Belloni, Saito, and Tissier
308
(17) K. C. Wang and P. G. Wahlbeck, J. Chem. Phys., 47, 4799 (1967).
(18) R. T. Grimlev and J. LaRue, Rarefied Gas Dvn.. Proc. ht. Svmo.. 5.
1455 (1969):
(19) D. W. Muenow and R. T. Grimley, Rev. Sci. lnstrum. 42,455 (1971).
(20) R. T. Grimley, D. W. Muenow, and J. L. LaRue, J. Chem. Phys., 58, 490
(1972).
(21) R. T. Grimley and L. C. Wagner, J. Chem. Phys., 58,402 (1973).
(22) See paragraph at end of paper regarding supplementary material.
(23) Unpublished work.
(24) T. A. Milne and H. M. Klein, J. Chem. Phys., 33, 1628 (1960).
~.
(25) J. W. Ward, R. N. R. Mulford, and M. Kahn, J. Chem. Phys., 47, 1710
119671.
(26) i. W. Ward, J. Chem. Phys., 47, 4030 (1967).
(27) G. F. Voronin, Zh. Fiz. Khim. 41, 1663 (1967).
(28) R. T. Grimley and T. E. Joyce, J. Phys. Chem., 73, 3047 (1969).
(29) Nat. Stand. Ref. Data Ser., Nat. Bur. Stand. No. 37 (1971).
(30) W. L. Winterbottom, J. Chem. Phys., 51, 56iO (1969).
(31) M. B. Liu and P. G. Wahlbeck, J. Chem. Phys., 59, 907 (1973).
(32) 6. A. Gottwald, Vak-Tech., 22, 106 (1973).
(33) E.A. Gottwald, Vak.-Tech., 22, 141 (1973).
Photodetachment of Electrons from Alcoholate Ions in Liquid Ammonia
J. Belloni,' E. Saito, and F. Tlssier
Laboratoire de Physico-chimie des Rayonnements,Associ.4 au CNRS, 9 1405 Orsay, France, and
DRA, SRiRMA, CEN Saclay, 9 1190 Gif-sur-Yvette, France (Received July 16, 1974)
Publication costs assisted by SRIRMA, DRA, Centre d'Etudes Nuci6aires de Saclay
Solvated electrons with a long lifetime (about 1 hr) are produced by photolysis of solutions of potassium
ethanolate in ammonia. The quantum yield of earn- produced by this photodetachment from the ethanolate anion is found equal to the theoretical yield of one electron formed per photon absorbed (for X 254 or
316 nm), thus indicating the negligible importance of reverse reactions of recombination, particularly in
the cage. In the presence of molecular hydrogen, @ea,,,- = 2 for both wavelengths; the formation of the supplementary earn- being explained by the quantitative reaction of the radical Et0 with H2 under these basic
conditions.
Since the early work by Ottolenghi and Linschitql it has
been shown2 that both radiolysis and photolysis of liquid
ammonia solutions of amide ions in the presence of hydrogen lead to the formation of ammoniated electrons whose
stability is comparable with that of electrons derived from
dissolved alkali metals. Nevertheless, the quantum yields
are small, of the same order of magnitude as those obtained
for photodetachment processes in other solvent^.^^^ This is
due either to a large amount of recombination, possibly favored by a cage effect, between the free radical and the solvated electron or to a rate of deexcitation of the excited
anion much greater than that of dissociation.
A solution of potassium ethanolate in liquid ammonia
showed anomalies a t room temperature which led us to suspect that, even without hydrogen, we should be able to observe the formation of stable solvated electrons by photolysis of this medium and a study of the system is briefly reported here.
The solutions were prepared as described previously2
taking all precautions concerning the purity of the reagents
and the cleanliness of the silica cells, since any impurity increases the rate of disappearance of the solvated electron.
The samples were given a preliminary test with a uv lamp;
if a short exposure gave a blue color, persisting for more
than 10 min, the sample was kept for further experiments.
The photolysis celI which served also for the spectrophotometric measurements (optical path 1.0 cm) was fitted
with a side arm to contain the stock solution. The volume
* To whom correspondence should be addressed at the Laboratoire de Physico-chimie des Rayonnements,Associb au CNRS.
The Journal of Physical Chemistry, Vol. 79, No. 4 , 1975
of the cell was calibrated before filling. The saturated ethanolate solutions were (3.1 f 0.2) X
M and the initial
solution in the cell could be diluted with known amounts of
ammonia distilled from the side arm.
For photolysis, we used a high-pressure mercury lamp
(Hanau Q 600) and interference filters (A 254 nm, AA = 19
nm, transmission T = 0.21; X 316 nm, AX = 16 nm, T =
0.17; A 365 nm, Ah = 4 nm, T = 0.34). Parker and Hatchard's ferrioxalate method5 was used for actinometry.
Spectra were recorded between 220 and 3000 nm with a
Beckman DK 1A double-beam spectrophotometer. For
these, as for the kinetic measurements made a t a selected
wavelength, both sample and reference cells (containing
also ethanolate ions) were kept in the dark to avoid errors
due to stray light.
At room temperature, the ethanolate ion has an absorption spectrum which starts around 330 nm and increases
gradually until it merges with the absorption band of ammonia near 230 nm. As is the case for other ions, this absorption is probably a charge-transfer to solvent band.334 At
250 nm, the extinction coefficient is 2 X lo3 M - l cm-l.
The ethanolate ion does not absorb a t 365 nm and it was
found experimentally that no photolysis occurred a t this
wavelength. Our studies were therefore confined to the
wavelengths 254 and 316 nm.
The spectrum of the solvated electron in liquid ammonia
is a large band covering the range between 600 nm, and beyond 2000 nm with a maximum a t 1750 nm. Depending on
the purity of the sample, the decay of the solvated electron
was more or less slow, in some cases the half-life being
Photodetachment of Electrons from Alcoholate Ions
309
Scheme I
C H 3CH2O
CHSCH20H + H
more than 1 hr. After photolysis, this decay was followed
for each run, thus allowing an extrapolation to the zero
time, i.e., that corresponding to the end of the irradiation.
Knowing the extinction coefficients, the concentration of
solvated electrons corresponding to the absorption of a
known number of photons could be determined and hence
the quantum yield.
In the case of the solutions without hydrogen, the quantum yield is
= 1 up to absorbed energies of 3 X 1019
photons l.-l, the corresponding concentration in ammoniated electrons being 5 X lom5M for both wavelengths 254
and 316 nm. For larger doses up to 8 X 1019 photons 1.-l
= 0.8.
the quantum yield decreases to
In the case of solutions in the presence of hydrogen, the
quantum yield is
= 2 for both lines mentioned above.
The initial yields remain constant as a function of ethanolate ion concentration in the range 10-4-3 X lod2 M.
Discussion
Applying Stein’s scheme of photodetachment of electrons from anions5 to our case of ethanolate ions in liquid
ammonia, we have
hu
EtO-*
EtO-
f
deexcitn
-
(Et0
I
+ e-)
-
Et0
-I-em‘
I
recambinat ion
As we obtain a quantum yield equal to the theoretical yield
of one electron per photon, we may suppose that the reverse reactions are negligible. In particular the deexcitation
of the excited state EtO-* must be very slow compared to
the photodetachment. Moreover, the recombination between the radical and the electron, solvated or not, hardly
proceeds at all, whether they are still in the cage or separated by diffusion. This recombination should be most
probable when the radical and e- remain close to each
other in the cage, but this seems not to hold in the case of
photodetachment. In effect, according to the experiments
of Rentzepis, et al., on Fe(CN)c3- in aqueous solution, the
electron is ejected to distances of several molecular diameters before the solvation shell of the initial ion is adjusted
to the new valency state.
Therefore we are led to conclude that the oxidizing radicals, either the initial ethoxy radical or the a-hydroxyethyl
radical which could be formed by rearrangement, react between themselves or with the solvent giving stable compounds a t rates much higher than that of recombination
CH3CH,0H +ern-
with earn-. At high absorbed doses, when the concentration
of earn- increases markedly, this competition may be reversed and we observed indeed a decrease of @e,m-. The
most probable product of the reaction of the oxidizing radicals would be glycol or its anion.
The quantum yield
= 2 found in presence of hydrogen can be explained by the hypothesis of a quantitative
reaction of the oxidizing radical leading to a second solvated electron. Such reactions are known for OH radicals in
water’ and for NH2 in liquid ammonia.2t8 These mechanisms involve either an intermediate formation of a hydrogen atom giving then a solvated electron in the basic solvent or the reaction of the anionic form of the radical with
the hydrogen. In the case of the ethanolate solutions, the
reaction “f” with the a-hydroxyethyl radical in its neutral
form seems improbable as the reverse reaction “r” is known
in liquid ammonia, i.e.
CH,kHOH
+ H2 e CH3CH20H + H
r
We suggest that the hydrogen molecule rather reacts directly with the anionic form of the a-hydroxy ethyl radical
CH&HO- or with the primary ethoxy radical CH&H20..
We summarize the whole mechanism proposed by
Scheme I.
Thus the solutions of alcoholate ions in liquid ammonia
allow us to study the photodetachment of electrons without
recourse to scavengers which are normally necessary in
media reactive toward the electron. The fact that the quantum yield attains the theoretical values shows that the recombination processes, even in the cage, do not occur.
Acknowledgment. We wish to thank Professor M. Magat
and Dr. J. Sutton for their helpful suggestions and discussions.
References and Notes
(1) M. Ottolenghi and H. Linschitz, Advan. Chem. Ser., No. 56, 149 (1965).
(2) J. Belloni and E. Saito, “Electrons in Fluids,” J. Jortner and N. R. Kestner,
Ed., Springer-Verlag, Berlin, 1973, p 461.
(3) G. Stein, Actions Chim. Biol. Radiat., 13, 119 (1969).
(4) M. J. Blandamer and M. F. Fox, Chem. Rev., 59, 70, (1970).
(5) C. A. Parker and C. G. Hatchard, Proc. Roy. Soc., Ser. A, 235, 518
(1956).
(6) P.M. Rentzepis, Chem. Phys. Lett., 2, 117 (1968).
( 7 ) C. J. Hochanadel, J. Phys. Chem., 56, 587 (1952); J. Jortner and J. Rabani, J. Amer. Chem. Soc., 83, 4368 (1961).
(8) J. Fradin de la Renaudiere and J. Belloni, Int. J. Radiat. Phys. Chem., 5,
31 (1973).
The Journal of Physical Chemistry, Vol. 79, No. 4, 7975