J. Phys. Chem. 1988,92,6958-6962
6958
small separations, e.g., R S 1.7 A). This strong cos3 0 anisotropy (dominant in the field of microwave pressure broadening) may turn out to be quite important in determining the reactive asymmetry effect.
Acknowledgment. We thank Dr. Y. M. Engel for assistance with the plotting of Figures 1 and 2. This work was made possible
by a grant from the U.S.-Israel Binational Science Foundation, Jerusalem, Israel. R.B.B. appreciates support from the National Science Foundation via Grant C H E 86-15286. The Fritz Haber Research Center is supported by the Minerva Gesellschaft fur die Forschung, mbH, Munich, FRG. Registry NO. HCN, 74-90-8; HCl, 7647-01-0.
Factors Affecting Charge Separation and Recombination in Photoexcited Rigid Donor-Imiator-Acceptor Compounds Michael N. Paddon-Row,* Anna M. Oliver: Department of Chemistry, University of New South Wales, P.O. Box I , Kensington N.S.W. 2033, Australia
John M. Warman,* Kenneth J. Smit, Matthijs P. de Haas, Interuniversity Reactor Institute, Mekelweg 15, 2629 JB Delft, The Netherlands
Henk Oevering, and Jan W. Verhoeven* Laboratory for Organic Chemistry, University of Amsterdam, Nieuwe Achtergracht 129, I018 WS Amsterdam, The Netherlands (Received: September 28, 1987)
The dipolar transients formed on photoexcitation of a series of molecular assemblies consisting of a dimethoxynaphthalene donor and a dicyanoethylene acceptor separated by rigid, nonconjugated hydrocarbon bridges have been investigated by using time-resolved microwave conductivity (TRMC). The edge-to-edge donor to acceptor separation, 4,is varied from 4.6 to 13.5 A in steps of approximately 2.3 A. The dipole moments of the corresponding excited-state intermediates increase from 26 to 77 D. The lifetimes of the charge-separated states increase markedly with increasing separation, from 1.0 to 740 ns in benzene. The lifetimes in pdioxane are in general approximately an order of magnitude shorter than in benzene. The lifetimes in cyclohexane are approximately an order of magnitude longer than in benzene for the two shortest compounds. The lifetime toward direct charge recombination to the ground state is quite well described by an exponential dependence 1.3 X lo-", and 1.1 X IO-'* s for cyclohexane, on R,Le. TCR = A exp(&/a) with a = 1.13 A for all solvents and A = 1.3 X benzene, and dioxane, respectively. For & = 9.4 A in cyclohexane and 13.5 A in benzene back electron transfer resulting in indirect charge recombination via the local excited donor state begins to dominate the overall decay kinetics. This is accompanied by the appearance of delayed donor fluorescence.
Introduction In all discussions or theoretical treatments of the site to site transfer of electrons the distance over which transfer must take place plays a crucial role. It is only recently, however, that serious attempts have been made to bring this parameter under experimental control by deliberately fming the relative positions of two, or more, electron exchange centers. This has been achieved using rigid, nonconjugated hydrocarbon frameworks as spacers or bridging elements.l-12 To date the number of systematic investigations of the effect of separation distance on electron transfer is, however, still very limited. In the present work we have investigated photon-induced charge separation and subsequent recombination in a series of compounds which were specifically designed and synthesized with the aim of gaining a better insight into the factors controlling long-distance electron transfer, in particular the effect of distance. The series consists of the five compounds shown in Figure 1, all of which have a dimethoxynaphthalene donor moiety and a dicyanoethylene acceptor held rigidly in a fixed orientation and at a fixed distance with respect to each other by an intervening hydrocarbon bridge. The bridge separating donor and acceptor varies from 4a bonds for the shortest compound in steps of 2 up to a maximum of 120 bonds which results in an edge-to-edge distance in the range from approximately 5 to 14 A.13 The main technique that we have applied to probe the electron-transfer process in the present work is that of nanosecond time-resolved microwave conductivity (TRMC).I4 This can 'Australian National Research Fellow.
provide quantitative information on both the extent of intramolecular charge separation and the reaction kinetics of dipolar intermediates which is often difficult to obtain in other ways. Fluorescence measurements are also included where these are helpful in interpretation of the TRMC results. A more limited, preliminary report of the present work has been published12as have studies using optical detection techniques."' (1) Pasman, P.; Rob, F.; Verhoeven, J. W. J. Am. Chem. SOC.1982,104, 5127. (2) Mes, G. F.; de Jong, B.; van Ramesdonk, H. J.; Verhoeven, J. W.;
Warman, J. M.; de Haas, M. P.; Honrnan-van den Dool, L. E. W. J. Am. Chem. Soe. 1984, 106,6524. (3) Leland, B. A.; Joran, A. D.; Felker, P. M.; Hopfield, J. J.; Zewail, A. H.; Dervan, P. B. J. Phys. Chem. 1985,89, 5571. (4) Miller, J. R.; Calcaterra, L. T.; Closs, G. L. J. Am. Chem. SOC.1984, 106. . - -,3047. - - .. . ( 5 ) Warman, J. M. Nature 1987, 327, 462. (6) Wasielewski, M. R.; Niemczyk, M. P. J . Am. Chem. SOC.1984, 106, 5043; 1985, 107, 1080, 5562. (7) Heitele, H.; Michel-Beyerle, M. E. Chem. Phys. Leu. 1987, 134, 273. (8) Bolton, J. R.; Ho,T. F.;Liauw, S.;Siemiarczuk, A.; Wan, C. S.K.; Weedon, A. C. J . Chem. Soc., Chem. Commun. 1985, 559. (9) Hush, N. S.;Paddon-Row, M. N.; Cotsaris, E.; Oevering, H.; Verhocven, J. W.; Heppener, M. Chem. Phys. Left. 1985, 117, 8. (10) Verhoeven, J. W.; Paddon-Row, M.N.; Hush, N. S.;Oevering, H.; Heppener, M.Pure Appl. Chem. 1986,58, 1285. (1 1) Oevering, H.; Paddon-Row, M. N.; Heppener, M.; Oliver, A. M.; Cotsaris, E.; Verhoeven, J. W.; Hush, N. S.J . Am. Chem. SOC.1987, 109, 3258. (12) Warman, J. M.; de Haas, M. P.; Paddon-Row, M. N.; Cotsaris, E.; Hush, N. S.;Oevering, H.; Verhoeven, J. W. Nature 1986, 320, 615. (13) Craig, D. C.; Paddon-Row, M. N. Aust. J . Chem., in press. (14) de Haas, M. P.; Warman, J. M. Chem. Phys. 1982, 73, 35.
0022-3654/88/2092-6958$01 .50/0 0 1988 American Chemical Society
Photoexcited Rigid Donor-InsulatorAcceptor Compounds
The Journal of Physical Chemistry, Vol. 92. No. 24. 1988 6959
t
.. . ..... .
.
. .
.....
cyclohexane
0
>r
.-c > .-+ 0
3 -0
benzene
c
0 0
..
CN
al > 0 3
2
.-0
para-dioxane
E
Flyre 1. The molecular structum of the donorinsulatoracceptor compounds used in the present work and referred to in the text by the number of intervening carbansarbon bonds as denoted on the left. Also shown at the bottom are the stmctures used for probing the prop erties of the uncoupled donor and acceptor. The pitions defining the edge-to-edgedistance R, are shown for the 120 bond compound and the values are listed in Table 1.
Experimental Seetion
The rigid donor-insulatoHcceptor compounds (DIADs') used in the present work are shown in Figure 1. Their method of preparation and purification have been described in a separate publication."." The solvents used were cyclohexane (Merck "uvasol"), benzene, andpdioxane (Fluka UV spectrosoopic grade). The solutions were bubbled with CO, gas (Matheson >99.99%) to remove air and to prevent possible spurious transients resulting from the formation of highly mobile electrons which can be formed in low-yield multiple photoionization processes. The optical densities of the solutions were between 0.2 and 1.0 and were measured after bubbling. The donor extinction coefficient at 308 nm is 3500 M-l cm-I. The acceotor is comoletelv transoarent a t this wavelength. The solutions were transferred after deaeration to an X-band microwave conductivity cell the design of which has bem reported fully previously.'4 The solution within the cell was flash photolyzed by using the xenon chloride 308-nm line of an cxcimer laser (Lambda Physik EMC 500 or Lumonics Hyper-X 420). The pulse uidths, fwhm. were between 5 and 8 ns. The pulse shape was repeatedly monitored with a fast pholodiode for use in s u b sequent data handling. The pulse intensity was monitored by using a solution ofpnitroaminobiphenyl (PNAB) in bcn7xne (OD I ) as an internal TRMC actinometer. The value of W X qnc = 1.8 X m' V - L s-I w as used for the product of the change in rotational charge mobility" on excitation, AM.. and the intersystem musing ellicicncy for the formation of the highly dipolar and long-lived (110 ns in an aerated solution) triplet state of PNAB. The primary colorimetric actinometer was a solution of The intensities used were between Abcrchrome 540 in 5 and 15 mJ cm-' per pulse. The X-band (8.2-12.4 CHz) microwave circuitry was identical with that described previously." The response time of detection .
I
-
(15) Paddon-Row, M. N.:Cotsaris, E.: Patney. H. K. Tefrohcdron1986, 42, 1779. (16) Hcller, H. 0.:Langan, J. R.1.Chcm. Soe., Perkin TMN 2 1981.341. (17) Warman. J. M.: de H a s , M.P.: Oevcring, H.:Paddon-Raw. M. N.: Oliver. A. M.: Hush. N. S. Chcm. Phy$. Left. 1986, 128.95.
I
0-
k 50n
s 4
Figure 2. Microwave conductivity transients resulting from flash phctolysis (8 ns, 9 mJ/crn', 308 nm) of ca. l ( r M solutions of the 60 compound in cyclohexane, bcnzene. and p-dioxane. The vertical scales
carrespond to different sensitivities for the different solvents. The data pictorially illustrate the strong solvent dcpendence of the charge-remmbinatian lifetime. was determined by the quality of the microwave cavity and was in the range 4-5 ns depending on the solvent used. The transient changes in microwave power reflected by the cell on flash photolysis were monitored by using a Tektronix 7912 digitizer. The metbod of quantification and fitting of the data has been described fully e1se~here.l~ The procedure takes into account the temporal shape of the pulse, the rise time of the system, and the depth distribution of products within the cell. The fluorescence measurements reported have mostly been reported elsewhere.e" Where new values are given, their method of determination was the same as that reported previously.
Results and Discussion Flash photolysis of solutions of the compounds shown in Figure 1 at 308 nm mults in local excitation of the dimethoxynaphthalene donor group. This has been shown to be followed by rapid transfer of an electron to the dicyanoethylene acceptor. The evidence for electron transfer has k e n provided by fluorescence studies in which both the quenching of the local donor emission and the appearance of long-wavelength charge-recombination hands have been observed.e" The occurrence of charge separation has been further substantiated by the measurement of large, transient changes in the dielectric loss of the flash photolyzed solutions using the time-resolved microwave conductivity (TRMC) detection technique."J' Using TRMC the dipole moment of the transient can be estimated if the lifetime is longer than a few nanoseconds. The previously reported values for the dipole moments of the first four compounds in Figure 1 in benzene solution were found to be in agreement with expectations based on complete charge separation between donor and acceptor." In the present work we have extended the TRMC study to the 12u bond compound and to saturated hydrocarbons and pdioxane as solvents. Rather surprisingly the lifetimes of the dipolar transients have been found to be extremely sensitive to solvent even within this limited sample of low dielectric constant media.
6960 The Journal of Physical Chemistry, Vol. 92, No. 24, 1988
Paddon-Row et al.
TABLE I: The Measured Dipole Momenta of the Tnnsients Formed on Photoexcitation of the Compousds Sbown in Figure 1, together Over the with Calculated Values for FuU C b g e -tion Edge-to-Edge Donor-Acceptor sep.rrrtiOa Distance, R e dimle moments (measd), D solute Re,” A eR,, D cyclohexane benzene p-dioxane 22 26 b b 4.6 4a 33 45 55 65
6.8 9.4 11.5 13.5
6a 8a 100 12a
36 63
37 55 68 77
c c
b b 65 78 0
@Reference13. bToo short lived. ‘Giant dipole not formed.
,“G 0
,
10
R,
TABLE II: Mean Lifetimes townrd chrpe sepuotioP and Charge Recombination for the Comporslds Shown in Figure 1 and the Solvents Listed
charge-separn time,” [CS quantum yield] cycloben- diethyl solute hexane zene ether 40 6a 8a
c d 48
10a
e
12a
e
C
C
d
d 21 196 [0.96] 2703 [0.64]
19 139 [0.97] 1370 j0.751
charge recomb time, ns pcyclo- benhexane zene dioxane c w.5l.f
e
[LO)‘ 6 32 360h
e
74Wh
297h
9 38 288
[2.5l.f 43
*From donor fluorescence decay. Where significantly different from unity, based on donor lifetime in cyclohexane, benzene, and diethyl ether of 5.43,4.13, and 4.82 ns. ‘Too short to measure. d 3 4 ps found for ethyl acetate, THF, and acetonitrile. CNot measurable. /From in-pulse signal dipole moment assumed. ZDelayed donor fluorescence observed. *By extrapolation to zero concentration. This is illustrated by the transient recordings shown in Figure 2 for the 6c compound in cyclohexane, benzene, and p-dioxane. In Table I are given the values of the dipole moments determined from the magnitude of the conductivity signals for those solute-solvent combinations for which the transient lifetimes were 5 ns or longer. For the 120 bond compound in benzene and dioxane account has been taken of the fact that the time for charge separation is comparable with the natural lifetime of the local excited donor.lOJ1The quantum efficiency for charge separation for the 12u compound can therefore no longer be assumed to be unity as for the shorter compounds. The values of the quantum efficiencies obtained from optical studies of donor quenching are given in Table I1 together with the lifetimes toward charge separation, TCS. The quantum efficiency for charge separation in dioxane was taken to be equal to the value found for diethyl ether since no optical measurements were available for dioxane itself. The value of the dipole moment is derived from the measured change in the microwave conductivity (dielectric loss) of the medium by using the relationship
In ( l ) , (i is the dielectric constant of the medium, N is the concentration of the dipolar intermediate with dipole moment p and rotational relaxation time T,, w is the radian microwave frequency, kB is the Boltzmann constant, and T is the temperature. The rotational relaxation times were obtained from the semiempirical relationship18 1.9 x 10-8?~3 7, e (2) T[ln ( 2 L / D ) - 0.51 with L = 2.12 X 10”M/pDZ
(3)
which has been found to describe quite well (f20%) known re~~
~
~~
(18) Warman, J. M.; de Haas, M. P. Interuniversity Reactor Institute Report No. 134-86-07, 1986.
I
,
5
15
(a)
F i e 3. The dipole moments experimentally determined in the present work for the charge-separated states of the compounds shown in Figure 1 as a function of the edge-to-edgeseparation distance between donor and
acceptor. The values were determined in the solvents cyclohexane (0); benzene (0); and dioxane (A). The full line corresponds to the dipole moment expected for complete charge separation over the edge-to-edge donor-acceptor distance. laxation data for rod-shaped dipolar molecules in the present solvents. In (2) and (3), q is the solvent viscosity in centipoise, L and D are the effective molecular length and diameter in nanometers, M is the molecular weight, and p is the effective density of the solute molecule in g ~ m - ~A .constant value for D of 0.55 nm is taken for all systems. The values of 9 and p are as follows: for cyclohexane 1.00 and 1.20; for benzene 0.65 and 1.06; and for dioxane 1.45 and 1.06. It is worth pointing out that any errors or uncertainties in the value of 7, are approximately halved in the eventual dipole moment derived since the measured signal is proportional to p2/7,. This also applies to uncertainties in the concentration or quantum yield of the dipolar intermediate. For the present solutes the termf(w7,) in (I), which for a Debye type relaxation is given by (4) can be taken to be equal to unity, i.e., 0 7 ~>> 1, for the microwave frequencies used (ca. 10 GHz). Where it was possible to measure the dipole moment of a given DIAD in more than one solvent the agreement between solvents was found to be good. This is shown by the data in Table I and is illustrated in Figure 3 where the dipole moment values are plotted against edge-to-edge distance in the molecule. The full straight line in Figure 3 corresponds to complete charge separation over a distance equal to the edge-to-edge distance between donor and acceptor, 4,as indicated in Figure 1. The fact that the dipole moment data lie in general above this line is confirmation that quenching of the donor fluorescence does in fact occur via electron transfer to the acceptor to produce a completely chargeseparated, “giant-dipole” state in all cases. The dashed straight line drawn through the points in Figure 3 would su gest that the charge is in fact separated by approximately 1.5 in excess of the edgeto-edge distance. Since.the dipole moment of the chargeseparated state of a given DIAD is independent of solvent it is possible to assume its value for those cases where the lifetime was too short to be measurable directly, as for example in the case the 6a compound in dioxane as shown in Figure 2. Using the fitting procedure described previously,14 including convolution over the laser pulse shape, it was then possible to determine the lifetime of the giant-dipole state necessary to explain the magnitude of an “in-pulse” conductivity transient. These lifetime estimates are given bracketed in Table I1 together with the values measured directly from the after-pulse decays. The decay rates represent the sum for all processes leading to the neutralization of the charge-separated state. Apart from a direct transition to the ground state, as evidenced by longwavelength charge recombination type emissions,” two other deactivation pathways have been found to occur. The first of these was initially apparent as a decrease in the lifetime of the dipolar transient with increasing solute concen-
x
The Journal of Physical Chemistry, Vol. 92, No. 24, 1988 6961
Photoexcited Rigid Donor-Insulator-Acceptor Compounds
10-6f
Figure 4. A reaction schematic of the processes underlying the effects investigated in the present work: (1) photon absorption by the donor
giving the local excited donor, LED, state; (2) LED direct decay to the ground state (accompanied by donor fluorescence); (3) electron transfer from LED to acceptor resulting in charge separation; (4) back electron transfer to give LED; ( 5 ) decay of the charge-separated, CS, state directly to the ground state. tration. This effect, which is particularly marked for the longer lived transients, has been shown to be due to electron-transfer quenching of the giant-dipole state by the neutral solute itself. The unimolecular decay time in such cases must then be obtained by extrapolation to zero solute concentration as described in a previous publication.” The rate constant for self-quenching by the 12a compound in benzene has been found to be 7 X lo9 M-’ s-’. All values given in Table I1 have been corrected for this self-quenching effect which is in fact only significant for lifetimes of a few hundred nanoseconds and longer. The second “complication” in the kinetics of charge recombination is the Occurrence of back electron transfer to re-form the local excited donor. This presents an indirect pathway for deactivation of the charge-separated state via reaction 4 followed by (2) as shown in Figure 4. This indirect pathway makes itself apparent in the form of a delayed local donor fluorescence with a lifetime equal to that of the TRMC transient. Such a delayed donor fluorescence is observed for the 8a compound in saturated hydrocarbons and the 12a compound in benzene. This extra pathway will result in a more rapid deactivation than if only the direct transition to the ground state were operative. If the equilibrium between local excited donor and giant-dipole state is established rapidly compared with the net decay time, then the overall unimolecular rate constant for recombination kCR,would be expected to be given by (5)
The Occurrence of this particular kinetic complication is of great interest since it helps to pinpoint the relative positions, in terms of energy level of the local excited donor and the giant-dipole state since for delayed donor fluorescence to occur the two states must be within a few tenths of an electronvolt of being degenerate. This aspect of the data is being investigated in greater detail with particular attention being paid to the temperature dependence of the kinetics under such conditions. These results will be reported in a future publication. Because of the occurrence of this indirect pathway for the 8a compound in saturated hydrocarbon solvents and for the 12a compound in benzene, these lifetimes will be omitted from the discussion of the distance and solvent dependence of direct charge recombination in the ensuing discussion. In Figure 5 the charge recombination lifetimes involving what is thought to be a direct transition to the ground state are plotted against the number of u bonds in the bridging unit, N,,. A logarithmic presentation of TCR was chosen because of the large range of the individual T values, from the shortest measureable value of 0.5 ns to the longest of 360 ns. Where values of TCR for more than two different No values are available for a given solvent the data are in fact found to obey a linear semilogarithmic dependence on Nuquite well. This is shown by the straight lines drawn through the benzene and dioxane data points. Furthermore the slopes of the straight lines for the two solvents are found to be the same within the accuracy of the measurements. A straight line of the
IO
I6
’
I
I
’
’/
I1
4
0
12
Number of u bonds
Figure 5. The dependence of the lifetime for direct (to ground state)
charge recombination, 7CR,on the number of u bonds separating donor and acceptor for the compounds shown in Figure 1 and for the solvents cyclohexane (0);benzene (+); and dioxane (m). The full lines correspond to a value of b = 1.00 in the expression .OcR = T*CR exp(bNu). same slope is also found to describe reasonably well the distance dependence of the two cyclohexane points. Empirically therefore, the data can be said to be in agreement with a dependence of the direct charge-recombination time given by 7 O C ~= T*CR
exp(6Nu)
(6)
The value of b corresponding to the slopes of the lines in Figure 5 is 1.OO. The value of T*CR is 1.3 X s for cyclohexane, 1.3 X lo-” s for benzene, and 1.1 X s for dioxane. The majority of theoretical treatments predict that the rate of electron transfer between weakly interacting sites separated edge-to-edge by a distance R, should depend exponentially on this distance according to
kd = Vet exp(-aRe)
(7)
The distance R, for the present compounds is taken from the center of that C-C bond of the naphthalene system which is shared with the bridge, to the first carbon atom of the bridge which forms part of the ethylinic bond of the acceptor. This is illustrated for the 12a compound in Figure 1. X-ray structural analysis has revealed a slight curvature of the bridges in these molecules.13 This results in a nonlinear dependence of the through-space value of Re on the number of a bonds, as indicated by the values in Table I. The deviation from the average distance per a bond of 1.13 8, is, however, small. Since b = 1.00 is found to give the best fit in eq 6, the distance dependence of charge recombination can be taken to be represented well by k°CR
= VCR exp(-0.88Re)
with the value of the preexponential frequency factor vCR being strongly dependent on the solvent. The dramatic sensitivity of the charge-recombination kinetics to the nature of the solvent, even within the present group of compounds of low polarity, contrasts with the relative insensitivity of the charge-separation kinetics to the nature of the medium which has been This is illustrated by the charge-separation lifetimes for a selection of solvents, covering a similar range of polarities as the present solvents as evidenced by solvent shifts of other giant-dipole excited states: which are plotted against N,, in Figure 6 . The point at 3 ps for the 60 compound in ethyl acetate determined by picosecond flash photolysis studies of the donor fluorescence decay has not been previously reported. A value of approximately 3 ps has also been found for the solvents tetrahydrofuran and acetonitrile. The solid straight line drawn through the values of the lifetime for charge separation in Figure 6 corresponds to an exponential dependence as given by (6) with the same value of 6 = 1.00 that is found to give the best fit to the recombination data. The value
6962
The Journal of Physical Chemistry, Vol. 92, No. 24, 1988
- 10.' I
u
ld'20
2
4
6
8
NUMBER
IO
OF
12
14
16
0 BONDS
The dependenceof the lifetime toward charge separation, sa, on the number of u bonds separating donor and acceptor for the compounds shown in Figure 1 and for the solvents benzene ( 0 ) ;diethyl ether (+); and ethyl acetate (0). F i g u r e 6.
of T * corresponding ~ to the solid line is 9.1 X s. The forward charge-separation process is therefore seen to be approximately 2 orders of magnitude faster than charge recombination even in the most favorable solvent studied here, Le., dioxane. The relatively small differences which are found between different solvents for the charge-separation process have been discussed in a previous article" and are ascribed to small differences in the free energy barrier for forward electron transfer. The conclusion may be drawn, at least for the present molecular assemblies, that the distance dependence of electron transfer is the same for the forward charge-separation process as for backward charge recombination with an a factor in (7) of 0.88 The much smaller absolute rates for the charge-recombination process can be understood, at least qualitatively, by considering the general theoretical e x p r e s ~ i o nfor ~ ~the ~ ~rate ~ of electron transfer between two states a and b for which the electronic coupling is weak i.e. k,,
2r -Ta2F h
(9)
In (9), Tab is the electronic transition matrix element which is the factor mainly responsible for the exponential dependence of the electron-transfer rate on distance. The term F is referred to as the Franck-Condon factor for the transition. It is related to the energy difference between the two states, Eab, and the molecular and solvent reorganization energies, L, and L,, by'9920
+
In (lo), L = L, L, and 2u2 is the energy dispersion parameter associated with the transition which is assumed to have a Gaussian line shape given by P(E) =
exp[-(E - E a b + L ) 2 / 2 a 2 ] (2ru2)'/2
(11)
with a maximum at E = E a b - L. In the high-temperature limit 2a2 = 4LkBT and (10) reduces to the well-known Marcus expression'*J9
(19) Redi, M.;Hopfield, J. J. J . Chem. Phys. 1980, 72, 6651. (20) Marcus, R. A.; Sutin, N.Biochim. Biophys. Acta 1985, 811, 265.
Paddon-Row et al. For the forward electron-transfer process leading to charge separation, the preexponential frequency factor for the distance dependence, vet, is of the order of magnitude of the maximum value expected. This indicates a Franck-Condon factor close to unity. As discussed in a separate publication" this is to a large extent due to the similarity in magnitude of the driving force for charge separation Eaband the molecular reorganization energy which was estimated to be 0.6 eV, and to compensatory changes in these parameters with increasing distance. For the recombination process, however, Eab is known, from the redox properties of the donor and acceptor%" and from the observed emission maxima of the present compounds, to be approximately 3 eV." This is much larger than the reorganization energy L, and hence according to (10) should result in a very much reduced value of F and a corresponding low rate of electron transfer as is found. In addition it is also known that the effective solvating powers of benzene and dioxane are considerably larger than would be expected on the basis of their dielectric constants alone. This is observed as anomalously large bathochromic shifts in the emission bands of highly dipolar excited states. Recent studies on the emissions of the giant dipole states of other rigidly separated donor-acceptor compounds have shown the shifts in benzene and dioxane to be approximately 0.3 and 0.6 eV, respectively, when compared with a saturated hydrocarbon solvent.l** The much more rapid recombination found in these two solvents for the present compounds can therefore be at least qualitatively understood in terms of a reduction in (Eab- L) and a resulting increase in the exponential energy dependent term in the Franck-Condon factor. At present a program of experiments is being carried out in which the temperature dependence of the recombination rate and its variation in solvent mixtures are being investigated in an attempt to test more thoroughly the quantitative agreement with theoretical models. This data and its consequences will be presented and discussed in future publications.
Conclusion Photoexcitation of donor-insulator-acceptor compounds in which the donor is dimethoxynaphthalene, the acceptor is dicyanoethylene, and the insulator is a rigid u-bonded hydrocarbon bridge results in the formation of completely charge-separated, giant dipole intermediates with high quantum efficiency. Dipole moments up to 77 D have been measured which corresponds to complete charge separation over a distance of more than 15 A. Both the rates of charge separation and charge recombination are found to obey a dependence on edge-to-edge separation between donor and acceptor, Re, given to a good approximation by k,, = vet exp(-0.88Re) with Re in A. The preexponential frequency factor for charge separation is relatively insensitive to solvent and is of the order of the maximum value expected, Le., 10'4-1015 s-l. The value of v, for charge recombination is found to be sensitive to the solvent even within the relatively nonpolar range of cyclohexane, benzene, and p-dioxane for which values of 7.7 X lo9, 7.7 X lolo, and 9.1 X 10" s-l are found, respectively. Two additional routes for deactivation of the giant-dipole state over and above the direct ground-state transition are found. One involves self-quenching via intermolecular electron transfer which occurs at higher concentrations. The other involves a return pathway via the local excited donor state which becomes apparent for the 80 compound in cyclohexane and for the 12a compound in benzene.
Acknowledgment. Support from the Australian Research Scheme (Grant (3315773) to M.N.P.-R. is gratefully acknowledged as is the award of a National Research Fellowship to A.M.O. Registry NO. 4 ~98501-38-3; , 6 ~98501-39-4; , 8 ~ 102827-99-6; , IOU, 102828-00-2; 120, 104078-17-3; cyclohexane, 110-82-7; benzene, 7143-2; dioxane, 123-91-1; diethyl ether, 60-29-7.
