electronic states;38however, the primary radiative loss mechanism is through the strongly allowed cl’Z,,+-XIZg+ transition. Hence, the branching ratio between emission on this transition relative to the cq/-a transition will determine the cascade contribution to the total a-state excitation rate. Measurement by Filippelli et have established that the branching ratio for emission from these to channels (cd-a/c,‘-X) has an upper limit of 0.006. When coupled to Cartwright’s relative electron excitation cross section for the c i state,35 cascade from the cq/ state is calculated to contribute substantially less than 1% of the total excitation of the a’n, state. These results are confirmed by the electron and optical emission cross sections for the a’& state measured by Ajello and S h e m a n ~ k y They . ~ established an upper limit on the cascade contribution of 5% based on the close agreement between the two cross-section measurements. Radiative cascade from other high-lying Rydberg states cannot be totally precluded. However, they do not appear to be a significant source of excitation of the a state. Vibrational Redistribution. The N2(a)-statevibrational distributions recorded as a function of pressure clearly show that extensive vibrational redistribution is occurring in addition to a slower electronic deactivation process. Our goal in conducting
-
-
Marinelli et al. the various kinetic analyses presented is to place bounds on the rate coefficients wherever possible so as to guide later laser-based selective excitation experiments. The electronic quenching analysis for the Nz(a) state provides a lower bound on the true average electronic quenching coefficient, since total cascade from the a’ state is not considered as an additional population mechanism. Recent measurements by Piper and c o - ~ o r k e rsuggest s ~ ~ that pure electronic quenching of the a’ state by N, is quite slow (1.9 X cm3 molecule-I s-I), and hence much of the excitation in the a’ state should decay via coupling to the a-state manifold. The measurements of Piper et al. for quenching of the a’ state are supported by the coupled relaxation measurements of the a state reported by van Veen and co-workers.22 In their two-photon excitation measurements of the a state at high pressures, they observed quenching of the fluorescence from the coupled a’-a cm3 molecule-’ system to have a rate coefficient of 2.3 X SS’. The only a’-state level for which we observe fluorescence is a’(O), which lies 1212 cm-’ below a(0) and is the lowest lying excited singlet vibrational level. This level is poorly coupled to the a state, and hence we observe its emission in our experiments. The averaged electronic quenching rate coefficient for our experiments also compares reasonably well to the low-pressure quenching results of van Veen and co-workers.z2 Their data on relaxation of N2(a,u=0,1) at pressures from 0.05 to 0.200 Torr give nearly identical rate coefficients of 2.1 X lo-” cm3 molecule-’ s-l, which are about a factor of 2 faster than our averaged values. While this result is significantly higher than our data would indicate, it should be stressed that our result is essentially an average over all of the observed vibrational levels and level-to-level variations in the relaxation rate may be present. The apparent deactivation rates observed for the a-state increase by nearly a factor of 10 over the range of vibrational levels studied. Golde and Thrush’* in a steady-state flow reactor study observed a similar apparent order of magnitude enhancement with vibrational level. Their observations were made at Torr pressures where substantial a-a’ coupling occurs.22 The present observations in the milliTorr regime provide a more accurate determination of the deactivation rates. Vibrational relaxation rates for vibrational levels u = 4, 5 , and 6 have been obtained because predissociation at levels higher than L. = 6 greatly reduces cascade into level 6 and because the high vibrational levels of the a’ state efficiently couple to predissociated levels of the a state. As a result, the pressure dependences of the population of a-state levels 4, 5, and 6 will be described by eq 14. The deactivation rates we obtain are near gas kinetic for the higher levels and are significantly faster than what would be expected for a simple vibrational-translational relaxation process. The observation of these accelerated rates alone argues that resonant electronic energy transfer between the a and a’ states must account for a significant component of the observed relaxation process. Furthermore, the fact that near gas-kinetic relaxation rates are observed argues that collisional cascade into these upper levels does not occur to a significant extent. If collisional cascade was important, then relaxation rates larger than gas kinetic would be required to obtain the apparent rates measured in the laboratory. As a result of intersystem cascade from the a’ state to the a state, the rate equation analysis leading to eq 15 is incomplete. For a-state vibrational levels u = 0, 1, 2, and 3 the vibrational relaxation rate coefficients cannot be accurately obtained from our data. The theories for intersystem electronic relaxation where both collision partners have internal degrees of freedom are poorly developed, and few conclusions can be drawn about the specifics of the relaxation process from the available data. Simple scalings with vibrational energy level, energy defects with collision partners, or Franck-Condon factors between initial and final states do not seem to readily fit our data. As a result a more detailed comparison with theory does not seem warranted. Conclusions
(38) Zipf, E. C.; McLaughlin, R.W. Planet. Space Sci. 1977,26,449-462. (39) Filippelli, A . R.; Chung, S.; Lin, C. C. Phys. Rev. A 1984, 29. 1709.
The global deactivation rate for loss of Nz(aln,) molecules by electronic quenching has been experimentally determined to be
J . Phys. Chem. 1988,92, 3431-3440 greater than or equal to 8.9 X cm3 molecule-’ s-l. Feed into the a state from higher electronic states does not appear to be significant on the basis of spectral surveys and N2(C) state normalization. A substantial redistribution of the vibrational populations within the a state as a function of pressure is observed. Total quenching rate coefficients for vibrational levels 4-6 approach gas kinetic values. Intersystem energy transfer from the N2(a’12[) state to lower vibrational levels of the a state cannot be distinguished from intrasystem vibrational relaxation within the a-state manifold. Future state-specific laser-excitation ex-
Reductive Quenching of Ru( bpy):’
3437
periments are now under way to distinguish these competing pathways.
Acknowledgment. We acknowledge useful discussion with L. G. Piper and G. E. Caledonia of PSI and R. Huffman of the Air Force Geophysics Lab. This work was supported by the U S . Air Force Office of Scientific Research under Task 23 10G4 and the The Defense Nuclear Agency under Project SA, Task SA, Work Unit 115. Registry No. N2, 7727-37-9.
at High Pressures
Monty L. Fetterolf and Henry W. Offen* Department of Chemistry, University of California, Santa Barbara, California 931 06 (Received: October 2, 1987)
The reductive quenching of R~(bpy),~+ by several aromatic amines was studied at high pressures (0.1-300MPa) in acetonitrile (CH3CN) and n-butyl alcohol (n-BuOH). The luminescence lifetimes are lengthened with increasing pressures. This results in positive activation volumes for the quenching rate constants k, ranging in value from =1 mL/mol for dimethylaniline quencher in CH3CN to 13 mL/mol for benzidine in n-BuOH at 25 OC. The pressure dependence of k, for Ru(bpy),*+/DMA in CH3CN was insensitiveto temperature in the 15-45 OC range. The pressure results are discussed in terms of the commonly accepted mechanism for electron transfer in the activation- and diffusion-controlled limits.
Introduction
Photoinduced electron transfer is an important mechanism for luminescence quenching.’v2 Both oxidative and reductive quenching have been studied with excited ruthenium metal comp l e ~ e s . ~In, ~this high-pressure study of luminescence quenching we focus on the tris(2,2’-bipyridine)ruthenium(II) complex paired with several electron donors Q chosen from related aromatic amines. The excited-state reaction in eq 1 has been shown to lead * R ~ ( b p y ) , ~++ Q
A Ru(bpy)3+ + Q+
to the typical plot of In k, against the quencher reduction potentials, in which a diffusion-limited plateau in the highly exergonic region precedes the predicted linear drop-off with increasingly positive activation barriers3-’ The reductive quenching of * R ~ ( b p y ) by ~ ~neutral + organic molecules leads to products of the same charge, so that back electron transfer to form the excited precursor complex can be ignored in the mechanistic description.8 The ordinary behavior of rate vs free-energy change for these reactants in both the diffusion- and activation-controlled regimes makes them instructive examples for pressure studies of bimolecular quenching reactions. A prior pressure study of photoinduced electron transfer of * R ~ ( b p y ) , ~with + a series of (charged) metal complexes revealed a complex pressure response of k,.9 These results could not be (1) Kavarnos, G. J.; Turro, N. J. Chem. Reu. 1986, 86, 401. (2) Marcus, R. A.; Sutin, N. Biochim. Biophys. Acta 1985, 811, 265. (3) Meyer, T.J. Prog. Inorg. Chem. 1983, 30, 389. (4) Whitten, D. G. Acc. Chem. Res. 1980, 23, 83.
(5) Ballardini, R.; Varani, G.; Indelli, M. T.; Scandola, F.; Balzani, V. J. Am. Chem. SOC.1978, 100, 7219. (6) Bock, C. R.;Connor, J. A,; Gutierrez, A. R.; Meyer, T. J.; Whitten, D. G.; Sullivan, B. P.; Nagle, J. K. J. Am. Chem. SOC.1979, 101, 4815. (7) Sandrini, D.; Maestri, M; Belser, P.; von Zelewsky, A.; Balzani, V. J. Phys. Chem. 1985, 89, 3675. (8) Kitamura, N.;Obata, R.; Kim, H.-B.; Tazuke, S. J . Phys. Chem. 1987, 91, 2033. (9) Ueno, F. B.; Sasaki, Y.; Ito, T.; Saito, K. J . Chem. SOC.,Chem. Commun. 1982. 328.
interpreted on the basis of electrostatic effects, which is not surprising when compared to the more numerous studies of ground-state electron-transfer reactions.IO An earlier report demonstrated that diffusion-controlled quenching between neutral organic molecules is diminished under pressure, as expected from the pressure-induced increase in solvent viscosity.” In contrast, the quenching of *Ru(bpy);+ by the neutral complex C ~ ( a c a c ) ~ , for which k, is less than the diffusional rate constant, resulted in an increase in k, with pressure.12 A recent study of the Mo6C114z-/IrC162- system suggests that the volume change associated with the precursor complex may be an important factor in pressure-induced changes of electron-transfer processes. l 3 We report on several *Ru(bpy),2+/amine systems in acetonitrile and n-butyl alcohol at elevated pressures and discuss the quenching results in terms of the classical theory of electron transfer. The effects of solvent polarity and pressure provide an interesting comparison for these bimolecular quenching reactions. Experimental Section
The solvents acetonitrile (CH,CN) from Burdick-Jackson (high-purity grade) and n-butyl alcohol (n-BuOH) from Aldrich Co. (Gold Label) were stored under nitrogen once opened. The source. of tris(2,2’-bipyridine)ruthenium dichloride was G. F. Smith Chem. Co. The five quencher molecules were used as received: N,N-dimethylaniline (DMA), N,N-diethylaniline (DEA), and N,N,N’,N’-tetramethylbenzidine (TMB) came from Aldrich Co., N,N-dimethyl-p-toluidine (DMpT) was obtained from Kodak Chemicals, and the free base benzidine (B) arrived as an Isopac from Sigma Chemical. The high-pressure equipment and operation have been described earlier.I4 The lifetime station15employs a nitrogen laser to excite (10) van Eldik, R. In High Pressure Chemistry and Biochemistry; van Eldick, R., Jonas, J., Eds.; Reidel: Dordrecht, 1987; p 333. (11) Turley, W.D.;Offen, H. W. J . Phys. Chem. 1984, 88, 3605. (12) Kirk, A. D.;Porter, G. B. J. Phys. Chem. 1980, 84, 2998. (13) Tanaka, H.D.; Sasaki, Y.;Saito, K. Sci. Pap. Imr. Phys. Chem. Res. (Jpn.) 1984, 78, 92.
0022-3654/88/2092-3431$0~.50/0 0 1988 American Chemical Society
