voi”me 39, nun&i
3
-’
kHEMICAL. &YSICS LETTERS
:
-. I‘May, I!376
-.-
. -_
Ei~~ONSQLVATION~ALCQ~QLS J .H . BAXENDALE Chemistry
Department,
AT
77K AFTERPULSE
-.’
tiDiOLYSiS
._
and P.H .G. ShRPE The Univbsi@.
Received4 February 1976
..
Madrester,
Ml3 GPL. UK
_
The presence of benzyl in I-propanol, 2-propanal and ethanol glasses at 77 K markedly affects the‘ changes’ which occur in the absorption spectra of the trapped electrons in the pure alcohols after their production by pulse radiolysis. The well known growth in the visible is considerably reduced while the decay in the far red is unaffected. ft is concluded that the spectral changes arise by “trap-hopping,” of the electrons induced by thermat excitation from ihe initial shallow traps, rather than solvent re-orientation.
1. Introduction Electrons produced in alcohols at low temperatures have initial absorptions with maxima in the near infrared, but these change with time (except at very low temperatures) until stable spectra are produced with maxima from 500-600 nm, similar to the room temperature spectra observed by microsecond pulse radiolysis [1] . The time scale for these changes lengtlrens as the temperature falls and also along the series methanol, ethanol, n-propanol. For I-piopanol the changes were originally [I] followed from 16.5K to 130 K, and recent work [23 using picosecond pulse rzdiolysis has extended the range to room temperature and above. The observations have been inter-. preted tims of solvent re-orientation,around the ele.@& and the time scales and temperature dependence tie in line with those obtained for these processes using dielectrii dispersion measuremen&. Similar changes occur when the alcohols are glasses at 77 K [3-S], but in tbe case of I-propanol, below 115 K, the kinetics of the spectral shifts, originally fust order, alter appreciably and it becomes more convenient to use a logarithmic time scale to present the observations [6]. With ethanol and I-propanol spectral changes are still occurring tiahy secpnds after the electrons have been introduced [5,6] . _Three mechanismshave been suggested to axplain this phenomenon in glasses at 77 5, viz. redistribution of +ctrons into deeper traps following thermal
excitation from the initial shallow ones, electron tunnelling from shallow to deeper traps, and the formation of deeper traps by molecular orientation under the influence of the field of the electron. Following . the work of Kevan [4] it has been widely accepted that even at 77 K the latter is responsible for t&e changes. The most important evidence to support this view is the report [4] that in the presence of solutes which scavenge electrons to give identifiable products, e.g. the benzyl radical from benzyl chloride, the spectral changes still occur and there is no evidence for electron capture which would be shown by an increase in the reaction product. This argues against the electrons beco.ming mobile as required by the trap redistribution mechanism. We have re-examined the effects of benzyl chloride on electrons in I-propanol, 2.propanol and ethanol glasses at 77 K and find that, contrar)i to this report, it has a profound effect on the spectral changes.
2. Experimental The alcohols were purified as described previously [I]. Best reagent grade benzyl chloride was used without further purification. The alcohols or solutions weie contained +I 2 “lollipop” cell with a path, length of ea.‘5 v and were deaerated by pumping and shaking before immersing the cell in liquid ni._.
_ 401
Volumc39,n~mber3’~‘
.-
-_
_:_
‘.
-.
CHEMICALPHYSICSLETTERS
:
:
,,
1 May 19‘76
&o&xi contained In an all silica &war for pulse &~a-
_: 1. d&on_ The p&e radiolysis and optical detection arrange:ment has been described previously [7]. In general
20 ns
electron pulses of 12 MeV electrons w&e-used
and the consequent light absorption changes meas-
ured either with a $hotomultipber, an EG ang G SHS100 silicon photodiode or a Philco-Ford L4521 Ge diode. AlI the detectors have IO-90% response times
of 5 ns or less. -Careful precautions were taken to avoid bleaching by the monitoring light and at the longer observation times this involved the use of narrow band interference and neutral density filters. It was always established that changing the light intensity two fold did not alter the rate of absorption changes.
3. Results and discussion ~ Figs. 1 and 2 show the spectral changes in pure lpiopanol and with 8.7 mM benzyl chloride present. In the pure alcohol the far red absorption has almost
Fig. 2. As fig. 1 with I-propanol containing 8.7 mM benzyl chloride. Insets! (2) CR0 trace at 550 nm, 5 &div, (b) 1050
nm, 1 psldiv, (cl 319 am, I &div. disappeared in 10 s and the simultaneous growth in the visible is about 50% of the original absorption. In the presence of the solute the initial absorption falIs by ca. 20% due to capture of electrons prior to trapping as was found at higher temperatures [l] , the far red decay is on approximately the same time scale, but there is very little growth in the visible. The examination of ethanol and 2-propanol was less extensive but figs. 3 and 4, and 5 and 6 show that here also the solute (17-4 m&l) decrease?%? init@ absorption and reduces the subsequent visible growth
quite markedly. It is clear that in these solutions there is an appreciable decrease in the integrated absorption with time which is more than occurs with the pure alcohols. kssuming no change in oscillator strength, this implies an increased loss of electrons in the presence of solute
Fig; 1. Spectral charges in I-propanol at 77 K followinga 20 ns.p&e of 10 MeV electrons:het: CR0 trace at 550 nm, 5 &ii!. ..
and hence a reaction with the solute. However, as shown by the inset in fig. 2, the initial absorption at 3 19 nm dueto the benzyl radical produced by capture of pretrapped electrons is almost unchanged over the time range covering the visible growth. Apart from capture by solute the only alternative electron
VoI&e
1 May 1976
CHEMICAL PHY$CCS LEERS
39, nuribei 3
.1
.I
-1
--1..
t
. 500
Fig. 3. Spectnl pulse of 10
changes in ethanol
at 77 K following
a 20 ns
MeVelectrons. Inset: CR0 trzce at SO nm, 0.2
&div. removal process is ion recombination.
However,
Klassan et al. [S] report that this does not occur to any appreciable extent in pure ethanol 6t 77 K during the spectral changes and it does not seem reasonable that the solute would increase it. The conclusion must be that solute capture is occurring but it does not lead to the benzyl radical, and if this is indeed the case
.
. 700
900
no0
Fig. 5. Spectral changes in 2-propanol following a 50 ns pulse of 10 MeV electrons. Insets: CR0 traces at (a) 550 nm, 10 Bs/div; (b) 600 nm, 0.15 &div.
then solvent re-orientation cannot be responsible for the spectraI changes. There are other indications in this direction. We have already pointed out in the case of !-propanal [l] that provided the higher temperature results can be extrapolated to 77 K, the relaxation times here would be many orders of magnitude smaller than those observed. Also, reasqnable agreement with the observed relaxation times at the higher temperatures can be
I
3M
Fig. 4. As fg. 3 with ethanol contain& 17.4 mhZbenzyl chloride. Inset: CR0 trace at 550 tirn, 0.2 ps/div.
A1I-. . . am.
ml
-920
I
llw
t3ao
Fig. 6. As fe. S with 17.4 mM Senzyl chloride. Itits:
CR0 traces at (a) 600 nm, 10 .us/div; (b) 600 nm. 0.15 gs/div.
403
Vohlme 39.nunil;er ,-_
3 :
..
:
CJ-&CALPHYSICS
LETTERS
I
..
I.&Say 1976
:_
1qb&ed [S] -,usingthe~simple model of a dipolq mol.&& ro&ing-in a visczou’s’m_e&-rr under the influence-of
the field of the electron.
A controlling
fat-
--tar e-the vkcosity of the-medium and the faster relaxation
df e&an01 arises from a lower viscosity
comrange. How[4] and on
.pared with propanol in this temperature tever, at 77 K the viscosities are reversed
the-same model ethanol should relax more slowly than 1-propanol, which is not the case. Further as mentioned above [6] , there isa distinct difference in the time dependence of the spectral shifts in lpropanol in going from 115 to 77 K which would be -consistent with a change in mechanism_ There is other evidence for acertain amount of mobility of the electrons in these systems. Thus Hase et al, [9]- report-a small loss of electrons as measured by ESR during the time of the spectral shift which occurs on warming irradiated ethanol from 4 K to 77 K. For I-propanol, fig. 1 indicates an appreciable loss in oscillator strength and presumably therefore of electrons during the spectral shift period. In both systems recombination must occur as a result of electron transport. Considering the alternatives to solvent re-orientation at 77 K, although the time dependence-is similar to that required by electron tunnelling [lo] , we find that the time scale for the changes shown for l-propano1 in fig. 1 is temperature dependent [6]. Thus at 93 K the far red absorption persists for only 0.2 s compared with 10 s at 77 K. This is not to be expected on a tunnelling mechanism [ 1 I] . We therefore con-
404
.
elude that the origin of the chan& isthermal tion from shallow i.e. “trap-hopping”.
excitatraps and recaptlrre in deeper ones, The extended time scale would
arise because., as the-average occupied trap depth increases with time,
there will be a decrease.in the extent of thermal excitation and therefore the rate of retrapping will become slower. It would seem mat solvent reorientation predominates at the higher temperatures but that due fo its lower activation energy, thermal excitition, which allows “trap-hopping”, takes over at the Iower temperatures.
References
(11 J.H. Baxendale and P. Wardman, J. Chem. Sac. Faraday Trans. I 69 (1973) 584. PI L. GilJes, J.E. Aldrich and J.W. Hunt, Nature 243 (1973) 70.
[31 J.T. Richards and JX. Thomas, J. Chem. Phys. 53 (1970) 318.
[41 L. J&van, J. Chem. Phys. 56 (1972) 838. [51 W.V. Klasson, H.A. Gillis, G.G. Teather and L. Kevan, J. Chem. Phys. 62 (1975) 2474.
[61 I.R. MiJler. private communication; J.H. Baxendale and P.H.G. Sharpe, to be published.
[71 J.P. Keene, J. Sci. Instr. 41 (1964) 493. VI K. Fueki, D. Fang and L. Kevan, J. Phys. Chem. 78 (1974) 393.
PI H. Hase, T. Warashina, hf. Noda, A. Namiki and T. Higashimura, J. Chem. Phys. 57 (1972) 1039,
[lOI J.R. Miller, J. Phys.Chem. 79 (1975) 1070. iill J.R. Miller, J. Chem. Phys. 56 (1972) 5 173.