..m,m.mm=-- ----=.--.W------

lllullLll=illll

ANL/RPG-94/1 IJ

Argonne National Laboratory9700 South Cass AvenueArgonne, Illinois 60439

RADIATION AND PHOTOCHEMISTRY SECTIONAnnual Report

October 1992 - November 1993Chemistry Division

Argonne National Laboratory

November 1993

RWork performed under the auspices of the Office of Basic Energy Sciences

Division of Chemical Science, U.S. Department of Energy.

UlI_"_BUI"IONOFTHISOOCUMENTIS UNLIMITEI_

RADIATION AND PHOTOCHEMISTRY SECTIONSection Head: Alexander D. Trifunac

Radiation & Photochemistry Group Low-Energy Accelerator Facility(LEAF) Group

Group Leader: Group Leader:

Alexander D. Trifunac, Senior Chemist Charles D. Jonah, Chemist

Staff: Staff:

David M. Bartels, Chemist George L. Cox, Engineering Specialist

Charles D. Jonah, Chemist Donald T. Ficht, Engineering Specialist

Yi Lin, Assistant Chemist Allan C. Youngs, Engineering Assistant

Myran C. Sauer, Jr., Chemist

David W. Werst, Assistant Chemist

Postdoctoral Fellows:

Keith Cromack

Elizabeth Piocos

Ilya Shkrob

Special Term Appointees:Robert H. Lowers

Klaus H. Schmidt

Patricia D. Walsh

Visiting Scientists:Ron Cooper, University of Melbourne, Melbourne, Australia

10/93Ping Han, Biomedical Engineering Institute, Tianjin, China

3/93- PresentVadim V. Krongauz, DSM Desotech, Inc., Fiber Optic Materials, Elgi, IL

3/93An-dong Liu, Radiation Research Centre, Beijing, China

3/90- PresentStephen Mezyk, Atomic Energy of Canada, Pinawa, Manitoba Canada

7/93

iii

TABLE OF CONTENTS

PAGE

REACTIVE INTERMEDIATES IN THE CONDENSED PHASE: RADIATION

CHEMISTRY AND PHOTOCHEMISTRY ................................................................................. 1

A. Chemistry of Ions in the Condensed Phase ........................................................................... 2

1. Zeolite Control of Radical Cation Reactions .................................................................. 2

2. Silica Sol-gel Matrices for Radical Ion Studies ............................................................... 6

3. Ion Recombination in Solids ........................................................................................... 7

4. Multiphoton Processes in Liquids .................................................................................. 9

B. The Role of Solvent in Chemical Reactivity ........................................................................... 13

1. H- and D-Atom Encounters with 02 in Water .............................................................. 13

2. Reaction of H Atoms with I- and Br. .......................................................................... 15

3. Diffusion of H, D, and Muonium Atoms in Ice ........................................................... 17

4. Solvent Relaxation Dynamics ....................................................................................... 19

PUBLICATIONS .......................................................................................................................... 25

SUBMISSIONS ............................................................................................................................ 27

PRESENTATIONS ....................................................................................................................... 29

COLLABORATIONS .................................................................................................................... 33

Research Highlights

October 1992 - November 1993

The following survey of the research activities of the Radiation Chemistry and Photochemistry

Section is representative of highlights of the past year rather than a comprehensive account. For

more detailed accounts of the activities described herein, see the articles listed in Publications and

Submissions at the end of the report, or contact the author(s) indicated.

vii

rll i i qlrll IN I |, i ii ii i i i i t

I. REACTIVE INTERMEDIATES IN THE CONDENSED PHASE:RADIATION CHEMISTRY AND PHOTOCHEMISTRY

A. O. Trifunac, D. M. Barrels, M. V. Barnabas, K. R. Cromack, C. D, Jonah, ¥. Lin, A.-D. Liu, P.Han, M. C. Sauer, Jr., I. A. Shkrob, K. H. Schmidt, O. W. Werst, P. O. Walsh

Outside Collaborators: R. Cooper, S. Mezyk, P. W. Percival, E. Roduner. C Romeroi ilmmi ...................... in I IN IIIII. IIII I I IIIII I III ,Ulllm In

"IT_cresearch described in this survey represents a catic, n processes that are photodriven m other matrices

multifaccted and comprehensive approach to the study occur in zeolites without photochemical excitation.

of transient intermediates and chemistry induced by This survey is presenied in two major parts. Part A

energetic radiation. The goal is to determine the funda- describes several research efforts in which we study

mental chemistry that occurs when ionizing or photo- ions, excited s'tates, and other transients in the con-

ionizing radiation interacts with comtcnsed-phase mat- dem,'ed Fhase. These include studies of radical cations

ter. The various short-lived intermediates, radicals, and ion-molecule reactions by the techniques of mag-

radical ions, electrons, and excited states all play a netic resonance, picosecond emission, and picosecond

role in transforming high energy from photons and absorption, which probethe role of ions in high-energy

particles into different stable chemical products, chemistry, and studies of ions and their reactivity in

The simultaneous goals ()f this research are (1) an specialized rnatrices such as freons and zeolites.

uilderst,mding of the chemical transformations irH,luced Part B outlines the work on the role of solvents

by energetic radiation via the delineation of tl_e overall in chemical reactivity, t lydrogen and deuterium atoms

framework of reactions: (2) the identification and have been used as probes of the short-time events in

description of reactions of transient species; and {3) the radiation chemistry of water. The nature of the

the development of novel tools to allow better pursuit solvation structure around hydrated electrons has been

of .such elusive reaction intermediates. State-of-tllc-ilrl cornpared with thiit of classical monatomic ions by

clectron accelerators, ultrafast laser techniques, and way of thermodynaxnic and transport properties. In a

novel detection means for the study of ultrafast proc study in which ions are produced by both laser excita-

esses and transient species arc c{mtinuously being lion and pulseradiolysis, we examine the dynamics of

developed and improved, solvent reorganization around transient ions.

"l'his work has revealed the uhiqtlitous role of fad- Some of the highlights of the past year are:ical ions iil their earliest observable time frame. Ion- {1) Observations of ion-molecule reactions of

molecule reactions play a very prominent role. The excited aromatic cations via flash photolysis and tran-

eilergy content of the radical cation determines its sient tie conductivity.

reactivity: for example, only electronically excited I2) Study of anion solvation reveals the signifi-

aromatic radical cations undergo ion-molecule cant role of solvent molecular structure,

reaclions. Dit'fusion and molecular transp_rt il'l gen- 13) Zeolite matrices have allowed observation of

ural, a:; well its charge transport, are being examined at a remarkable range of (-'7I-t8 radical cation chemistry,

very low temperatures lot the first time. including the first direct EPR observation of the elu-

The use of novel i11icrorenctor/lllatrix materials sire quadricyclane radical catioli.

such as zeolites allows unprecedented coiltrol of ioil (4) The study of lt-atonl reiictioll with ()2, I- and

chemistry. We call observe aild control multitudes of Cl)lylparJsl)ll of lhe diffusion of !1, l), and illuoniuili in

radical cation lr;.insfornlalions, illaily of wllich were ice, conlinue to expand the use of this runlarkahle

previously observect in the gas phase. Many radical probe,

2

A. Chemistry of Ions in the Condensed PhaseA. D. Trifunac, C. D. Jonah. D. W. Werst, M. C. Sauer, Jr., K. R. Cromack, M. V. Barnabas, Y. Lin,

A. D. Liu, I. A. Shkrob, R. Cooper

1. Zeolite Control of Radical Cation is formed by radiolysis (7 rays or fast electrons) at a

Reactions A. D. Trifunac, D. W. Werst, temperature between 4.2 and 77 K. By carefully con-

M. V. Barnabas, K. R. Cromack trolled annealing of the sample and by observing the

EPR spectrum as a function of temperature, we areRadical cations of simple organic molecules are able to follow the reaction dynamics of the radical cat-

very short-lived in the condensed phase and thus are ion and observe both unimolecular and ion-molecule

studied by fast time-resolved methods or in solid reactions. The radical cations reported here all showmatrices that are capable of providing a large degree of reaction behavior in zeolites different from that in freon

stabilization. The matrix used for radical cation stud- matrices.

ies must possess at a minimum two attributes: it must

trap electrons indefinitely, and it must be relatively a. Elimination Reactions of Branched

chemically inert. Ion-molecule reactions of radical Alkane Radical Cations

cations are ubiquitous in the condensed phase. Thecontrol of ion-molecule reactions requires isolation of The radical cations of methyl-substituted butanes

the radical cations from parent molecules or other have been previously studied in freon matrices, and,

reactive neutrals. Matrix molecules in which the except for the hexarnethylethane radical cation, their

radical cations and unoxidized substrate molecules are thermal and photoinduced reactions have been enumer-

J diluted must be unreactive. Examples of inert matrices ated. In the absence of light, ion-molecule reactions

that prevent rapid ion recombination include solid take place to produce alkyl radicals by deprotonation

inert gases with added electron acceptors, such as of the parent cations. Photoexcitation produces exclu-halocarbons and SF6. sively the olefin radical cation by elimination of CII4.

Zeolites possess several additional attributes con- We studied the reactions of _he radical cations ofducive to the control and study of radical cation reactiv- 2,2,3-trimethylbutane /TMBI and hexamethylethane

ity. Zeolites are crystalline solids and as such do not (HMEI in Na-ZSM-5 zeolite with a SIO2/AI203 ratio

restrict the range of temperature studies as has been of 40.

experienced with other matrices because of low melt- Irradiation of TMB (.5-5 wt.%/ in ZSM-5 at 77 K

ing point. The size-selective ability of zeolites plays a gave rise to the six-line spectrum/a = 31 G). This sig-

central role in isolating radical cations and neutral nal persisted to some degree to at least 170 K. Decay

substrate molecules. Their pore dimensions limit the of the TMB °+ spectrum when the sample was annealed

amplitude of possible molecular motions such as con- to 150-200 K was accompanied by the appearance of a

formational changes and molecular rearrangements, multiplet spectrum with a hyperfine coupling constant

Depending on the size of the substrate molecules, the of 17.2 G. The new signal was assigned to themotional freedom within a zeolite cavity might be tetramethylethylene radical cation. Thus, TMB .+

intermediate between that characteristic of a frozen undergoes thermal elimination of Ct-I4 in the zeolite.

matrix and that of the liquid phase, or it may lie near In contrast, methane elimination from TMB "+ in theeither extreme. The remarkable range of synthetic freon matrix requires photoexcitation.

possibilities with aluminosilicates provides another Coexisting with the TME "+ signal in the annealedvariable, that is, tunability of matrix ion interactions. TMB samples is the EPR spectrum of the neutral

Not only can pore sizes and channel dimensions be radical,/CH313CC'/CH3_2/a = 22.5 G_, that forms as

changed, but, within an isomorphous family of zeo- the product of the ion-molecule reaction between

lites, polarity or acidity can be varied as well. In this TMB "+ and neutral parent molecules. This radical

way, specific electronic interactions between the eventually becomes the dominant species in thematrix and radical cations can be enhanced, spectrum above 200 K. Sctaeme I-I shows the

We have studied several hydrocarbon radical cat- processes that TMB .+ undergoes in ZSM-5.

ions in pentasil-type zeolites by variable-temperature HME "+, which lacks a tertiary C-tl bond to facili-EPR. The radical cation of the adsorbed hydrocarbon tate elimination of CH4, undergoes bond scission at the

3

central C-C bond. This reaction has not been reported parent molecules. On the other hand, when the 1,4-

in freon matrices. The seven-line spectrum (a = 31 G) cyclohexadiene radical cation in CFCi 3 was irradiated

of HME "+ that is observed in the zeolite after irradia- with light (_L> 700 nna) to excite the intramolecular

tion at 77 K is replaced upon annealing, >125 K, by charge-resonance transition, it was shown to undergo

the spectrurn of the 2-methylallyl radical. The 2- isomerization to the 1,3-cyclohexadiene radical cation.

methylallyl radical is the expected deprotonation prod- Figure I-I shows the EPR spectrum observed at

uct of the isobutene radical cation fragment that results 100 K after 77 K irradiation of a 2% sample of 1,4-

from dissociation of HME "+ at the central C-C bond. cyclohexadiene in Na-ZSM-5 (SIO2/AI20 3 = 170).

This reaction sequence is depicted in Scheme I-2. The same spectrum was obtained when the same zeo-

The spin density in methylbutane radical cations lite was loaded with 1,3-cyclohexadiene and then irra-

is highest in the central C-C bond. For HME °+, this diated (Figure 1). The spectra are well simulated with

is also the most labile bond. We have shown that the same hyperfine coupling constants used for the

scission of the central C-C bond does not occur in 1,3-cyclohexadiene radical cation in the freon matrix.

methylbutane radical cations that possess one or more Thus, the isomerization reaction 1,4-cyclohexadiene .+

tertiary C-H bonds. The central C-C bond in HME .+ -- 1,3-cyclohexadiene .+ proceeds without the aid of

is probably weaker than in less substituted butane light in the zeolite.

cations because of greater steric repulsion between When the samples containing the 1,3-cyclohexa-

methyl groups. The dissociation mode may also diene radical cation are annealed above 110 K, an

depend on the relative facility of a concerted elimina- irreversible change occurs to produce the cyclohexa-

tion process, such as methane elimination, dienyl radical. The neutral radical is lhe product of ion-

molecule reactions between the 1,3-cyclohexadiene

b. Isomerization of a Diene Radical Cation radical cation and unoxidized substrate molecules,eq. 1. This reaction proceeds whether the substrate

The radical cation of 1,4-cyclohexadiene is stable molecules are 1,4-cyclohexadiene or 1,3-yclohexadiene.in the freon matrix at 77 K. Shida et al. showed that The product radical is the same in either case.

annealing in CFCI 3 led to the decay of the parent rad-

ical cation and to the formation of the cyclohexadienyl 1'3C6H8°+ + 1,4-C6H8 (or 1,3-C6H 8) - C6H7"radical as a result of ion-molecule reactions between (1)

the 1,4-cyclohexadiene radical cations and neutral

(CH3)3C_CH (CH3)2 Y'7._L_'_'_'_'_'_'__TMI3+ _'CH4 tiP.-(CH3) 2 C °+=C(CH3) 2

SchemeI-1 "H+"_'"I_ (CH3)3 C_C" (6H3)2

Ecy, 77 K "+ -C "+ CI...I2(C1-t3) 3 C_C (C!-13) 3 _p,,,. HUE 4Hlo _ (OH3) 2 =

"H+ 1

Scheme I-2 CH3"-_("CH

2

zeolites is the quadricyclane radical cation. Quadri-

a j d cyclane °+ is a member of the extensively studiedC7H8 °+ family of cations. Investigations of C7H8 °+

species have been carried out in the gas phase and in

the condensed phase and have revealed many excess-

energy and photo-driven processes, including severalisomerization reactions. Under most conditions, oxi-

dation of quadricyclane yields the rearranged radicalcation of norbornadiene, which is estimated to be

more stable by approximately 10-12 kcal mo1-1

Previous efforts to observe quadricyclane °+ in freon

b matrices at 4.2 K were unsuccessful.Utilizing our newly developed experimental cap-

_ ___._ ability for low-temperature in situ irradiation of sam-

ples in the EPR spectrometer, we have investigated

the reactions of the quadricyclane radical cation, 1.+,and the related norbornadiene radical cation, 2°+, from

4.2.K and at higher temperatures. In zeolites, we have

made several observations that are unique when com-

pared to studies of C7H8 radical cations in other

media: (1) A reverse Diels-Alder reaction gave rise to

the cyclopentadiene radical cation in a thermal process,

Figure 1. (a) EPR spectrum observed at 100 K after 77 K contrary to the expectation that this reaction is onlyirradiation of 2%l,4-cyclohe×adiene adsorbed on Na-ZSM-5. (b) EPR spectrum observed at 90 K after 77 K photochemically allowed. (2) The bicyclo[3.2.0]hepta-irradiation of 1.5% 1,3-cyclohexadiene adsorbed on Na- 2,6-diene radical cation is produced in small yieldZSM-5. upon oxidation of quadricyclane or norbornadiene.

(3) Some evidence has been gained for the stabiliza-

tion of the quadricyclane radical cation in zeolites at

When different isomers produce distinguishable very low temperatures.

product radicals (e.g., as in the case of linear alkenes), Over most of the temperature range from 4.2 K to

the radical cation isomerization (thermal or photo- above 200 K, the dominant EPR signal after radioly-

driven) followed by ion-molecule reaction can differ- sis of 1 or 2 in zeolites is that due to the norborna-

entiate between two possible reaction mechanisms" diene radical cation (8.0 G(4H), 3.0 G(2H)). As the

proton transfer from the cation to the neutral and sample temperature is slowly increased, various

hydrogen-atom transfer from the neutral to the cation, reaction products are observed. One thermally acti-

Experiments are usually ambiguous on this point, and vated reaction produces the cyclopentadiene radical cat-

we expect to be able to make use of isomerization ion, 4 "+, (24 G(4H), 14 G(2H)). The cyclopentadiene

reactions to study the outcome of reactions of a given radical cation was observed after 200 K annealing of a

radical cation with different reaction partners. Ion- sample in Na-ZSM-5 with a SiO2/AI20 3 ratio of 170.

molecule reactions with the necessary asymmetry will However, it was not observed in ZSM-5 zeolites of

reveal whether the radical product is derived from the lower polarity (i.e., higher SIO2/A120 3 ratio).

radical cation (proton-transfer mechanism) or the neu- The cyclopentadiene radical cation was observed

tral substrate molecule (hydrogen-atom transfer), from the radiolysis of quadricyclane and norborna-diene. Its formation represents a (3+2) cycloreversion,

e. Transformations of C7H8 °+ Family which, as discussed by Bauld, is a thermally forbid-

den radical cation pericyclic reaction. This reaction has

It is hoped that zeolites with their unique proper- never been observed in the condensed phase before. It

ties will allow observations of many unstable radical clearly proceeds thermally in the zeolite. If the exo-cations that have eluded observation in other matrices, thermicity of the electron transfer from substrateOne such radical cation that we have studied in molecules to the matrix holes activated the reaction,

then the formation of 4"+ would be independent of imately 200 K. We did not observe an excited state

temperature and would be observed immediately fol- species by means of an EPR signal distinct frorn that

lowing irradiation. This was not the case. of 2"+. However, this does not rule out the existence

The bicyclo[3.2.0]hepta-2,6-diene radical cation, of such a species, for the signal may have been hidden3"+, (41.5 G(H), 31.6 G(H), 22.5 G(H), 4.5 G(H)) is by the EPR signal from 2"+.

produced upon irradiation, albeit in small yield Scheme I-3 depicts a possible mechanism involv-

(<5%), of quadricyclane or norbornadiene in polar Na- ing 1"+ and 2'+ and the product radical cations. SomeZSM-5 zeolites, even at 4.2 K. In contrast, the 77 K final remarks can be made regarding preliminary

radiolysis of quadricyclane and norbornadiene in observation of the quadricyclane radical cation in polar

freons gave the 3"+ product only in the case of the zeolites at low temperature. In contrast to EPR spectra

former, especially under conditions of high dilution, obtained at 4.2 K in freon matrices, which are broadProduction of 3"+ from norbornadiene was possible in and unresolved, our EPR spectra obtained after

the freon matrix only with photobleaching, irradiation at 4.2 K of zeolites containing quadri-

These results raise questions about the state of the cyclane or norbornadiene show well-resolved structure,

radical cation precursor of the various radical cation as shown in Figure 2. The spectra sharpen at higher

products that are observed in zeolites and the influence temperatures and can be simulated (using the hyperfineof the zeolite matrix on reactions. The fact that coupling constants for the norbornadiene radical

processes are observed that in other media are photo- cation, 7.98 G(4H), 3.28 G(2H), 0.58 G(2H)) with

induced implies the involvement of excited states. In linewidths of 0.3-0.4 G, compared to 5-6 G at 4.2 K.

the case of the transformation to 3"+, it is possible

thal the exothemlicity of hole transfer from the matrix

to substrate molecules provides the necessary excita- _tion to drive the reaction. However, only a small "1 u_.__a...d3 . Thus, ,4"percentage of parent cations react to give "+

energy relaxation to yield the ground-state parent /_ 7cation must compete favorably with rearrangement. %.

In the freon matrix this competition is very sen- _,sitive to the local environment. The strong concentra- %

ti°ndependence°fthcyield°f3"+infre°nmatrices _ _ _"_...._/was explained by others by assuming that energy dis- 2

sipation was aided by the proximity of other substratemolecules. In the zeolite, there is also the possibility '4

that the rate of energy relaxation is site dependent.

Perhaps 3.+ is produced only from a small percentage Scheme I-3of the adsorption sites in which the excited radical

cation is metastable because of weaker coupling. Theobservation of 3'+ from 2 as well as 1 in the zeolite

implies either that the exothermicity of hole transfer is

greater in zeolites than in freons, which is not likely,

or that certain zeolite adsorption sites provide weaker

coupling for energy transfer to the matrix than in , /! _ -freons,

Additional evidence that the nature of the zeolite

interaction with the substrate radical cation influences

this reaction is the fact that the observation of 3"+ and

4 "+ depends on polarity. The percursor of 4"+ is also

thought to be an excited radical cation because the re- Figure 2. EPR spectrum observed at 4.2 K after irradiationat 4.2 K of norbornadicnc iraZSM-5 zeolite. The brokenverse Diels.-Aldel reaction is thermally forbidden. How- line shows the low-field portion of the differenceever, in this case, it must be assumed that the zeolite spectrum of quadricyclanc and norbornadicne-loadedis capable of stabilizing the excited state up to approx- samples.

Notably, there ,are differences in the EPR spectra of the gel. All steps of gel preparation were carried out

observed at 4.2 K in irradiated quadricyclane and nor- between room temperature and 60 °C.

bornadiene samples that cannot be accounted for by An advantage of this sol-gel method for synthe-

line broadening. In quadricyclane-loaded samples, the sizing inorganic matrices is that the low-temperature

full width of the spectrum is somewhat greater than in route allows organic modification of the material. The

norbomadiene-loaded samples, and features are observ- simplest way to introduce organic components is by

able near the outer wings that are not present in the doping organic molecules into the pores of the gel.

spectra from norbornadiene-loaded samples (Fig. We have synthesized silica sol-gels doped with a vari-

I-2). These differences vanish when the quadricyclane- ety of aromatic compounds (e.g., p-terphenyl, naph-

loaded samples are annealed above approximately 30 K, thalene) by simply adding aromatic solutes to the alco-

even afterrecoolingto4.2 K. The extra features become !1ol solvent used in gel preparation. Absorption and

more prominent with increasing zeolite polarity, fluorescence spectroscopy were used to verify that the

The difference spectrum obtained from the 4.2 K aromatic species were incorporated into the sol-gel.

spectra of the quadricyclane-loaded and norbomadiene- The sol-gel method also allows considerable vari-

loaded samples shows similarities to the spectrum sim- ations to covalently modify the matrix with organic

ulated from theoretical coupling constants for 1°+ functional groups. For example, the use of alkyl- or

This is exciting evidence that we may have made the aryltrialkoxysilane, RSi/OR)3, precursors yields gels

first observation of this very unstable radical cation, with silicon coordinated to three oxygen atoms and

which has eluded observation by so many other one R group. We have synthesized transparent sol-gels

researchers. We are pursuing this issue in still more using CH3Si(OCH3) 3 and C6HsSi(OC2H5)3 alone or

polar ZSM-5 zeolites and hope to further enhance the diluted with the tetraalkoxysilane precursor. In this

presumed I°+ signal, way, many properties of the matrix can be varied, such

as the polarity of the pores. We have also prepared

2. Silica Sol-gel Matrices for Radical Ion transparent sol-gelsfrommixturesoftetraalkoxysilanes

Studies A. D. Trifunac, D. W. Werst and polyethylene glycol (MWav = 200 g). The

addition of the polyethylene glycol results in less

Our EPR studies of radical cations in zeolites shrinking and therefore a gel of lower density. This

show the utility of such inorganic matrices for stabi- dramatic effect seems to indicate that the glycol mole-

lizing and investigating reactive intermediates formed cules participate in condensation reactions and become

by high-energy processes. Ont: shortcoming of zeolites covalently bonded to the network; that is, they are not

is their opacity to light. This disadvantage precludes merely trapped in the silica framework.

their use for most optical methods other than attenu- We have tested the silica sol-gels for resistance to

ated reflectance spectroscopy, which has limited applic- optical damage when irradiated with ionizing radia-

ability to zeolite samples. This year, we initiated the tion. The UV cutoff (-250 nm in the unirradiated ma-

development ofaltemativesilicate matrices with excel- terial) shifted slightly to longer wavelengths after an

lent transparency in the visible and ultraviolet, absorbed dose of several megarad. The transparency of

Porous silica matrices were synthesized from the sol-gels throughout the rest of the spectrum, to

organosilane precursors. For example, hydrolysis of 900 nm, was unaffected. It is important that the matrix

alkoxysilanes leads to condensation reactions and the used to study intermediates generated by ionizing

formation of a polymeric silica network, eqs. 2-3. radiation not give rise to interfering absorbances

induced by the radiation.

(RO)3Si-OR + H20 _ (RO)3Si-OH + ROH (2) Future experiments will investigate the formation

-Si-OH + HO-Si- --,- -Si-O-Si- + H20 (3) of trapped radical ions in sol-gel matrices. Preliminarystudies of transient radioluminescence from pulse-

An alcohol solvent is used because the silane precursor irradiated sol-gels doped with aromatic molecules

is not miscible with water. The product after polymer- indicate that the yield of exited states of the substrateization is a wet gel, containing water and solvent. If a molecules from ion recombination is low in the

monolithic, transparent product is desired, the gel is unmodified SiO2 material. In contrast, sol-gels modi-dried slowly to prevent cracking during solvent evolu- fied with polyethylene glycol give rise to a strongtion, which is accompanied by considerable shrinking delayed fluorescence component that is due to ion

recombination. The glycol-modified sol-gels are The experirnental method used to investigate ion

promising media for carrying out fluorescence-detected recombination in low-temperature solids was time-

magnetic reson,'mce (FDMR) studies of radical cations, resolved FDMR. Fluorescence-detected magnetic reso-

A separate effort was begun to develop sol-gel nance detects electron spin resonance in the geminatematrices for EPR studies of matrix-isolated radical radical ion pairs produced by radiolysis that undergo

ions. The synthetic methods used are similar, except spin-selective electron transfer. Because the yield of

that monolithic, transparent samples are not needed. In radioluminescence upon ion recombination depends on

fact, small particles are more easily introduced into the the multiplicity of the product excited state, and thus

EPR spectrometer and low-temperature cryostat. Pow- on the spin phasing in the pairs, spin transitions in

dered or granulated sol-gels are obtained by mechani- the radical ion pairs are detected as an increase or

cal grinding of large particles or, in some cases, by decrease in the fluorescence intensity. The amplitude

spontaneous cracking of the gel during rapid drying, of the FDMR signal is directly related to the lifetime

Porous silica sol-gels are expected to have some proper- of the pairs (i.e., recombination rate) and rate of spin

ties in cornmon with other inorganic matrices, such as mixing induced by an applied microwave field.

zeolites, used for studying trapped radical ions. The rate of single-step electron tunneling depends

Obviously, the amorphous sol-gel matrix differs from exponentially on distance, K(r) = K0exp(-r/'L). In orderzeolites in that it does not have a well-defined pore to observe an FDMR signal, the ion-pair lifetimes

size. Nevertheless, the pore size and distribution can must be on the timescale of the spin dynamics within

be controlled to a degree by preparation conditions, the pairs. Very fast (<10 ns) recombination allows

This, together with organic modification of the sol- insufficient time to perturb the spin states of the

gels, offers a variety of ways to modify the physical reacting radical ions with the applied microwave field.environment of radical cations and study the influence In the event that recombination is very slow (many

of such environmental changes on their behavior, microseconds), spin relaxation begins to erase all

One encouraging test of the ability of the silica memory of the original spin phasing of the pairs, and

sol-gels to stabilize radical cation species involved no FDMR signal will be observed. In the rigid solid,the substrate molecule, N,N,N',N'-tetramethyl-p- where pair-separation distances become fixed, only a

phenylenediarnine (TMPD). TMPD is very easily fraction of the radical ion pairs, representing a narrow

oxidized, and, in the polar reaction mixture used to slice of the ion-pair-separation distribution, can be

synthesize the TMPD-doped sol-gel, the TMPD radi- FDMR _ctive according to a single-step electron

cal cation formed without irradiation. TMPD "+ tunneling model.

remained stable in the dried sol-gel even at roorn tetrt- One prediction of the single-step tunneling

perature. This was verified by both EPR and optical mechanism is that there should exist an optimal sepa-

spectroscopy, ration distance for FDMR; that is, the FDMR ampli-tude should have a maximum value for some separa-

3. Ion Recombination in Solids tion distance and decrease t'or smaller or larger

A. D. "l'rittm_c, D. W. Wer.st, I. A. Shkrob distances. However, the experimentally obtained d'stri-

bution of ion-pair-separation distances is difficult to

A combination of experiments and theoretical control. The distribution will depend on certain vari-

modeling was carried out to investigate the mecha- ables under experimental control, however, such as

nism of charge recombination in the solid state at low scavenger concentration or the molecular structure of

temperatures. The main question is, in the absence of the solvent. A novel approach used in our work was to

molecular diffusion, itow does rapid radical ion artificially narrow the separation-distarlce distribution

recombination proceed'? The simplest mechanism is by forming bridged radical ion pairs in detergent

back electron transfer via single-step tunneling. Other molecules organized in micelles in aqueous media.mechanisms must invoke some mode of charge migra-

tion without molecular displacement, such as resonant a. FDMR in Frozen Aqueous Micellarcharge transfer or thermal detrapping of electrons. Solutions

Such hypotheses may be applicable in some syslemsin which radical ion recombination occurs on the FDMR experiments were carried out in the tem-

submicrosecond time scale, perature range 30-80 K in aqueous solutions of 5-10

wt.% polyoxyethylene alkylphenols, known as Triton less optimal for FDMR detection. Therefore, bridged

X detergents. The molecules comprising the Triton ion pairs that undergo intramolecular recombination

series have an aromatic head group (1,1,3,3-tetra- account for the vast majority of FDMR active pairs in

methylbutylphenoxy) and a polyoxyethylene tail. The the micellar solutions.

polydisperse detergents used in our study had average Scavenging experiments and tests using model

numbers of oxyethylene units equal to 5 to 40. The compounds all support the above mechanism. In every

proposed mechanism of FDMR in these frozen case, intermolecular electron transfer generated a

micellar solutions is depicted in Figure 3. smaller FDMR signal than electron transfer in bridged

Interaction of fast electrons with the micelles pairs. This was found in homogeneous alcohol solu-

produces positive holes localized somewhere along a tions (frozen) of the detergent molecules as well as the

polyoxyethylene tail and ejected electrons. Some of aqueous micellar solutions. Figure I-4 shows thethe electrons attach to the aromatic head group of the dependence of the FDMR amplitude on the number of

detergent molecule on which its parent hole resides; oxyethylene units. It was maximum for n = 9.5 and

that is, a bridged gemina*e ion pair is produced, decreased for shorter or longer tails. This is the func-

Nonbridged pairs are also produced, but our results tional dependence that is expected if the ion-pair-indicate that the recombination times of these pairs are separation distribution is shifted to larger separation

distances for longer polyoxyethylene tails.e

(1)

°o 5 lo is .0 .5 30 2 ¢0 ,

n

(3) Figure 4. Dependence of the FDMR amplitude in micellar

solutions on the number of oxyethylene units, n, inTriton X detergentmolecules in (i) aqueousand (ii) 2-

_- propanolsolutions(at -10 K).

b. FDMR in Low-Temperature Organic

Solids: Theory and ExperimentThe behavior of FDMR in frozen organic solids ise"

more complicated than in the artificially confined pairs

I (4) generated in micelles. The distribution of separationdistances between geminate pairs of ions is broader,and less is known about these distributions for solids

*(_ than for liquids. If one assumes a correlation betweenelectron thermalization distances in liquids and in

(5) solids, then one can rank different organic compoundsaccording to the relative scale of electron-transfer radiiinvolved in ion recombination. In fact, the FDMR

[ intensities measured in a large number of frozen

I

FLUORESCENCE

Figure 3. Reaction steps that lead to the FDMR signal in alcohols, ethers, olefins, alkanes, etc., correlateaqueous micellar solutions, approximately with known electron thermalization

distances in the respective liquids. This finding sug- eral mechanism proposed for solutions of aromatic

gests that electron tunneling is a primary mode of ion molecules in alkanes (or alcohols), where ArH and RH

recombination in the solid state on the submicro- represent the aromatic molecule and the alkane

second timescale that is probed by FDMR. solvent, respectively, is as follows.

As an aid in gleaning more information from thebehavior (kinetics, microwave power dependence, ArH + hv -+ IArH (4)

concentration dependence, etc.) of the FDMR signal in IArH + hv ---_ ArH .+ + e- (5)solids, a theoretical framework based on a single-step

electron-tunneling mechanism and the Liouville ArH. . +hv _ ArH. +* (6)approach for the description of spin and reaction dynam-

ics in geminate pairs was developed. Although the ArH.+* + RH _ Ar. + RH2+ (7)

theoretical model reproduces many qualitative features

of the experimental data, such as t-I decay kinetics, This mechanism, combined with previous obser-

narrow radial distribution of FDMR active pairs, and va;2ons of riM+ in cyclohexane and trans-decalin, indi-

spin coherence transfer, its quantitative predictions cates that HM+ is the protonated solvent, RH2 +.

seem to underestimate the degree of charge mobility. Observations in radiolysis studies and subpicosecond

For this reason, two modifications to the single-step two-photon ionization studies that alkane radical cat-

electron-tunneling model are being considered. These ions have short lifetimes (due to reactions other than

are (1) resonant charge transfer involving the solvent charge recombination), and the chemical stability of

holes and 12) the existence of electrons in weakly ground state ArH° + with respect to reaction with thebound states. The latter would increase the estimate alkane solvent, are consistent with transfer of charge to

forthe tunneling radii in reactions of negative charges, the solvent via ion-molecule reaction 7.For neat alkanes, the same evidence supports the

4. Multiphoton Processes in Liquids production ofprotonated solvent by reaction 8.

M. C. Sauer, Jr., A. Liu, K. H. Schmidt, RH, + + RH _ R. + RH2 + (8)A. D. Ttqtunac

Using transient photoconductivity measurements,

Alkane liquids and solutions of aromatic corn- we have carried out a detailed reexamination of the

pounds are known to be ionized by intense UV light wavelength and intensity dependence of the yield offrom sources such as excimer lasers, In cyclohexane HM+ in both neat solvents and solutions of aromatic

and trans-decalin, positive ions with anomalously molecules to determine whether such observations are

high mobility (HM+) have been observed by such consistent with this mechanism. The general procedure

photoionization. High-mobility cations have also been was to measure the yields of free electrons and HM+ at

observed in the radiolysis of these and other saturated 248 and 308 nm as a function of laser pulse intensity

hydrocarbons with C 6 rings Icis-decalin, methyl- for neat solvents and for solutions containing aromatic

cyclohexane). Investigations aimed at identifying compounds. In neat cyciohexane and trans-decalin, the

HM+ have been a continuing subject of research in our yields of electron _tnd HM+ have the same intensity

laboratories, and evidence that suggests that HM+ is dependence; for ionization of these two alkanes, two

the protonated solvent molecule has been examined, photons are required at 248 nm, and three photons at

Recent observations via transient optical absorp- 3(18 nm. In solutions containing aromatic solutes,

tion spectrophotometry in our laboratories have shown where the major fraction of the light absorption is by

that the quantum yield of aromatic radical cations the solute, yields of free electrons and HM+ are

created by such laser photoionization decreases with markedly higher. At both 248 and 308 nm, the

increasing light intensity. The conclusion was reached intensity dependences indicate that two photons are

that absorption of a photon by the aromatic radical required for ionization, but that three photons arecation promotes a reaction of the radical cation with required to create HM+. This is consistent with the

the solvent. On the basis of the results of product explanation, based on our product analysis studies,

analysis and effects of intensity on the absorbance that the aromatic excited state reached with one photonobserved due to the aromatic radical cation, the is ionized by a second photon. The radical cation then

reaction was concluded to be proton transfer. The gen- absorbs a third photon and reacts with the s_lvent to

t

10

form HM+. The observed decrease in the quantum 10................................

yield of the anthracene radical cation in 2-propanol ," iCyclohexane248hal

with increasing intensity can be explained similarly. _ _ --_

In the neat solvents, the production of HM+ can be "_ 10o ...... _, _ ......... _ _ _'_' :ascribed to the reaction of the radical cation of the --"

solvent, produced by simultaneous two- or three- E

photonabsorption, withasolventmolecule, whichis _ 10" .. , '_ _.., "consistent with a body of results bearing on the fate of S. _.-. IH_I__ ,j__. A i_.._

Ealkane radical cations in radiolysis. _ 10.2 ........ i .......... Ao - • i

(0

a. Results with Neat Solvents _,

Figures I-5 and I'6 sh°w representative results °n _ 1°3_ _i!i ' _ii;i:"ioi:" '_, ,',_ ..........__,, ,.I;,

neat solvents. Three different conductivity signals cFree ions] io s]

were measured: (1) the signal at 50 ns after the begin- o [F10"4 .... 2 3 4 56789 2 3 4 56789 2 3 4 567

ning of the laser pulse in neat liquids or solutions 10 100 1000

degassed with Ar, primarily due to the electron (e-) IncidentI (mJ/cm2)and the high-mobility positive ion (HM+); (2) the Figure 5. Intensity dependence of ionization in cycle-

signal at 50 ns after the beginning of the laser pulse in hexane at 248 nm. The upper set of points is for a

neat liquids or solutions degassed with SF6, primarily solution degassed by bubbling with argon, and the mid-dle and lower sets with SF6. The various symbols rep-due to HM+, but contributed to by normal-mobility resent different conditions of laser operation or treatmentions; and (3) the signal due to scavenged free ions. of the data. The curves are calculated as described in the

The conductivity signal (3) is due to the ions that text, for a two-photon ionization process.

have escaped geminate recombination and is measured

after scavenging of electrons by SF6 and conversion of 10°H M+ to a normal mobility cation, but before -""

Ehomogeneous recombination begins.. These signalswere determined from measurements of conductivity -_vs. time. __ 10' -

In Figs. I-5 and I-6, the experimental values of sig- E0_"1:3

nals (1) and (3) are plotted without alteration. To obtain 5

HM+ vs. light intensity, signal (3) was subtracted -'= '._,_'-.. 10-_

from signal (2). This is an approximation because part _"

of the signal at 50 ns is due to geminate ions other

than HM+, and the free ion.signal due to ions other "E'= __ [_ffslei _'[Fr

than HM+ is smaller at 50 ns than signal (3). The _ 10_ _'t:

._ ionsl .._ t_ ' ..;resulting error is negligible for the neat solvents, but

"ois expected to cause the signal labeled HM+ to be too =o

large for solutions containing aromatic solutes, oThe conductivity signals were divided by the inci- 10" " ; ; _1oo

dent light intensity and plotted vs. incident light inten- Incident I (mJ/cm 2)

sity on a log-log scale. For a two-photon process, a Figure 6. Intensity dependence of ion_ization in cyclo-line ofslopeunity is expected (as is observed)for such hexane at 308 nm. The upper set of points is for aa plot if the fraction of the light absorbed is negli- solution degassed by bubbling with argon, and thegible. The experimental results are compared with a middle and lower sets with SF 6. For the upper and middle

curve calculated by the finite element analysis method, sets, the signals are at 50 ns after the beginning of the

which is not described in detail here. Briefly, in this pulse; decay due to recombination is not appreciablebecause of the low concentration of ions. The free ionscalculation, parameters for two-or three-photon ioni- (lower set) were measured at 2 Its. The curves arezation (a2 and a3) are chosen by trial and error to calculated as described in the text, for a three-photonmatch the electron conductivities observed at low ionization process.

11

intensity, and the fraction of escaped ions (ffi) is The ratio of HM+ signal to electron signal is of

adjusted until the measured power absorption is interest because, if we accept the literature values of

matched. The parameters derived in this manner for the mobilities, it is a measure of the fraction, filM+,ionization of the neat solvents are given in Table 1. of the ionization events that produce HM+. Values of

The same calculated curve, normalized to account for filM+ of 0.35 for neat cyclohexane and 0.6 for neat

the different mobilities, is reasonably consistent with trans-decalin (for both 248 and 308 nm) were obtained

all three conductivity signals in Figure 5, which by analysis of the results.indicates that every ionization results in a constant The value of fiiM+ is less than unity, suggesting

ratio of HM+ to electron, independent of intensity, that not every escaped positive ion is HM+. Thisconclusion depends on the accuracy of the mobility

values. The radiolysis/microwave-conductivity methodTable 1. Coefficients for production of free ions and

the probabilities, ffi, for separation into free ions in by which the mobilities of HM+ in cyclohexane andtrans-decalin were determined would result in a low

neat liquids, a value of the mobility if not all of the positive ionswave-

length produced by radiolysis were HM+. If the true values ofliquid (.nm) a2 a3 ffi b griM+ are higher than reported, the values we would

cyclohexane 248 1.9x10 -20 0 0.0060 obtain for filM+would be even smaller. Therefore, we" 308 0 1.3x10 -31 0.0030 must conclude that at least 65% and 40% of the

trans- 248 1.4x10 -20 0 0.010 escaped positive ions are not HM+ for cyclohexane and

decalin trans-decalin, respectively." 308 0 1.3x10 "30 0.020 A possible reason for [HM+]/[e] values less than

aTo obtain the rate of production of free ions during a unity is that HM+ is formed by proton transfer orlaser pulse in moles literl ,_ec-I a2 is multipl;ed by the charge transfer from an excited cyclohexane radical cat-square, and a3 by the cube, of the intensity in mJ cm-2 sl. ion that also decays by a competing process, for exam-

The total rate of ionization is greater than this by I/ffi. pie, reaction 11. Evidence of such reactions has been

t_The values of ffi given in this table are for a field obtained for alkane liquids and solutions even understrength of 7.1 kV cmj. The values of ffi are therefore conditions where RH .+ contains little, if any, excita-higher than the separation probabilities at zero fieldstrength by factors of 1.41 for cyclohexane and 1.36 for tion,trans-decalin.

b. Results with Aromatic Solutes

For cyclohexane at 308 nm (Figure 6), the slope The addition to cyclohexane of a small concentra-of the plot is nearly 2 and therefore consistent with a tion of a solute such as anthracene or pyrene, which

three-photon ionization. This is expected, because two absorbs light significantly at the laser wavelength,

photons have insufficient energy to ionize cyclohexane, causes a marked increase in the yield of ionization.

The results on neat solvents can be described by The yield of HM+ is likewise enhanced.

reactions 9-11. Figure I-7 shows the intensity dependence at

248 nm of the electron yield (Ar-saturated solutions)RH+2hvor3hv _ RH '+*+e" (9)and the HM+ yield (SF6-saturated solutions) obtained

RH .+* + RH ---* R. + RH2 + (10) for 2 x 10 .6 M anthracene in cyclohexane. These

results strongly support the three-photon nature of the

RH "+* _ H2 + olefin "+ (11) production of HM+ in this system. This can be seen

RH2 + is the high-mobility cation, HM+. The qualitatively by noting that at low intensities theresults of other studies establish that reaction 10 slope of the curve for the electron or the free ions is

close to the value of unity expected for a two-photonoccurs, although excitation may not be required, inwhich case reaction 10 becomes identical with reaction process, whereas the slope for HM+ is near to two, the

value expected for a three-photon process. The curves8. The results of the present study are completely

consistent with reaction 10, and this is supported by through the experimental data in Figure 7 werecalculated by using the method described briefly in the

the preponderance of relevant radiolysis studies.following.

12

100-_............... ............... :....................................... excited states of anthracene and the radical cation of

12xl 0B M Anthracene in Cyclohexano 248 nmI anthracene was calculated using the absorption coeffi-E-_ . , i cients given in Table 2. The yield of free ions is

_ 10 . . • _ determined by the values chosen for these absorption

E _] coefficients and the quantum yield for separation into_ 1 free ions, _fi. The yield of HM+ depends on _fi and_= the absorption coefficient chosen for the aromatic rad-

ical cation. Therefore, the proposed mechanism, whereE

0.1 absorption of a 248 nm photon by the anthracenecoradical cation leads to formation of HM+ in cycle-

.-_>" hexane, is well substantiated by the agreement of the

•__ 0.01 calculated and experimental results for the concentra-

-o tions of ions and the calculated and experimentalO

o power absorptions.o.001 ........ _ ........ b ................ I Similar results (not shown) were obtained foro.1 1 lo loo looo

Incident I (mJ/cm 2) 5 x 10-6 M pyrene in cyclohexane for 308 nm excita-Figure 7. Intensity dependenceof ionization in 2x10-6 M tion. The results show that the yield of HM+ increasesanthracenein cyc]ohexaneat 248 nm. The upper set of more rapidly than the yield of ionization, just as waspoints is for a solution degassed by bubbling with argon, observed for 248 nm excitation. Again, a three-photon

and the middle and lower set with SF6 (triangles) or CO2 mechanism for HM+ formation is indicated.(diamonds or squares). In the upper set, the circles are

The ratios of HM+ conductivity to electronobtained by correcting the pluses for ion recombination.The curves are calculated as indicated in the text; conductivity at 248 nm obtained from Figure 7 andmobility values from the literature were used except that the analogous results at 308 nm show that the relative

the mobility of HM. had to be increased by 20% to HM+ yield is much smaller for the 308 nm photo-

obtain a match with the experimental points, ionization. An analysis of the conductivity ratios at

100 mJ cm -2 shows that values of filM+ of 0.9 and

Two-photon ionization ofthe solvent was included 0.12 can explain the 248 and 308 nm results,

in the calculations by using the parameters derived respectively. The low value at 308 nm appears to befrom the results on neat cyclohexane and given in primarily due to incomplete conversion of Py'+ (the

Table 1. Absorption of light by the ground and pyrene radical cation) to HM+, because of the low

Table 2. Parameters in experiments and calculations on cyclohexane and 2-propanol solutions, a

parameter b 2 _tM A/chex 5 pM Py/chex 2_/.tM A/2-prop

known quantitiesC:

laser wavelength (nm) 248 308 248

Eground state (M'! cml) 1.2 x 105 1.14 x 104 1.2 x 105

ksinglet --- ground state (s- 1) 5.0 x 107 1.44 x 106 5.0 x 107

ksinglet- triplet state (s-I ) 9.2 x 107 6.5 x 105 9.2 x 10.7derived quantities:

Esinglet state (M-! cml) d 3.0 x 104 4.6 x 103 3.6 x 104

£triplet state (M-I cm-l) d 3.0 x 104 4.6 x 103 3.6 x 104£radical cation (M-I cm-l) d 2.0 x 104 1.26 x 103 1.3 × 104

_fi 2.2 x 10-3 9.0 x 10-4 3.2 x 10-2

Oradical cation .... HM+ 1.0 1.0 1.0

aA, Py, chex, and 2-prop represent anthracene, pyrene, cyclohexane, and 2-propanol, respectively.t'The designations for e's, @'s, and k's refer to the aromatic solute, except for tl)fi. The e's are at the laser wavelength.CThevalues of e and k given are from the literature. Excimer formation in pyrene solutions is known to occur, but would notbe significant for the concentrations used here on the time scale of interest.dAt the laser wavelength.

13

absorptivity of Py.+ at 308 nm. At 248 nm, a much 4-

riM]larger fraction of An"+ (the anthracene radical cation)is 3- M Anthracene in 2-Propanol248

converted to HM+; this is evident from inspection of _[:: _. E_29_9

the results for HM+ in Figure 7, where leveling off of

the HM+ signal at higher intensities is seen. The large _ =, I I _ .....| U s Iidifference in the 248 nm values of filM+ for neat - 10"

cyclohexane and anthracene in cyclohexane shows that Eo ,-B- o o6- _ _,.n_:!....................the creation of HM+ when An .+ absorbs a photon is 5 __ ...... _z .......,....jr'- II .........

considerably more efficient than formation of HM+ in ---_,..

the simultaneous two-photon ionization of _ 3- .::" _...........: ::":" ('._ (.3 _",

cyclohexane. In other words, in the latter process, the '- c.,......

formation of some other positive iota must account for _ corea large fraction of the ionizations. Ea_ 10 B- \,

Figure I-8 uses results frorn an earlier study of '- ,0 't - ":;.

248 nm laser ionization of anthracene in 2-propanol, o _- ;,5"

and shows that those results can also be explained by , _ _ _,; _,;_ _ ,_ _ _ _ _ _t ' _ ......._the same model as used for the alkane solutions. The lo 10o

solid curves shown in Figure 8 vvere obtained using Incident I (mO/cm 2)

the parameters given in Table 2. Figure 8. Intensity dependence of ionization in 5×10 .5 Manthracene in 2-propanol at 248 nm. The upper set of

The intensity effects (which indicate the "bleach- points is for a solution degassed by bubbling with argon,ing" of the ground state aromatic radical cation) show and the lower set with SF 6. The concentrations werethat HM+ is formed via an excited state of the are- obtained from the e,_ol absorption at 84(1 nm and the

matic radical cation, and are completely consistent anthracene radical cation absorption at 720 nm for the

with the formation of the high-mobility species by upper and lower sets, respectively. The curves arereactions 4-7. Reaction 7 may involve charge transfer calculated as described in the text.

or proton transfer on the basis of the observed inten-

sity effects. Evidence that IJM+ is protonated cycle- was used {the aryl radical reacted mainly by abstrac-

hexanc comes from product analysis studies, where the tion of H from cyclohexane). This assignment is

conclusion was reached that the aryl radical was consistent with radiolysis results bearing on proton

formed, based on the increase in products resulting transfer refened to earlier, as well as with a number of

from cyciohexyl radicals when anthracene was present magnetic resonance studies that indicate that radical

as well as from the observation of corresponding cations of alkanes undergo reaction 4, and do so by

amounts of C14D9H when perdeuterated anthracene proton transt'er rather than hydrogen atom abstraction.

B. The Role of Solvent in Chemical ReactivityA. D. "Fritbmtc, D. M. Bartels, C D. Jonah, Y. Lin, K. fL Schmidt, S. Mezyk, E. Roduner, P. W. Percival

1. H- and D-Atom Encounters with 02 in to previous EPR methods in that the destruction of

Water D. M. Barrels spin-phase coherence by H-atom scavenger follows

simple pseudo-first-order kinetics. This bypasses the

Several recent studies trom this laboratory have problem of very complicated kinetics caused by the

taken advantage of electron pulse radiolysis combined chemically induced dynamic electron polarization

with pulsed, time-domain EPR detection to measure (CIDEP) phenomenon in the longitudinal magnetiza-

reaction rates of tt and D atorns in water. Comparison tion. EPR detection of H atoms is advantageous in

l of 1{ reaction rates in water with the gas-phase values terms of its great specificity; the doublet spectrurn ofis a prornising venue in the fundamental study of sol- ca. 510 gauss splitting is impossible to mistake for

vent effects in chemical reactions. The technique, any other radical. Detection of the D-atom triplet isbased on the attenuation of free induction decays .just as specific, and in a recent publication we

(FIDs) in the presence of radical scavenger, is superior demonstrated that H/D isotope effects can be

14

detennined by sitnultaneous measurement of theH and time scale of the FID. Rate constants are easily

D reaction rates in a 90% D20/10% 1-120mixture, determined by incremental addition of scavenger to

It has been argued, based on the similarity of produce a linear plot of damping rate vs. scavenger

muonium and hydrogen atom rate constants in the concentration.

diffusion-limited regime, that there is essentially no Results for all experiments in H20 are plotted

isotope effect in the diffusion of H or Mu (or D) in Arrhenius feral in Figure 10. The rate constants are

atoms in light water. Measurement of absolute H and similar to but slightly larger than reaction rates

D diffusion coefficients in H20 and D20 separately reported by other workers, as to be expected for a spin

has found roughly a square-root of mass difference, exchange process. Rzsults for both H and D encoun-

but this experiment cannot differentiate atomic mass ters with O 2 in 90% D20/10% H20 are presented in

and solvent isotope effects. The present study was Figure 11. The rate constant for D is consistently 5-

inspired as an attempt to clarify this issue by simul- 10% lower than the rate for H. The rate constant fortaneous measurement of diffusion-limited H and D the H + O2 dephasing in 90% D20 is consistently

reaction rates in the same H20/D20 mixture, to avoid less than the rate in pure 1:-I20,and seems to have a

any possible effects of viscosity or density differences lower activation energy. However, care should be

between light and heavy water. Molecular oxygen was taken not to overinterpret this comparison given thechosen as a reaction partner because it is easy to available S/N ratio.

replenish its concentration in a recirculating flowsystem by continuous bubbling. Previous reports of

the H + 02 reaction rate suggested that the reaction to ,.0 - [HatomFIDinwater,pH=2 I

form He2 is not actually diffusion limited, but nearly il i .- J''-Nitrogen saturation I" _ ] 39% Air saturation

so. It was not anticipated that this would matter in the 0.s-

spinPresentexchangeC°nteXtbetweenbeCausetheeitherparamagneticChemicalreactionHand 02or o.0...............:!!:_:__'i%_?i-.:i._i' _..,_:,._,::::;;:,..i_:::;!:__.:i!,_)._!!_,.i_:_).i_i-if:, '_

altenualionSpecieswillexperiment.Causeloss of phase coherence in the FID -o.s - +i...::".:" : :_i +: _:_:!:..i:i_:)_: '

H and D atoms were generated within an EPR :

cavity by pulse radiolysis of aqueous solutions with ..0 , , , ,

5-25 ns pulses of 3 MeV electrons from a Van de 0 _ r,,ea(seex_0a l 4 SGraaff accelerator, hnmedialely following creation of

the radicals, a 30 ns r_/2 microwave pulse is applied to Figure 9. Typical Fll)s of the H-atom low-field Inq = +1/2)the EPR cavity. After a ca. 130 ns receiver "dead line immediately following pulse radiolysis of pH 2time", FIDs of the It atoms can be observed, as shown aqueous solution. The FID is attenuated in the partiallyin Figure 9. One thousand shots were averaged to air-saturated solution by the prEsEnccof 02.

produce the signal-to-noise ratio illustrated. The

signals are well described as a simple damped cosine,and tim information of interest is contained in the

reduction of damping time constant with scavenger ,. mconcentration (Figure 9). The effective T2 will be ..mtr lea= (14.08 :t: 0.59) kJ/mole I" .- m [A = (6.09 + 1.29) x 1012 [

given by ""_, ,. " ram.. /Jo i '_ "l/T2Ieff) = i[I"2 + ks[S] + _ kex[Ri] (12) .-- "-.

i 6' 2xlOt°'_ ill+ "l'_i ,,." "..o =1

where 1/'1"2 is the (small) natural dephasing rate, a =at.pks[Sl is the scavenging rate of interest, and

E,klexlRi] represents the contribution of all spin io'0- ,,I l l l l l

exchange and reaction due to second-order radical- 2.9 3.0 3.1 3.2 3._3 3.4 _.5

radical encounters. In general, the free radical 1000/Temp (K1)

concentrations are sufficiently low that the latter term Figure 10. Rate constants for bimolecular dephasing of His virtually constant, if not negligible, on the 5 Its atoms by Oe in H-,O.

15

,. 2. Reaction of H atoms with I" and Br"

.-. [forH + 02 D. M. Barrels, S. Mezyk

'_ 3. " "- .. - lea = (I 0.5 :i: 1.2) Id/mole

_'_" -r [A = (1.3 + 0.fi) x 10 t2 The study of the radiation-induced chemistry of

"_ "" , iodine-containing species in the aqueous phase has6+'2xlOtO " .. -. been stimulated by the recognized safety hazard of

"_, _..__ radioactive iodine release following a nuclear reactor

[for D + O_ -I "''"_ accident. There now exists a significant body of

lea = (11.0 + 1.2) kJ/molel - "__ information on the reactions of iodo-substituted spe-

10 ,°- ]A = (1.4+0.7)x 1012 ] _I....... cies with a variety of free radicals in aqueous solution,I I I I I I

2.9 3.0 3.1 3.2 3.3 3.4 3.5 However, the paucity of data for H-atom reactions

1000/Temp (K t) with even the simplest of these species has meant that

models for predicting the volatility of radioactiveFigure 11. Dephasing rate constants for both H and Datoms in 90% D20/10% H20. iodine under accident conditions are incomplete.

Furthermore, even when such data are available, as for

the H-atom reaction with I-, the large scatter in the

reported rate constants has made their inclusion inthese modeling schemes essentially futile. To begin to

It has been asserted, based on the similarity of rectify this situation, we have undertaken measurement

many H atom and Mu reaction rates in the diffusion of rate constants for the reaction of H atoms with Ilimit, that there is virtually no difference in their dif- and Br-. Our FID attenuation method is particularlyfusion rates. The factor of nine difference in mass advantageous for this purpose because of the simple

should appear as a three times larger reaction rate for pseudo-first-order scavenging kinetics generally

Mu if the system behaved like a dilute gas. A roughly obtained. Important insight into the nature of the_/2 difference between H diffusion in H20 and D dif- products is also gained from H-atom spin exchange

fusion in D20 was also reported, but it was suggested and CIDEP in the presence of the halide ions.

that the primary mass effect might be the librational The FID attenuation experiment was carried out as

frequency of the water molecules, rather than the mass described above. Hydrogen-atom longitudinal mag-

of the diffusing atoms. We are now in a position to netization was also recorded as a function of time afterclarify this issue. First, given the comparison of H the radiolysis pulse for several different scavenger con-

and D spin dephasing in Figure l l, there definitely is centrations, by scanning the microwave pulse delay

an atomic mass effect on the diffusion rates in water, relative to the radiolysis pulse, and integrating the

In water of a given isotopic composition, the lighter first 100 ns of the FID signal in a gated integrator.

atom moves faster. However, the effect is signifi- Rate constants for the reaction of H + I- were

cantly less than a classical square-root-of-mass depen- measured by FID attenuation at several temperatures in

dence. Secondly, there is also a difference between the the range 10-57 °C, and results are plotted in

dephasing rates for H atoms in H20 and in 90% D20 Arrhenius form in Figure 12. At 25 °C, the rate con-

water, and an apparent difference in the activation stant (extrapolated to zero dose) is 2.8 .+_0.4 × l08 M

energies that does not parallel T/q. The great impor- I sec-l. The reaction rate is found to be remarkablytance of water librational motions in controlling the insensitive to the temperature, with fitted activation

diffusion should not be discounted, but the diffusion energy of only 1.8 + 4.6 kJ/mole. Rate constants for

coefficient is complicated by several simultaneous the reaction of H + Br were also measured over the

effects. There is clearly room for better theoretical 10-58 °C temperature range, and the results are also

understanding of these light-molecule diffusion included in Figure 12. At 25 °C the rate constant is

processes in water, and more high-precision mea- 1.7 + 0.3 × 106 M-I sec -1 As in the iodide case, thesurements of diffusion rates over a wide temperature reaction of I-! with Br- seems to have only a small

range. The issue is far from esoteric, as many iso- activation energy of 6.3 + 6.2 kJ/mole.

topic tracer techniques used in geology and oceanog- The CIDEP of the H atom in the presence of theraphy depend critically on gas-transport properties in halide ions was also recorded as a function of time,water, and the results were quite curious in light of the rate

16

• I-.. I _ --""19- Z_ • ,r -_._'ca . _" 2- I H high field• H + I rate .._,,.,-.

1 8- Ea = 1.8 + 4.6 kJ/mole ._.

:_17- _ 0-

._1 6- H + Br" rate _" 3E-2 =

• . ._w,""15 <1:•-J - T_ "I" H ow field line

14 I -- ]; -I 121- 4"i i , m........,'" ,T [I]- 03.1 3.2 3.3 3.4 3.5 I I I I I '"i

-1 s 0 2 4 6 8 10

lIT (K xl0 ) Time (lasec)

Figure 12. Arrhenius plots of Lnl krxn j vs. I/T for the Figure 13. Effect ot i on lt-alom C'II)EP in pH 3 per-reaction of H atoms with !- and Br in aqueous solution, chloric acid soluti(m at 25 °C following a 25 ns radioly-

sis pulse. In the presence of 5.45 x 1()'_ M iodide, thechemical lifetime of the H atoms should be only 680 ns,but magnetization persists beyond I() tas.

constants determined by FID attenuation. In Figure 13

we show the H-atom magnetization in the presence of

5.45 x 10-3 M iodide. The decay of H magnetization times observed in the halide solutions differ dramati-

is anomalously slow, given theexpected660 ns chemi- cally from the rate constants deduced in the FIDcal lifetime, and the measured FID decay time of attenuation experiments, i

480 ns. Observable H-atom magnetization persists We suggest that this anomalous CIDEP result in

beyond 10 Its. Similar behavior is found in the Br- the halide scavenging systems is consistent with an

scavenging systems. Qualitatively similar results, equilibrium H + X- --_ HX-, which is established on a

where the rnagnetization decay time lagged the calcu- microsecond time scale. In this case, the FID decaytime will be controlled primarily by loss of spin

lated chemical decay by factors of two to four, were

obtained in the presence of both I- and Br- in solu- coherence in the forward reaction, but additional CIDEP

lions of pH 1, 2, and 3. The same phenomenon was can be generated by self-recombination of the residual

also found in iodide-containing solutions al neutral H atoms. Back reaction rates on the order of 106 sec -1

pH. No indication was found that the F1D attenuation would be required to maintain a sufficient l t-atom con-

rate might be a function of pit. centration to qualitatively explain the CIDEP results.In theirpulse radiolysis/lransient absorption studyThe H-atom magnetization decay may slightly lag

the chemical decay if polarization is generated in of the production of I2- in iodide-containing acid solu-

subsequent H + product free radical cross recombine- lions, Elliot el el. demonstrated thal the observed for-

tions. However, it is unreasonable to expect this marion kinetics were consistent with the following

behavior when the product is HBr- or HI-. Small lin- mechanism:ear free radicals lend to have very fast spin relaxation H + I- ;:---!II (13)

limes due to the spin-rolation coupling mechanism, tti-+I- ::: t112= (14)The presence of the heavy atom in the I IX- molecules

Iti2 =+(tt+)aq -- ti2+12- (15)should enhance the spin relaxation even further, to the

point that spin coherence does not persist between re- Under the conditions of their experiments (large I,

encounters of ttX and H radicals. In this limit, ([]+)aq concentrations), reaction 15 was shown to beCIDEP will not be generated, Even if H-atom CIDEP rate-limiting (kH = 2.3 + 0.4 × 107 M-I sec-I), with

could be efficiently generated in such encounters, the the It atoms, tti-, and ti12= reaching a preequilibriumchemical conversion of It to the product radical will on a submicrosecond time scale (K7 = 0.82 + 0.13

ensure that the It magnetization decay only slightly MI). They considered only the forward reaction of

lags (10-20%)the chemical decay. The CIDEP decay 13, whose rate they deduced must be greater than or

17

equal to about 4 × 108 M-1 sec-I, in good agreement ice by electron pulse radiolysis over the temperature

with our FID attenuation measurement (2.7 x 108 range 150-260 K. The dependence of this transverse

M-I sec-l). It appears that the data collected by Elliot spin relaxation on the ice H/D composition clearly

et ai. demands K19 > 25 M-I . Values of k. 19 on the demonstrated that the mechanism is purely electron-

order of several times 106 sec -I, as required to explain nuclear dipole relaxation, modulated by diffusion of

our CIDEP results, are perfectly compatible with the the atoms through the lattice. An unfulfilled goal of

mechanism shown above, the original study was the experimental measurement

The discovery of reaction -13 can resolve the huge of longitudinal relaxation times, TI, which would

discrepancy between the kinetic study of Elliot et al. allow unequivocal determination of the motional cor-

and the steady-state radiolysis/product analysis study relation time, x , from solution of the simultaneous

of Deeble et al., which reported only very slow equations:reaction of H with I-. In the reaction scheme assumed

by Deeble et al., the H atoms may react irreversibly 1 2 ( 2x "_

with either the halide ion or with 2-propanol-d8 to TI -_)L _ _ (16)1+ o_2_'c2give lid product. This competition should, in princi-

ple, reduce the amount of HD product. However, rate

constants sonic two orders of magnitude smaller than I 2/" 1: '_

those detemlined in our study were deduced for both ---=T2O_L_x+ -2 2) (17)1+ 0_¢_ions. Clearly, if the HX- product dissociates again ona short time scale, the liberated H atoms will all even-

tually react with 2-propanol-ds, and the added halide where o_o is the electron larmor frequency (2rt x 9.27

ion will produce essentially no change in the HD GHz in our experimentl, and o)2 is the second moment

product yield. The effective rate for k19 found in the of the random magnetic field created by the lattice

competition experiment is naturally much smaller than nuclei. The measurements of T2 were already madethe correct value, difficult by the interference of Heisenberg spin

In conclusion, our measurements have succeeded exchange, a second-order spin relaxation mechanism

in establishing the H atom + halide reaction rates, and resulting from diffusive encounters of the H and D

reconciling previous literature results in temas of the 1t atoms. Initial attempts to measure the much longer '1"1+ X =ftX- equilibrium. The measured activation relaxation time proved futile because signal-to-noise

energies for the forward reaction are found to be far was insufficient at the very low H-atom densities

smaller than typically determined for reactions of required to avoid the second-order spin relaxation

small free radicals in aqueous solution, even though Pr2ocess. Consequently, the T2 data was modeled withthe rates are orders of magnitude smaller than the dif- Ct_Las an additional fitting parameter.

fusion limit. It appears that the only reasonable expla- Analysis of the 'I'2 results suggested that it might

nation for this situation is in terms of a quantum- be possible to measure the TI for H atoms under the

mechanical curve-crossing mechanism, involving most favorable circumstances, in H20 ice near the

orthogonal electronic states. (If the mixing of the melting point. We report here our measurement of both

states involved spin-orbit coupling, the heavier I T I andT2 at-6 °C, from which we deduce diffusional

atom would naturally react much faster than Br, as correlation times for H and D atoms in hexagonal ice.

observed./ However, it is entirely unclear which elec- tt atoms were generated in a random polycrys-

tronic states might cross, given the gas-phase energet- talline ice sample in an EPR cavity by pulse radioly-

ics and correlation diagrams of HX molecules, sis with 3 MeV electrons from a Van de Graaff accei-

Apparently, the water solvent plays an intimate role in erator. T2 measurements were made by the two-pulse

the process. This problem awaits further study, spin-echo method. Ti measurements were made by adynamic inversion recovery method. A 180" inversion

3. Diffusion of H, D, and Muonium pulse was applied at a time to after the radiolysis

Atoms in Ice D. M. B_lrtels, P. W. Perciv_d pulse. The longitudinal magnetization was then mea-sured at a later time to+tree by applying a 90 ° probe

In a previous year, we reported EPR spin-spin pulse, and integrating the first microsecond of the

relaxation times, T 2, for H and D atoms generated in resulting FID.

18

In Figure 14, we illustrate the magnetization of H amplitude of the spin echo. A weighted average of this

atoms in ice by plotting the initial FID amplitude as a and eight other measurements gives 6.9 + 1.0 ItS asfunction of time following pulse radiolysis at -6 °C. the I' 2 value for Hatorns in ice at -6 °C

As can be seen, the H-atom EPR spectrum is imme- By substituting the relaxation times determined at

diately observed in a spin-polarized state, with low- -6 °C, T I = 23 + 3 Its and T 2 = 6.9 + 1.0 Its int_ eqs.field emission and high-field absorption (an E/A pat- 16 and 17, it is possible to derive values for WEand

tern) characteristic of CIDEP from random (F-pair) the correlation time at that temperature, namely

recombination reactions. The prompt appearance of

this polarization implies significant cross recombina-

tion in radiolysis "spurs". Under the relatively low- 1.0- _.,,,,,.

dose conditions of this experiment, subsequent CIDEP "_"" 0.5- _ atomhighfield line]generation from second-order recombination is slow,

and the polarization decays slowly rather than growing "o=quickly as demonstrated in our previous work. ._ 0.0__I1_

Unexpectedly, a small signal was still present imme- 1:::diately before each radiolysis pulse, left over from the ,a:. 0.5 -

pulse applied 2.8 ms previously. ClI.i.

Because the absolute signal amplitude persists .1.0- / I I /much longer, inversion recovery and spin-echo exper- 0 i 0 0 2 0 0 3 00 4 00

inaents were performed on the high-field line. Time (l.tsec)Inversion pulses were applied at to = 50, 250, and 450

Its following radiolysis, and also at 200 Its before eachFigure 14. tt-atom magnetization in ice at-6 °C, recorded

radiolysis pulse to establish a baseline. In Fig. as a function of time following pulse radiolysis (5 ns, ca.1-15, we show the (unexpected) small inverted (Srec) 400 rad pulses repeated at 360 Hz). The high-field (mI =and noninverted (Sre f) magnetization signals for the -I/2) line is polarized in absorption, and the low-field (ml"baseline" scan at to = -200 Its. Because the reference = +1/2_ in emission (negative signal). The magnetization

present immediately after the pulse is generated bysignal is very constant over this period, a recoveryCIDEP in cross-recombination reactions in radiolysis

time Tr= 23.3 Its can be obtained from a simple expo- spurs. Just prior to each radiolysis pulse, a small signalnential fit to the inversion recovery. The inversion remains from H atoms created in the previous pulse.pulse rotates the initial magnetization signal into itsnegative, so the true zero level can be deduced as

halfway between the inverted and final signal levels, ,,-, 301 A.J_ A_'"as shown. (This information was then used to estab- ._ _ _, _ ,. A i _AA_

lish the zero level in Figure 14, as well.) An identical _ 20-1 _/)v" v ", ___

was performed with the radiolysis dose (electron'_

charge per pulse) cut by half. The recovery times "E

?-2Ir,'''23.3, Iaveraged 23 Its. We appear to have successfully found ,a: . 10- =a low-dose limiting dynamic inversion recovery time aof 23 + 3 Its. Strictly speaking, however, this should u.. 2 0-

be regarded as a lower limit for T r because we were i i i i i i0 20 40 60 80 100 120

not able to test the polarization recovery at even lowerradiation dose. Recovery Delay (l.tsec)

In Figure 16 we plot a typical spin-echo envelopedecay curve recorded at 50 Its after a slightly larger Figure 15. Dynamic inversion recovery of the H atom12 ns, 2.8 nC radiolysis pulse, signal averaged at high-field line, in ice at -6 °C. On alternating radiolysis180 ns Hz. The larger radiation dose could be used shots, a 70 ns microwave inversion pulse was applied to

because T2 is shorter than T I , but signal/noise is produce the "recovery" signal Src c. This experiment isperformed on the very small signal remaining some 2.8

comparable to the T i measurement because a shorter ms after a 40(I tad radiolysis pulse. 3"he true "zerW' levelgate width (200 ns) is used to measure the peak is inferred from halt the initial difference in Sref-Sre c.

19

G)"0" 400

E 300 - _ 1 9-

< '0 IH + f rate

_) 1 8- lEa = 1.8 + 4.6 kJImoleO 200 mJ¢ouJ '7,_. 1oo :Z 17-

"R "_ JH + Br" rate--_._ .L _/'k._,_.., _e_,. ,-

o , , " ',*", ," , -*-_T,, ___ ,_,_l e- j_ = 8.3_ 6.2kJIn_ol,0 10 20 30 40 50 ¢: _/

Echo Delay (psec) .j 1s-14 ......

I I I I IFigure 16. Spin-echo decay envelope of the H-atom 3.1 3.2 3.3 3.4 3.5

high-field line in -6 °C ice at 50 Its following pulse 1/1" IK"1 xlO 3 )radiolysis with ca. 700 rad radiation pulses. An aver-age of nine similar curves at different time delays and ondifferent days gave 6.9 + 1.0 for our best estimate of T2.

2 Figure 17. Diffusional correlation times for free H and D6OL=3.3+0.7× 1015s -2and1;=3.7+0.7 × 1011 s.

- 2 - atoms in several isotopic mixtures of ice-lh. TheThis value of(o was then used in eq. 17 to translate correlation times were calculated from transverse spinprevious measurements of T 2 into motional correla- relaxation times on the basis of the calibration point at -6

tion times of H over a ravage of temperatures. °C, for which the spin-lattice relaxation time was alsoFurthermore, after correcting (oL for isotopic compo- measured.

sition, the same procedure was used to deduce correla-tion times for H and D in other ices. The resulting 4. Solvent Relaxation Dynamics

values of 'r are plotted in Figure 17. Y. Lin, C. D. JonahThe clearest difference is between H atoms and D

atoms. Their behavior is best compared in the same The effects of solvation and solvent dynamics on

lattice (90% D20), but even the other data sets reflect chemical reactions and especially on charge-transfer

the higher activation energy and smaller % for H. processes have long been a subject of great importanceSurprisingly, below ca. 200 K the D atom diffuses in radiation chemistry. Much past work has involved

faster than the H atom. This unexpected inverse iso- comparing equilibrium solvation effects, probed by

tope effect was explained by using transition state the- measuring the shift of optical absorption spectra, with

ory. The adiabatic "bottleneck" at the transition state chemical reactivity. However, chemical reactivity is abetween minimum energy sites is more severe for H dynamic process, and solvation effects will vary as a

than D because of its higher zero-point vibrational function of time. The dynamics of the solvent can be

energy. It is quite possible that the inverse isotope observed by monitoring solvent response to the fast

effect is a very general phenomenon in the diffusion of formation of different charge distributions (either a

hydrogen in solids, dipole or an ion) on a probe molecule. SolvationThe high rate of H and D (and also Mu) diffusion dynamics were first studied by observing the shift of

in ice in the upper temperature range is a consequence the optical spectrum of the simplest ion, the electron.

of the open channel structure of the ice Ih lattice. In the last decade, picosecond time-resolvedGiven the hopping rate at 273 K of 2.8 × 1010 s 1 absorption and emission spectroscopy have been used

and the distance of 3.67 A between minimum energy to measure the relaxation of the solvent around ansites, one can calculate a one-dimensional c-axis dif- excited state. Until our work, there had been little

fusion coefficient of 1.9 × 10-5 cm2 s-1. Remarkably, research on the solvation of charged species in room

this diffusion coefficient is only about a factor of two temperature fluids. This is a major deficiency because

smaller than that for liquid water close to the freezing many chemical reactions in the condensed phase

point, involve the creation or destruction of ions.

20

We have applied pulse radiolytic (PR) techniques solvent molecule. Variable-temperature experiments

to study the solvation of anions in polar media. In our will be very useful for altering the solvation pathways.

approach, a high concentration of the probe molecule, Because there are considerably different activation

benzophenone, which reacts very efficiently with the energies for different solvent molecule motions,

presolvated electron, is placed in the solution. Because measurements at different temperatures will make it

the benzophenone reacts with the precursor of the possible to probe the different processes.solvated electron, the anion is formed very quickly.

The advantage of the pulse radiolytic technique over 1000 -the laser method is that it allows observations to be

made on the ground state of the anion rather than on a + dipole i /"short-lived excited state and, thus, one can measure ,', eloetron

j.

long solvation times. Moreover, complications due to x anion • /+

changing emission lifetimes as a function of _ 100 .,,,,_/()tJwavelength, which can occur using laser techniques, _,,,are not encountered.

a. Pure Solvent Solvation Dynamics10

The relaxation of the solvent polarization due to a " ._1 fllll l I m t lllll I l I t IJJLI

sudden change in the charge distribution of a system 10 100 1000

has recently received much attention. Experimental "q (ps)

and theoretical work have revealed the importance of Figure IH. Solvalion time of tile benzophcnonc anion, thethe molecular structure of the solute and solvent in electron, and a dipole versus tile longitudinal relaxation

solvation dynamics, time of tile solvent, "q.

We studied the dynamics of anion solvation in

several dipolar, noncondt, cting solvents and, in what b. Solvation Dynamics in Electrolyte

may be considered a maior advance, rneasured the sol- Solutions

vation time of anions in room temperature solvents.

Previous work involved dipole solvation or electron We have extended our work on the solvation

solvation. Figure 1-18 compares the solvation time of dynamics of anions in dielectric fluids by studying

anions, electrons, and dipoles to the solvent iongitudi- solvation dynamics in electrolyte solutions. Although

nal relaxation time 'rL. The solvation times obtained ionic effects are well known in chemistry, a quantita-

from electron solvation and anion solvation experi- rive theory does not exist for high ion concentrations.ments are much faster than the solvation times of There will be an additional energetic relaxation process

dipoles. These results clearly show the importance of in ionic solutions. Adding to the stabilization by

measuring ion solvation rather than assuming that the solvent molecular reorientation will be the energysolvation times will be similar to dipole solvation, lowcring caused by the migration of electrolyte ions

The similarity of the solvation time of the anion and under the influence of the charge of the anion -an

the electron (despite the quantum nature of the latter) attraction for the electrolyte cations and a repulsion for

shows up clearly, as do the differences between the the electrolyte anions. This will increase the field at

solvation around an anion and a dipole (despite the the anion and be reflected in the absorption andsimilar molecular frameworkl. The substantial differ- emission spectra of the probe ion.

ence shown in these measurements indicates that there We have studied the dynamics of the bcnzophc-

are fundamental differences between the solvation of none anion solvation from 100 ps to 50 ns in solu-

charged entities and dipoles, tions of Li+, Na +, and Ba+2 in acetonitrilc, propylcneWe will expand this work to measure different carbonate, and in methanol. The solvation of the anion

probe molecules and different temperatures. We plan is too fast to measure in methanol or acctonitrilc. The

to study substituted benzophenone molecules that will absorption naaximum of the species with equilibrium

change either the geometry around the anion or the solvation is 720 nm in acetonitrile and 630 nm ininteractions between the probe molecule and tile methanol.

21

In the presence of an electrolyte in acetonitrile, the 740 -

absorption maximum of the radical anion is sig- 720 __AI_ 1_ 0.25Mo.5M

nificantly blue shifted and the dynamics are signifi- 0.75M1.0M

cantly slower. The transient absorption spectra of the 700 - exp_fit

benzophenone anion in 0.5 M NaCIO4 in acetonitrile

are shown in Figure 19. The spectra were run at 2 ns, _ 680 ____.5 ns, 10 ns, and 15 ns (from right to left), respec-

tively. Data have also been taken using the strobo- _" 660- o

scopic pulse radiolysis system, and the spectral _,_,+ o ""_

responses of the two systems have been corrected. The 640 - "_'"_-.,_ a_ •relaxation as a function of time for an anion in this

salt-acetonitrile solution, as well as for several others, 620 - I I I I I I Iis displayed in Figs. 1-20 and 1-21. As is clear from 0 2 4 6 8 10 12 14

these data, the relaxation time is considerably slower Time(ns)than in the pure acetonitrile. For example, the 0.5 M Figure 20. The peak of the absorption spectrum of the

benzophenone anion vs. time for four differentNaCIO4 solution has an exponential relaxation time of concentrations of NaCIO4 in acetonitrile.2.4 ns, at least two orders of magnitude slower than

the slowest time possible for the pure acetonitrile. As

one can see from the figure, this behavior is typical for [_1ionic solutions. These results show that the relaxation 720

I • Na"processes arising from the rotational motions of the ..] \ I A Li+solvent dipoles are much faster than the solvation 700 I _ I • Ba+2from ion translation or migration, which occurs in the .C 680electrolyte solutions.

The two experimental observables that change in 660our experiments are the spectral shift (or the final _,spectral position) and the time for the spectral relax- 640 -.a

ation. These parameters depend both on the salt iden- 7tity and the salt concentration, as can be seen in 620 /Figure 21. For example, for NaCIO4 the spectral shiftincreases and the solvation time decreases with an ' ' ' _ ' _ _0 2 4 6 8 10 12 14

Time(ns)Figure 21. The peak of the absorption spectrum of thebenzophenone anion vs. time for 0.5 M NaCIO4, LiCIO4,and BaCIO4 in acetonitrile.

0.14

0.12increase in salt concentration. These results are sum-

O.lO marized in Table 3.

_o.oa In addition the salt and concentration,to identity

a change in the solvent also changes the rate and

o.oe amount of the spectral relaxation. These results aretabulated in Table 4 for NaCIO4 in a variety of

o.04 solvents. There appears to be an inverse correlation0.02 between the amount of shift with the neat solvent and

5c _ s_o st_o s_o 7t_o 7_0 a_ the additional shift with the added salt. For example,wavelength in methanol, where the spectral shift is large, the

Figure 19. The spectra of the benzophenone anion in 0.5 additional shift caused by the addition of the salt isM NaCIO4 in acetonitrile at 2 ns, 5ns, 10 ns, and 15 nsafter formation. The spectra show a blue shift as a small. In acetonitrile, where the pure solvent shift isfunction of time. small, the shift caused by the salt is large. These data

22

Table 3. Shift in the maximum of the spectrum, A,, c. Molecular Dynamics Simulations ofand solvation time, Xs, of the benzophenone anion for Solvation in Pure Fluidsdifferent concentrations of NaC104 in acetonitrile.

concentration (M) An(cm'l) Xs Ins) Theoretical studies were undertaken to model the

0.25 1543 , 3.4 results of experimental measurements of anion solva-0.5 1809 2.4 tion times in alcohols. In alcohols, spectra of the

0.75 1834 2.3 equilibrium-solvated anion are virtually the same for

1.01 1859 1.9 n-butanol and n-decanol, and greatly different for n-

butanol and 2-butanol. A continuum theory would

have predicted the oppositenthe spectra would be the

same for the two butanols and greatly different for the

n-butanol and n-decanol. We suggested that theseTable 4. Solvation time and spectral shift for 0.25 M results arise from the packing of the OH bonds around

NaC10 4 in various solvents, the anion and, using Monte Carlo techniques, showed

solvents Xs Av(cm'l) that this _.xplanation is reasonable.

methanol <20 ps -248 Monte Carlo calculations do not probe the kinet-

n-propanol -70 ps -248 ics of solv_tion; instead, they only determine the

acetonitrile 3.4 ns 1543 equilibrium configurations. To calculate the solvation

propylene carbonate 7.9 ns 919 times, a molecular dynamics model was developed.

Molecular dynamics simulations were performed for

216 solvent molecules and a single charged solute inthe center of the simulation box. The dimension of

may indicate that there is a limiting spectral shift, and, the box is 23 nm. The simulation employed a cubic

thus, additional solvating power may not shift the periodic boundary condition with the minimum-image

spectrum further. On the other hand, it may indicate convention. Initially, the solvent molecules were

that in a strongly solvating solvent (e.g., methanol) placed on a face-centered cubic lattice with lattice

the electrolyte ion cannot replace a solvent molecule, parameter 0.31 nm. The systems were studied in both

Two uther groups have recently studied the influ- a constant-energy and constant-temperature ensemble.

ence of ions on the energetics and dynamics of dipole The classical equations of motion were solved using a

solvation by observing the time-dependent shift of the leapfrog scheme for linear rigid molecules. Time steps

fluorescence spectrum. These experiments qualitatively of 2 fs were used in the integration. All the

agree with our observations in that the solvation in simulations reported here were run at 300 K.

ionic sol,'ion is much slower than in tile pure A linear rigid array of n atoms was used as asolvent, and that the relaxation dynamics depend not model of the solvent molecules. The dipole was

only on the solvent studied but also on the identity modeled as partial charges, q+ and q-, on adjacent

and concentration of the salt. The data have been atoms with the positive charge always on the end ofinterpreted using either a Debye-Falkenhagen ion- the molecule. The solvent molecules are electrically

atmosphere approach (DF) or an ion-association (IA) neutral. The short-range part of the intermolecular

model. The DF model is a continuum model whereas interaction was modeled by a Lennard-Jones potential

the IA model emphasizes specific interactions. We between the atoms, and the long-range coulomb inter-

have analyzed our experimental measurements using actions between the charge groups of the solvent were

these two different models, and, in agreement with calculated using a spherical approximation to the

previous workers, found that, at best, only qualitative Ewald summation. The interaction parameters for both

agreement exists between theoretical predictions and solvent and solutes are given in Table 5.the experimental data. The quantitative discrepancy of Both the orientation and the distance of the model

the existing models and experimental observations solvent molecules from the central ion have been

requires the development of more sophisticated probed. We have carried out calculations for both a

molecular-level models that also can incorporate the positive and a negative central ion. Figure 1-22 shows

continuum structure of the fluid, the average orientation of the solvent molecules in

23

Table 5. Characteristics of the different solute and solvent properties simulated. For the solvent, o(,_,) will be the

exclusion volume of the central ion in angstroms and q will be the charge in atomic units. For the solvent, o(A) is

the diameter of each "atom" of the solvent, and E/k is the Lennard-Jones potential parameter for the "atom".

solute properties solvent properties

solute o(A) q(a.u.) solvent o(A) E/k(K) # of atoms

Q0 2.1 0 $2 3.5 190 2

Q1 2.1 1.0 $3 3.5 190 3

Q2 2.1 2.0 $4 3.4 190 4

Q3 2.1 -1

the first solvation shell, <cos(_.r-)>, where _ is the 1.0- _j_,...--,

direction of the dipole in the molecule, and f is the unit _JJ_" --

vector from the central ion to the center of the mole- 0.5- j. ..... ,.._J'",f ,_.,.,,..-m_''

cule. This average would be 1 if all the nearest mole- /," f_-_¢rf j,_" .-,..,.e"_'_'_'0'_...... _ ........... 0

cules pointed directly toward the central ion, 0 if the _" ./_ _d_._ _.molecule is perpendicular to the line from the central _, 0.0- _,_

ion, and -1 if the positive end of the molecule points 8 _._,_._"

directly away from the central ion. For an anion, the _/,molecules are directed towards the central ion, with -0.5 ¢/

stronger alignment for the longer molecules. The _!alignment around a cation is weaker and in the reverse _,direction. The results reflect the balance of energetic -1.0 -0'.5 0_0 0_5 1'.0 1'.5 2'.0

constraints: the advantage of close approach between 120Figure 22. The average angle between the molecular

the negative end of the solvent dipole and the central dipole and the line between the center of the solventcation without colliding, and the repulsion betweeJ_ molecule and the central charge. The lines are drawn onlythe cation and the positive end of the solvent dipole, to guide the eye.

The distance distribution is given in Figure 23 for

solvent molecules with lengths of two, three, and fourmolecular units. R is the distance from the central ion 14

to the center of the dipole, and g(R) is the number of 12

solvent dipoles in a unit volume at a distance R from 10

the central molecule relative to the number that would R

be in the same volume if the solution were _homogeneous (for a homogeneous solution g(R)would be 1). Note that all three model solvent

molecules have their dipoles at the same distance from 0- ," _ _the central ion. This is consistent with the _ _ / I I

experimental data in that the shifts of the spectra are 2 4 6 a _0 _2

independent of the length of the molecules. R(A)Figure 23. The radial distribution function g(R) as a

We calculated the potential at the solute molecule, function of distance R (in angstroms). R is measured fromThese results are shown in Figure 24. For a negative the central ion to the center of the dipole. $2, $3, and $4central ion, the electric potential is virtually the same are solvent molecules of length two, three, and four

at the central ion, independent of the length of the sol- elements, respectively.

vent molecule. This agrees very well with the experi-mental results, which show that the spectrum of the

equilibrium-solvated benzophenone anion is virtually To study the solvation dynamics, the response of

identical for different primary alcohols. Thus, the the solvent to an instantaneous change of the soluteresults show that the local structure _ther than the long- charge distribution is monitored. The solvent is equi-

range structure is dominating the solvation energetics, librated around a neutral solute molecule. The new

24

40- [ ua g_i molecule. That is, the solvent molecule is rotating andlis41 (

" 1 continues to rotate to reach the lowest potential.20- o

,_ • However, for longer molecules, the inertial rotation is_, small, because a molecule collides with other solvent

0- • molecules and the relaxation becomes diffusional.These calculations are beginning to tell us about

the origin of the observed dynamics and energetics.

-20 Future work will probe the specific interactions

' , , • , , -i between the solvent molecules and the central probe-' .0 -0.5 0.0 0.5 1'.0 1.5 2.0 molecule, which give rise to the spectral shift.

Q0

Figure 24. The potential at the central ion as a function ofthe central charge for solvent molecules of length two,three, and four elements.

charge distribution is imparted to the solute molecule,

and the kinetic energy, rotational energy, and the

potential at the solute molecule <V(t)> are observed

for approximately 2 ps. To achieve reasonable statis-

tics, this process is repeated about 500 times.As can be seen in Figure 25, the solvation dynam-

ics are slightly different for the relaxation around acation and an anion. Common to all the systems

studied, the dynamics show two components, a fast

response of 0.1-0.2 ps and a slower relaxation of 1 ps

or greater. The fraction of the fast relaxation is greaterfor the short molecules than for the longer molecules.

The short timescale can be thought to be due to

inertial dynamics (mainly rotational) of the solvent1.0-

0.8-

0.6-

O0.4-

0.2-

0.0-- "1 I I I i

200 400 600 800 1000

Time (fs)Figure 25. The solvation correlation function as a func-tion of time for solvent molecules of three and four

elements and as a function of the charge of the centralion.

25

PUBLICATIONS

DIFFUSION OF ATOMIC IIYDROGEN IN ICE-IhD. M. Bartels, P. Han, and P. W. PercivalChem. Phys. Lett. 210, 129-134 (1993)

ELECTRON SPIN RESONANCE STUDY OF IRRADIATED POLYSTYRENE FIBERS

A. Pla-Dalmau, M. Barnabas, A. D. Bross, A. D. Trifunac

Polymer Preprints Am. Chem. Soc. 34, 343-344 (1993)

RADICAL CATIONS OF QUADRICYCLANE AND NORBORNADIENE IN POLAR ZSM-5 MATRICES:THERMAL RETRO DIELS-ALDER REACTION

M. V. Barnabas and A. D. Trifunac

J. Chem. Soc. Comm. ]0, 813-814 (1993)

EARLY EVENTS IN RADIATION CHEMISTRY AND IN PHOTOIONIZATIONA. D. Trifunac, D. M. Loffredo, and A. D. LiuJ. Rad. Phys. Chem. 42,967-972 (1993)

ELIMINATION AND ION-MOLECULE REACTIONS OF HEXAMETHYLETHANE AND TRIMETHYLBUTANERADICAL CATIONS IN ZSM-5

M. V. Barnabas, D. W. Werst, and A. D. Trifunac

Chem. Phys. Lett. 204,435-439 (1993)

QUADRICYCLANE AND NORBORNADIENE RADICAL CATIONS IN SILICALITE: COMPARISON WITHFREON MATRICES

M. V. Barnabas, D. W. Werst, and A. D. Trifunac

Chem. Phys. Lett. 206, 21-24 (1993)

PHOTOIONIZATION AND ENSUING ION-MOLECULE REACTIONS OF POLYCYCLIC AROMATICHYDROCARBONS IN ALKANE AND ALCOHOL SOLUTIONS

A.-D. Liu, D. M. Loffredo, and A. D. Trifunac

J. Phys. Chem. 9_2_7,3791-3799 (1993)

EPR MEASUREMENT OF THE REACTION OF ATOMIC HYDROGEN WITH Br- AND I- IN AQUEOUSSOLUTION

D. M. Bartels and S. P. MezykJ. Phys. Chem. 97_7,4101-4105 (1993)

TRAPPED HOLES ON TiO2 COLLOIDS STUDIED BY EPRO. I. Micic, Y. Zhang, K. R. Cromack, A. D. Trifunac, and M. C. ThurnauerJ. Phys. Chem. 27, 7277-7283 (1993)

MULTI-PHOTON PROCESSES IN CYCLOHEXANE AND TRANS-DECAHN AND THE FORMATION OF

HIGH MOBILITY CATIONSA.-D. Liu, M. C. Sauer, Jr., and A. D. Trifunac

J. Phys. Chem. 97, 11265-11273 (1993)

ELECTRON CAPTURE RATE CONSTANTS FOR SOLUTES IN ALKANE LIQUIDS MEASURED BYTRANSIENT PHOTOCONDUCTIVITY

M. C. Sauer, Jr. and K. H. SchmidtRad. Phys. Chem., 43,413 (1994)

26

SOLVENT EFFECTS IN NONPOLAR SOLVENTS: RADICAL ANION REACTIONSD. W. Werst

Chem. Phys. Lett. 20__22,101-107 (1993)

EARLY EVENTS FOLLOWING RADIOLYTIC AND PHOTOGENERATION OF RADICAL CATIONS INHYDROCARBONS

D. W. Werst and A. D. Trifunac

Rad. Phys. Chem. 4._3_1,127-133 (1993)

PICOSECOND DYNAMICS OF BENZOPHENONE ANION SOLVATIONY. Lin and C. D. Jonah

J. Phys. Chem. 9__7_7,295-302(1993)

THE DEPENDENCE OF THE BENZOPHENONE-ANION SOLVATION ON SOLVENT STRUCTUREY. Lin and C. D. Jonah

J. Phys. Chem. 9__66,10119-10124 (1992)

TRANSIENT ABSORPTION SPECTRA OF AROMATIC RADICAL CATIONS IN HYDROCARBONSOLUTIONS

A.-D. Liu, M. C. Sauer, Jr., D. M. Loffredo, and A. D. TrifunacJ. Photochem. Photobiol. A: Chem. 6__7_7,197-208 (1992)

INVESTIGATION OF ELECTRON TRANSFER BETWEEN THE HEXAARYLBIIMIDAZOLE AND VISIBLESENSITIZER

Y. Lin, A.-D. Liu, A. D. Trifunac, and V. V. KrongauzChem. Phys. Lett. 19.__8_8,200-206(1992)

MECHANISM OF THE FORMATION OF TRANSIENT AROMATIC RADICAL CATIONS IN ALCOHOLS:LASER FLASH PHOTOLYSIS AND PULSE RADIOLYSIS STUDIES

A.-D. Liu, M. C. Sauer, Jr., C. D. Jonah, and A. D. TrifunacJ. Phys. Chem. 9__66,9293-9298119921

STUDIES OF ION SOLVATION USING PULSE RADIOLYSISC. D. Jonah and Y. Lin

Radiation Research: A Twentieth-Century Perspective, Vol. II: Congress Proceedings,W. C. Dewey et al., Eds., Academic Press, Inc., 1992, pp. 63-68

EXPERIMENTAL INVESTIGATION OF THE DYNAMICS OF BENZOPHENONE ANION SOLVATIONC. D. Jonah and Y. Lin

Chem. Phys. Lett. 19___1.1,357-361(1992)

27

SUBMISSIONS

CHEMICAl_. CONSEQUENCES OF NON-HOMOGENEOUS ENERGY DEPOSITION BY IONIZINGRADIATION

C. D. Jonah

Radiation Protection Dosimetry, in press

THE RATIO OF' TRIPLET TO SINGLET EXCITED STATE FORMATION FROM ION RECOMBINATION IN

THE RADIOLYSIS OF AROMATIC SOLUTES IN ALKANE LIQUIDSM. C. Sauer, Jr. and C. D. Jonah

Rad. Phys. Chem., accepted, 1993

FDMR IN LOW-TEMPERATURE SOLIDS. PART I. THEORYI. A. Shkrob and A. D. Trifunac

J. Phys. Chem. ( !993)

FDMR IN LOW-TEMPERATURE SOLIDS. PART II. OBSERVATION OF WEAKLY-BOUND ELECTRONS

I. A. Shkrob, D. W. Werst and A. D. Trifunac

J. Phys. Chem. (19931

VERY LOW TEMPERATURE STUDY OF RADICAL CATIONS OF QUADRICYCLANE ANDNORBORNADIENE IN Na-ZSM-5 ZEOLITES

K. R. Cromack, D. W. Werst, M. V. Barnabas, A. D. Trifunac

Chem. Phys. Lett. (1993)

PHOTOINDUCED HOLE TRANSFER FROM TiO2 TO METHANOL MOLECULES IN AQUEOUS SOI,UTIONBY ELECTRON PARAMAGNETIC RESONANCE

O. I. Micic, Y. Zhang, K. R. Cromack, A. D. Trifunac, and M. C. ThurnauerJ. Phys. Chem. (1993)

ENCOUNTERS OF H AND D ATOMS WITH 02 IN WATER: RELATIVE DIFFUSION AND REAC'TIONRATES

P. llan and D. M. Barrels

Proceedings, International Workshop on Ultrafast Reaction Dynamics and Solvent Effects,P. Rossky and Y. Gauduel, Eds. (19931

THE DYNAMICS OF ANION SOINATION IN ALCOHOLSY. Lin and C. D. Jonah

UndcrstatJditTg Chemical Reactivity, J. Sitnon, Ed., Kluwcr Publishing (1992_

SOIAD-STATE FDMR FROM SUNDRY POINOXYETttYI.ENE ALKYL PHENYI_ ETHERS: A VERIFICATIONOF THE SINGLE-STEP ELECTRON TtlNNEI.ING AS A CAIISE OF TItE CHARGE MOBII3TY IN FROZENRIGID SOLIDS AT 4-30 K

I. A. Shkrob and A. D. Trifunac

J, Plays. Chem. (1993)

OBSERVATION OF AN AROMATIC RADICAL ANION DIMER: (CIoFg)2"-D. W. Werst

J. Am. Chem. Soc. (1992)

28

29

PRESENTATIONS

SPIN DYNAMICS, REACTION, AND DIFFUSION OF Ii AND D ATOMS IN WATER AND i¢'F.David M. Barrels

Argonne National Laboratory Chemistry Division Monday Morning Seminar SeriesJanuary 18, 1993

RADICAL IONS IN RADIATION AND PHOTOCHEMISTRY. MAGNETIC RESONANCE AND OTIIERSTUDIES

A. D. Trifunac

Plenary Lecture, The First Australian-Asian Conference on Radiation Science and Nucler MedicineFebruary 14-20, 1993

STRUCTIJRE AND DYNAMICS OF RADICAL CATIONSA. D. Trifunac

Invited seminar, IJniversity of Sydney, AuslraliaFebruary 15, 1993

RADICAl. CATIONS IN PHOTOIONIZATION AND RADIOLYSISA. D. Trifunac

Invited seminar, University of Melbourne, AustraliaFebruary 25, 1993

tllGI! ENERGY CllEMISTRY: PHOTOIONIZATION AND RADIOI.YSISA. D. Trifunac

Invited seminar, Joint Physical/Fheoreli,,:al and Organic Chemistry Series, Australian National University,Canberra, AustraliaMarch 2, 1_)93

TRANSIENT SPECIES IN RADIATION AND PI1OTOIONIZATIONA. D. Trifunac

Invited seminar, Universily of Brisbane, Brisbane, AustraliaMarch 15, 1993

RADICAL IONS: INTERMEDIATES IN llIGtt-ENERGY CttEMISTRYD. W. Wersl

Radiation Research Society - 41st Annual Meeting/North American llyperthemfiuSociely 12th Annual Meeting, Dallas, TXMarch 2(I-25, 1993

ANION SOINATION IN AI.COtlOLSC. D. Jonah and Y. Lin

Invited Speaker, Radiation Chemistry Laboratory, Japanese Atomic Energy Research lnslilute, Osaka, Japan,March 29, !993

ANION SOI.VATION IN ALCOItOI.SC. D. Jonah and Y. Lin

lnviled Speaker, Kyoto Institute ot Technology, Kyoto JapanApril 15, 1993

CI.t/STER FORMATION IN "FILEIRRADIATION OF PAI_.LADIUM AND PAIA_ADIIIM-SIINER SOLUTIONC.D. JonahInvited Seminar, Radiation Chemistry Laboratory, Japanese Atomic Energy Research Institute, Osaka, Japan,April It), 1993

CONSEQUENCES OF THE NON-HOMOGENEOUS DEPOSITION OF ENERGY IN RADIOI.YSISC. D. Jonah

Invited Seminar, Radiation Chemistry Laboratory, Japanese Atomic Energy Research Institute, Osaka, Japan,April 22, 1993

30

ANION SOLVATION IN ALCOHOI.,SC. D. Jonah and Y. LinInvited Seminar, Hokkaido University, Sapporo, JapanApril 26, 1993

CONSEQUENCES OF THE NON-HOMOGENEOUS DEPOSITION OF ENERGY IN RADIOI.YSISInvited Seminar, Tokyo University, JapanApril 30, 1993

CONSEQUENCES OF THE NON-HOMOGENEOUS DEPOSITION OF ENERGY IN RADIOLYSISInvited Seminar, Beijing Normal University, Beijing, ChinaMay 3, 1993

ANION SOI.,VATION IN ALCOHOLS

Invited Seminar, Beijing University, Beijing, ChinaMay 6, 1993

COMPARISON OF EARI.Y EVENTS IN RADIATION AND PIIOTOCIIEMISTRYA. D. TrifunacInvited talk, Miller Conference, Bowness-on- Windemere, UKApril 3-8, 1993

A I,ASER DRIVEN PICO tAND SUBPICO??t SECOND EI,ECTRON I.INACA. D. Trifunac and C. D. JonahInvited talk, Miller Conference, Bowness-on-Windemerc, UKApril 3-1',I,1993

RADICAL CATIONS: KEY INTERMEDIATES IN iilGIt ENERGY CItEMISTRY AND IN ZEOLITESA. D. Trifunac

Chemistry Division Monday Morning Serninar Series, Argonne National l,aboratoryApril 26, 1993

TltERMODYNAMICS AND TRANSPORT PROPERTIES OF TIlE IlYI)RATF.I) EI.ECTR()ND. M. Barrels

IBM Research Facility, Zurich, SwitzerlandMay 5, 1993

EPR, SPIN DYNAMICS, AND DIFFUSION OF It ANI) D ATOMS IN WATER AND ICI:.D. M. Barrels

Invited Speaker, Physical Chemistry Seminar, University of Zurich, SwitzcrhmdMay 6, 1993

EPR, SPIN DYNAMICS, AND DIFFIJSION OF !t AND I) ATOMS IN WA'IER AND ICED. M. BartelsInvited Speaker, Radiation Chemistry Seminar, Interuniversity Reactor hlslittllc, Delft, The NelhcrhmdsMay 10, 1993

IIYDROPtlOBIC BARRIER IN TIlE REACTION OF !I ATOMS WITIi 02 IN WATERD. M. Barteis

Invited Speaker, International Research Workshop on I31lrafast Reaction I)ynamics and Solvent f'_ffcctsAsnieres-sur-Oise, FranceMay 12, 1993

ELECTRON SPIN RESONANCE STIJDY f)t: IRRADIATION F'OI.YSTYRENI-:, FIBERSA. Pla-Dalmau, M. Barnabas, A. D. Bross and A. I). TrifunacACS National Meeting, Chicago, ILAugust 22-27, 1993

31

EPR MEASUREMENT OF THE DIFFUSION OF H AND D ATOMS IN ICED. M. Bartels

ACS National Meeting, Chicago, ILAugust 22-27, 1993

THERMODYNAMICS AND TRANSPORT PROPERTIES OF THE HYDRATED ELECTRONDavid M. Bartels

Invited Seminar, Department of Chemistry, Johns Hopkins University, Baltimore, MDOctober 4, 1993

DYNAMICS OF ION SOLVATIONY. Lin

Chemistry Division Monday Morning Seminar Series, Argonne National LaboratoryNovember i, 1993

THERMODYNAMIC, TRANSPORT AND REACTIVITY PROPERTIES OF THE ttYDRATED ELECTRONDavid M. Barrels

University of Notre Dame, Notre Dame, INNovember 18, 1993

RADICAL ION REACTIONS STUDIED BY MATRIX-ISOLATION AND TRANSIENT METttODSDavid Werst

Chemistry Division Monday Morning Seminar Series, Argonne National LaboratoryNovember 29, 1993

33

COLLABORATIONS

R, Cooper, University of"Melbourne, AustraliaStudies of transient species created in gases, liquids, and solids are being carried out. Recent eftorts have concentratedon the spectra and dynamics of excited species created in solids, such as sapphire, which are invcstigatcd via time-resolved measurements of emitted light using the techniques of pulse radiolysis and laser flash photolysis. Sapphire isa prime contender for a lining material of the containrnent vessel in fusion reactions, and these studies are useful inunderstanding radiation effects.

B. Garrett, B:lttelle Pacific Northwest Laboratory, Richland, WATheoretical modeling of H-atom reactions in gas phase and water is being pursued.

S. Mezyk, Atomic Energy of Canada, Ltd,, Pinawa, CanadaReaction rates of H atoms with I- and other iodine-containing solutes are being measured tbr the purpose of predictingthe spread of radioactive iodine in potential nuclear accidents.

P. W. Percival. Simon Fraser Llniversity and TRIUMF, Burnaby, British Columbia, CantataInvestigation of I-I and D atoms in ice by time-resolved EPR has allowed determination of diffusion rates, and thedevelopment of a model H atom-H20 interaction potential. The aim of the collaboration is improved understanding ofH-atom transport and reactivity in solid lattices.

E. Roduner, Physical Chemistry Institute, Zurich University, Switzerland

Reaction of H atoms with benzene in water has been investigated in a test of the validity of transition state theory tbrreactions in aqueous solution. Precision measurements of the hyperfine coupling of H and D atoms in water have beenmade to elucidate the nature of hydrophobic hydration and effects of water "structure" on reaction rates.

C Romero, University of Santiago, ChileThe properties of the hydrated electron are being studied via molecular dynamics simulations. A computational pro-

gram has been developed to simulate aqueous LiCI solutions, and the results are compared to the experimental opticalabsorption spectrum. Water molecules are treated classically, and the electron is given a full quantum treatment. Thisis an important step towards simulating chemical reactions in ionic solutions.

; II

.. II