4292
J. Phys. Chem. 1983, 87, 4292-4294
cluster model employed here. . Acknowledgment. Research out at Brookhaven National Laboratory under contract with the U.S. Department of Energy and supported by its Office of Basic Energy Sciences. We acknowledge helpful discussions with Professor G. A. Jeffrey who suggested using the concept
of bifurcated hydrogen bonding to devise a model for dielectric re1axat:on in ice I. We-also benefitted from discussions with Professor J. Nagle. Professor N. R. Kestner kindly consented to inclusion of material from ref 24 prior to publication. Registry No. Water, 7132-18-5.
Subnanosecond Time-Resolved Optlcal Absorption Studies of Electron Solvation in Ice John M. Warman,’ Marlnus Kunsi, It7t8fUt7/VWSity Reactor
1t7St&d8, h&k81W8g
15, 2629 JB Delft, The Netherlands
and Charles D. Jonah Chemistry Divislon, Argonne National Laboratory, Argonne, Illinois 60439 (Received: August 23, 1982)
Optical absorption studies of pulse-irradiated ice have been carried out with subnanosecond time resolution. A growth in the solvated electron absorption is observed with a risetime of 450 ps at -5 “C. The activation energy for relaxation of the matrix following localization is 0.3 eV. Below -35 “C localization at trapping sites M NH4F considerably increases the overall becomes the rate-controlling step in solvation. Addition of solvation time by reducing the initial rate of dry electron localization at defects. A second absorbing species is produced immediately in the pulse with a yield GtW = 9 x 10 M-3 cm-’ (100 eV)-’, which is independent of temperature. This is tentatively ascribed to an exciton state.
Introduction It has been known for many years that electrons formed on irradiation of crystalline ice can become trapped in the medium to give a ”solvated” state with a strong, broad absorption with a maximum at approximately 680 nm.ld The shape of the absorption is very similar to that found in liquid water and the wavelength of maximum absorption is found to change continuously with temperature with no evidence for a discontinuity at the melting point. The structure of the fully relaxed, “solvated” state of the electron would appear therefore to be the same in both phases of H20. At present a polarized cavity model is favored for the solvated state with the electron occupying a vacant position in the H 2 0 lattice and dipole orientation and bulk polarization providing the energy of solvation. Direct information on the dynamics of the solvation process in water has to date proven elusive due to the extreme rapidity with which it occurs. Thus, repeated attempts to observe a growth in the absorption of the solvated state in aqueous systems with increasing time response of detection have only ever led to upper limits for the solvation time. The present upper limit for water at room temperature is 0.3 ps determined in a recent laser flash photolysis s t ~ d y . Even ~ in ice, for which the di-
electric relaxation time is lo6 times longer than in water, no growth of the electron absorption has been seen in experiments with a time resolution on the order of 10 ns.46 With fast conductivity detection techniques it has, however, proven possible to measure the decay rate by localization of highly mobile, conduction-band electrons in ice following pulsed i o n i ~ a t i o n . ~The , ~ mean lifetime of these “dry” electrons has been found to be approximately 100 ps and 1 ns at -5 and -35 “C, respectively. Since full solvation cannot occur more rapidly than the initial localization step, these times must give lower limits to the time necessary for the overall solvation process. A growth of the optical absorption in ice would therefore certainly be expected to be observed if measurements could be carried out with subnanosecond time resolution. A question of particular interest which could be answered by such measurements is the following: which is the rate-controlling step in solvation, the finding of a suitable defect site at which dry electrons become localized or the subsequent relaxation of the surrounding lattice? In the present paper we describe results of optical pulse radiolysis experiments on ice samples with 25-ps duration pulses of irradiation and picosecond time resolution detection which have allowed us to answer this question. A preliminary report has already been published.’0
(1)V. N. Shubin, V. A. Zhigunov, V. I. Zolotarevsky, and P. Dolin, Nature (London), 212, 1002 (1966). (2)I. A. Taub and K. Eiben, J. Chem. Phys., 49,2499 (1968). (3) G. Nilsson, H. Christensen, P. Pagsberg, and S. 0. Nielsen, J. Phys. Chem., 76,lo00 (1972). (4)G. V. Buxton, H. A. Gillis, and N. V. Klassen, Chem. Phys. Lett., 32. 533 (1975):Can. J . Chem.. 55. 2385 11977). (5)H. A. Gillis, G. G. Teather,’and C. K. Ross, J . Phys. Chem., 84, 1248 11980). (6)‘G. Nilsson and P. Pagsberg, Chem. Phys. Lett., 74,119 (1980). (7)J. M.Wiesenfeld and E. D. Ippen, Chem. Phys. Lett., 73,47 (1980).
Experimental Section The measurements were carried out by using a 25-ps pulse of 20-MeV electrons from the ANL linear accelerator (8) J. M. W m a n , M. P. de Haas,and J. B. Verberne, J. Phys. Chem., 84,1240 (1980). (9)M.P. de Haas, M. Kunst, J. M. Warman, and J. B. Verberne, J . Phys. Chem., this issue. (10)J. M.Warman and C. D. Jonah, Chem. Phys. Lett., 79,43 (1981).
0022-3654/83/2087-4292$01.50/00 1983 American Chemical Society
The Journal of phvsical Chemlstty, Vol. 87, No. 21, 1983 4293
Electron Solvation in Ice I
1
I
I
I
1
I
I
1
I
I
I
I
I
0 0
WAKLENOTH (nml
0 0
0.5
1.0 Time (nanoseconds)
I 1.5
10
20
30
40
Time (nr)
I
2.0
Figure 1. The time dependence of the yield X extinction coefficient product at 660 nm resulting from pulse Irradiation of Ice at -5 " C 0, experimental points: - - -, calculated dependence if localization rats controlling step in solvated electron formation: -, calculated for a mean relaxation time of 450 ps following localization. Insert: Absorption spectra at -5 "C taken at endof-pulse, 0, and 500 ps after the pulse,
Figure 2. The time dependence of the yield X extinction coefficient product at 660 nm resutting from pulse irradiation of ice at the temperatures (in "C) shown. The lines are calculated by using the parameters glven in Table I .
TABLE I : Electron Localization and Solvation Data for H,O Ice
+.
(dose per pulse = 1 krd). Optical absorption transients were detected with Cerenkov light and the stroboscopic technique reported previously." For kinetic measurements the absorption was measured at 660 nm. The samples, made from triply distilled deionized water, were contained in 1-cm square Supracil cells and were degassed on a vacuum line prior to freezing by lowering at 1cm/h into a bath at -10 "C. A windowless, fast flow, gas coolant cryostat was used for temperature control. Dosimetry was carried out by measuring the absorption of the solvated electron in water under the same conditions of irradiation and cell geometry as for the ice samples. The end-of-pulse value of Ge for water was taken to be Gem = 5.63 X lo4 M-' cm-' (100 eV)-'. From the known dosimetry the values of Gtxwere calculated for the ice samples. The parameters G and tk are respectively the yield of a given absorbing product in molecules per 100-eV absorbed and the decadic extinction coefficient of that product. For more than one absorbing product GeX= CiGiEj.
Results and Discussion
In Figure 1is shown the transient absorption at 660 nm resulting from pulse irradiation of ice at -5 "C. In the insert to the figure the absorption spectra immediately following the pulse and at a time of 500 ps are given. A growth in the absorption over a few hundred picoseconds is clearly visible but in addition a relatively large absorption corresponding to Gem 9 X 103M-' cm-' (100 eV)-' is found to be present immediately at the end of the pulse. Since the localization of electrons at -5 "C has been measured to take 100 pss,s it is apparent that this "immediate" absorption must be due to the formation of a species, denoted X, other than the solvated electron. The separate nature of the species is substantiated by experiments at lower temperatures and by the effect of NH4F as will be mentioned later. (11) C. D.Jonah, Reu. Sci. Insstrum., 46, 62 (1975).
10 6 2.5 1.0
-5
- 11 - 20 - 29
2.2 1.6 1.0 0.7
17.5 12.3 10.5 7.2
If we take into account species X, the optical data are found to be well described by the following reaction scheme: H20 e-
--
e- 9 H30+, x
+ H30+-!%
X
H20 + H
(2)
(4)
kX
(1)
products
(5)
From conductivity measurements the values of k R and kTNT are known8p9and the value of k x is found to be 1.2 X lo9 s-l from measurements at lower temperatures where the electron yield is considerably lower, see Figure 2. Two further parameters are required in order to fit the absorption transient using the above scheme, namely, the initial yield X extinction coefficient product for electrons Go(-)em and the rate constant kmlfor the lattice relaxation process, (4). Since the overall rate of solvation, kWl,is controlled by (3) and (4) in series it is given by 1 -- 1 1 _ ksol ~ T N T krel If the rate of relaxation is taken to be much greater than the rate of localization, Le., kTNT
