Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

430 Charged Particle and Photon Interactions with Matter

at3500cm

−1

due to the physisorbed water decreases on heating to reveal two OH surface bands

at 3780 and 3680cm

−1

. These bands are due to the terminal and bridged OH groups, respectively

(Merle-Mejean

et al., 1998). Absorption at the low energies is usually due to solid lattice modes.

Proton

NMR has been used to identify the OH groups on the surface of SiO

. Silica can have

several different OH groups depending on whether silane or siloxane groups are present. Thomas has

used this technique to observe three different OH groups (Thomas, 1993). Ceramic oxide powder

surfaces have also been probed with neutron and Raman scattering to characterize the OH sur-

face sites (Ozawa et al., 1997). However, highly penetrating probes cannot distinguish whether the

observed sites are on the surface or in the bulk. Water trapped in interstitial cavities or co-crystallized

during powder formation will be observed with NMR, EPR, x-ray, or neutron techniques.

The dissociation of water at a particular site on the ceramic oxide is due to the specic activity

at that site. For instance, ZrO

possesses acidic, basic, oxidizing, and reducing properties (Tanabe,

1985). Both Lewis and Bronsted acid sites have been identied on ZrO

(Hertl, 1989). A wide

variety of ceramic oxides exhibit both acid and basic character (Boehm, 1971). Sustained heat-

ing will drive the water off the surface site leaving a Lewis acid due to the unsaturated cation site.

Attachment of an available anion will quickly follow. Irradiation may also lead to dissociation of

a surface

group leaving a reactive cation site.

Surface-bound

OH groups essentially form a layer of H atoms covering the ceramic oxide. The

next layer of water molecules is highly inuenced by this interface and forms a structured water–ice

layer through hydrogen bonding, Figure 16.1. Infrared observation of Al

surfaces with increas-

ing amounts of relative humidity show that the next 2–3 water layers adsorbed on top of the mono-

layer of OH groups are very ice-like. The infrared spectra show a very structured ice-like layer

due to the hydrogen bonding network. A further increase in relative humidity results in a more

disordered transition layer in which the water is more liquid-like (Al-Abadleh and Grassian, 2003;

Thomas and Richardson, 2006). A somewhat similar result was observed in PuO

(Stakebake, 1967,

1981;

Stakebake and Dringman, 1968a,b).

Water

has three degrees of freedom so a typical IR spectrum should reveal two stretching modes

and a bending mode. The two stretching modes are close together and often merge into one broad

peak. Figure 16.2 shows that the absorbances due to the stretching and bending modes are readily

observable in amorphous ice at 3280 and 1680 cm

−1

, respectively (Hudgins et al., 1993). A combina-

tion peak is also observed at 2220cm

−1

and a libration due to hindered rotation is seen at 790cm

−1

Many of the features observed in amorphous ice are also observed from water adsorbed on ZrO

Figure 16.2, and other ceramic oxides. These results suggest that the rst few water layers on these

oxides are very ice-like; that is, they have hindered mobility. The amount of physisorbed water

increases with increasing water content and the spectra reveal more bulk water features. Figure 16.2

shows the DRIFT spectra for ZrO

on the addition of a drop of water to the surface to give “wet”

ZrO

. The distinctive water adsorption peaks for physisorbed water are broadened compared to that

for ice due to the disorder of the liquid-like phase. The change from the OH surface layer through

the transition ice layer to the bulk water layer is not abrupt. Different ceramic oxides are expected

to have different thicknesses of the transition ice-like layer. However, almost all ceramic oxides can

be expected to have a layered structure of water near its surface. These layers are expected to affect

the radiation chemistry of the associated water molecules and to inuence the reactions of other

molecular

species migrating toward the interface from the liquid phase.

16.4 radiolysis oF aQueous suspensions

The main stable reducing species in the radiolysis of water is H

, while the main oxidizing species is

. Both products are due to intra-track reactions of species formed by the water decomposition.

At about 1μs following the passage of a gamma ray, the yield of H

is about 0.45 molecule/100eV

of energy absorbed in liquid water (Pastina et al., 1999). The yield of H

under similar radiolytic

conditions is about 0.7 molecule/100eV (Pastina and LaVerne, 1999). These yield values are maxima.

Radiation Chemistry of Water with Ceramic Oxides 431

At longer times, the reactions of H

with OH radicals and H

with hydrated electrons or H atoms

lead to a series of reactions to reform water (Pastina and LaVerne, 2001). These back reactions can

be suppressed if the H

escapes from the liquid water phase, for instance by vaporization into a

headspace or purged away by a gas ow. Selected scavengers are routinely added to aqueous solu-

tions in radiation chemistry studies to stop the back reactions and allow the measurement of stable

product yields. Essentially no H

is observed at long times in the radiolysis of closed systems of pure

bulk

liquid water with gamma rays (Allen, 1961).

large number of the early studies on aqueous suspensions focused on the photolysis of water

with metal colloids because of the importance in solar energy conversion and in environmental

remediation by photooxidation (Rabani et al., 1988). Many of the rst papers on the radiolysis of

aqueous solutions of metals and metal oxides were logical extrapolations of these early systems,

especially the TiO

and SiO

suspensions. An obvious advantage to examining aqueous suspensions

is that pulsed radiolysis techniques can be utilized to examine transient species. The radiolysis of

acidic solutions of alcohols containing TiO

nds that excess electrons can be transferred from

the aqueous phase to the solid. Both the hydrated electron and the alcohol radical promote elec-

trons to the TiO

conduction band. Conduction band electrons can then react with added solutes.

These results are important because they prove that the transfer of energy and charge through the

water–oxide interface is possible (Safrany et al., 2000; Gao et al., 2002, 2003). Interestingly, these

studies show no effect due to particle size. The excess electrons in the conduction band can even

reduce aqueous Pt to form Pt clusters on the TiO

surface (Kasarevic-Popovic et al., 2004). The Pt

clusters can then reduce aqueous H

to H

at the surface of the TiO

(Behar and Rabani, 2006). The

scavenging of hydrated electrons or radicals by TiO

particles is not that surprising. With the right

redox potentials, the particle is just behaving like a typical aqueous solute. However, the potential

for

initiating new chemistry at the surface of the particles makes these systems very interesting.

one increases the concentration of particles in aqueous suspensions, more energy will be

deposited initially into the solid phase. A fundamental inquiry concerns whether that energy can

migrate to the surface and become available for initiating chemistry in the aqueous phase. Meisel

and coworkers used pulse radiolysis techniques to examine the hydrated electron yields in the aque-

ous phase with variation of SiO

concentration and particle size (Schatz et al., 1998). Some of

the data of that work for 7 nm SiO

is shown in Figure 16.3. The system was probed using pulsed

electron radiolysis techniques with a xed sample size. With increasing concentration of SiO

, the

relative electron density of the sample increases and the amount of energy deposited by the electron

beam pulse increases proportionately. In converse, the volume of water in the sample must decrease

with increasing SiO

concentration. The upper and lower solid lines of Figure 16.3 show these

values, respectively. Hydrated electrons, which are obviously in the water phase, have yields that

track with the electron density of the total medium. If energy deposited initially within the oxide

remained there, then the expected hydrated electron yields should vary with water content as predicted

by the lower curve in Figure 16.3. The results show that not only do the electrons escape the SiO

but the ionization yield within the oxide must be very similar to that within water. Monte Carlo track

calculations indeed predict that the average ionization yield of SiO

is 4.11 ions/100 eV, which is in

excellent agreement with the measured yield of hydrated electrons in water (Milosavljevic et al.,

2004). The pulse radiolysis experiments found no signicant variation in hydrated electron yields

with particle sizes of 7–22nm, suggesting that the electrons readily escape into the aqueous phase.

No fast time resolved experiments on the observation of the hydrated electron have been published,

but preliminary experiments have shown that the hydrated electron is produced within 8ps in aque-

ous suspensions of up to 40% SiO

, suggesting that electrons escape the SiO

with more than just

thermal energy (LaVerne et al., 2003). In other words, the carrier of the charge or energy within the

particle

must migrate very quickly to and through the interface with water.

Concurrent

with the production of an electron, a hole must also be formed in the solid phase.

Holes within the solid phase are difcult to examine in aqueous suspensions, but should the holes

migrate to the particle surface, they may react with solutes adsorbed on that surface. The oxidation

432 Charged Particle and Photon Interactions with Matter

( )SCN

−

and

Fe CN( )

−

on the surface of SiO

by radiolytically generated holes has been exam-

ined using electron pulse radiolysis techniques (Dimitrijevic et al., 1999). The oxidized products

are shown in Figure 16.3 and they can be seen to follow the volume fraction of the water in these

systems. These results suggest that the holes produced in the SiO

make it to the surface, but do

not cross into the aqueous phase. At the surface, they can initiate the oxidation of surface-bound

species (Dimitrijevic et al., 1999). Most commercial suspensions have substantial concentrations

of stabilizers that may react at the surface and interfere with the desired chemistry. Long time

studies of the hydrated electron will exhibit a pseudo rst-order decay due to the reaction with

additives.

Molecular hydrogen is the most common stable product examined in aqueous suspensions. The

driving force for many of these experiments is the same as in photochemical studies—the conver-

sion of radiant energy into chemical energy. A series of experiments have observed the change in the

production of H

in aqueous suspensions as a function of added TiO

and Al

(Yamamoto et al.,

1999; Seino et al., 2000, 2001a,b; Chitose et al., 2003). In all cases, an increase in the amount of H

generation over that of water alone is observed. Particle size ranged from 7 to 200nm with slurries

of 0.01% to 2% weight of ceramic oxide. The production of H

is independent of particle size for

particles less than a few tens of nm, but a substantially reduced yield of H

is observed at 200 nm.

Somewhat similar results are observed with SiO

suspensions, where above-normal yields of H

are

observed for 7 and 12nm particles and a decreased yield is observed for 22nm particles (Meisel,

2003). Yields are found to be linearly dependent on surface area suggesting that H

is formed at

the water–oxide interface (Seino et al., 2001a,b). However, large surface areas can be attributed to

small particle sizes and the observed increase in H

with decreasing particle size may be due to the

increased probability for H

precursor escape to or through the interface. High dose rates decrease

yields, presumably by increased second order reactions. The slurries of TiO

are very efcient

oxidizing agents in phenol degradation (Chitose et al., 2003). Studies with aqueous suspensions con-

taining a variety of ceramic oxides show that the efciency of H

production increases with particle

surface area and that the yields are greater than that of normal water (Cecal et al., 2008). In all of

the studies on suspensions, an increase in H

yield is found with increasing particle concentration.

Aqueous suspensions of TiO

with molar concentrations of methanol are found to give more H

0 10 20 30 40

0.6

0.8

1.0

1.2

1.4

Normalized values

SiO

(wt%)

Relative electron

density

O volume

fraction

Figure 16.3 The radiolysis of aqueous SiO

suspensions relative to that for pure water as a function of SiO

weight percent: (◾) hydrated electron concentration (Schatz et al., 1998); (

⚫

) hole production (Dimitrijevic et al.,

1999); (solid line) electron density; (dotted line) water volume fraction. (Reprinted from Schatz, T. et al., J. Phys.

Chem. B, 102, 7225, 1998; Dimitrijevic, N.M. et al., J. Phys. Chem. B, 103, 7073, 1999. With permission.)

Radiation Chemistry of Water with Ceramic Oxides 433

than without methanol (Jung et al., 2003). However, in these systems, methanol is probably being

sacriced at the surface of the particles and it is the source of H

rather than a modication in

the

water chemistry.

16.5 radiolysis oF adsorbed water

The yields of H

in the radiolysis of suspensions and slurries of ceramic oxides are enhanced com-

pared to that found in bulk liquid water. They also show a direct dependence on the particle surface

area suggesting that the water–oxide interface has an important effect on the radiation chemistry

of the surrounding water. However, even the most concentrated slurries have far too much water to

observe the chemistry occurring at the water–oxide interface without using special techniques such

as probe molecules on the particle surface. Pulsed radiolysis studies on aqueous suspensions suggest

that many of the electrons formed in SiO

migrate into the bulk water while the holes tend to remain

on the particle surface (Schatz et al., 1998; Dimitrijevic et al., 1999). However, the actual yield of

electron-hole pairs formed in ceramic oxides is not known. The recombination of an electron-hole

pair on a water molecule at the surface of a particle can lead to H

production by a dissociative pro-

cess (LaVerne and Pimblott, 2000). Even a single low-energy electron can lead to H

formation from

water by a dissociative attachment reaction (Rowntree et al., 1991; Kimmel et al., 1994; Corbut et al.,

1996). Such a process has been attributed to the observation of H

from water on silicon surfaces

(Klyachko et al., 1997). Exciton (electron-hole pair) formation and migration to the solid surface is

another potential source of H

production. The results are conclusive in that chemistry can occur

at the water–oxide interface, but the details remain unresolved. An excellent method to examine

radiation effects occurring on or very near the interface of water and ceramic oxides is to remove

most of the excess water and observe the chemistry at surfaces with a few layers of adsorbed water.

The rst studies on the radiolysis of adsorbed molecules examined the decomposition of organ-

ics on a variety of mineral solids, especially silica gel, in part because of interest in application to

petroleum rening. Allen and coworkers found the decomposition of the organics to be enhanced

in the adsorbed state as compared to the liquid or gas state and attributed the increase in radiolytic

yields to an energy-transfer process involving an exciton (Caffrey and Allen, 1958; Sutherland and

Allen, 1961; Rabe et al., 1964, 1966). They noted that the enhancement effects were greatest with

wide band gap solids and almost nonexistent with semiconductors. Sagert and coworkers noticed

that replacement of silanol by siloxane groups on silica gel had a large effect on radiation yields.

This work and several others attributed the increase in the decomposition of the organics to

electron-transfer processes from the solid to the adsorbants (Rojo and Hentz, 1966; Wong and

Willard, 1968). Khare and Johnson used an inorganic matrix system to denitively show that the

transfer of energy can depend on the band gap and illustrated the role of excitons in the process

(Khare

and Johnson, 1970).

Most

of the early studies on the radiolysis of adsorbed water were performed in the former Soviet

Union. The rst published studies on the radiolysis of adsorbed water were performed on semicon-

ductors and silica gel (Bubyreva et al., 1966; Krylova and Dolin, 1966). This work has been followed

by others examining the radiolysis of water adsorbed on a variety of surfaces and the production

of H

has been observed (Garibov et al., 1982a,b, 1984, 1987a,b, 1990, 1991a,b, 1992; Rustamov

etal., 1982; Garibov, 1983; Nakashima and Tachikawa, 1983, 1987; Nakashima and Aratono, 1993;

Nakashima and Masaki, 1996). All of these studies found that the type of oxide has a huge effect

on H

yields. The production of H

is by far the most common process examined in the radiolysis

of adsorbed water. This product is stable and relatively easy to measure. The main stable oxidizing

species produced in the radiolysis of water, H

, is known to react thermally at ceramic oxide

surfaces (Hiroki and LaVerne, 2005). Few studies have been reported on the radiolysis of H

the presence of ceramic oxides even though the process may be extremely important in nuclear

power

reactors.

434 Charged Particle and Photon Interactions with Matter

A very thorough examination of the effects of various types of ceramic oxides on the production of

from adsorbed water has been performed (Aleksandrov et al., 1991; Petrik et al., 1999, 2001).

These studies nd that most ceramic oxides fall into three categories when the yield of H

is deter-

mined with respect to the energy deposited in the water phase alone. These categories are as

follows: oxides in which the yield of H

is lower than that of water (MnO

, Cr

, Co

, and ZnO)

with yields of 0.001–0.01 molecule of H

per 100eV; oxides in which the yield of H

is comparable

to that for water (SiO

, V

, NiO, CuO, CdO, TiO

, and the alkaline earth oxides) with yields of

0.1–1 molecule of H

per 100eV; and oxides in which the yield of H

is greater than that of water

(La

, MgO, ZrO

, and Er

) with yields of 50–100eV per 100eV. The researchers conclude that

the physisorbed water has little effect on H

production and that surface OH groups also have little

effect. They claim that most of the H

is due to chemisorbed water, the highly structured ice-like

layer

between OH group surface layer and the physisorbed liquid-like layer.

few observations should be made about the determination of H

yields from adsorbed water.

The previous section on the radiolysis of aqueous suspensions noted that the density of the sys-

tem, and thereby the energy absorbed, increases with increasing oxide concentration. The dose in

aqueous suspensions is usually reported with respect to the energy deposited to the total system,

oxide and water. Product yields are expected to change with variation in oxide concentration simply

because of density and volume effects. The yields from the radiolysis of adsorbed water often report

the dose with respect to the energy absorbed by the water alone. This method is often favored in sys-

tems

of adsorbed water because a direct representation of energy transfer between the water–oxide

phases is obtained. A yield of H

lower than that observed in liquid water suggests energy or charge

transport into the oxide from the liquid or some sort of quenching process by the oxide surface. A yield

of H

greater than that observed in liquid water indicates energy transfer from the solid to the liquid or

other

enhancement in water decomposition.

The

amount of water on an oxide surface is usually only a few weight percent at most and difcult

to measure accurately (LaVerne and Tandon, 2002). A major problem in dosimetry with respect to

the energy deposited directly into the water is quantitatively determining the amount of adsorbed

water for each type of ceramic oxide and in each experimental condition. Figure 16.4 shows a typical

result for the yield of H

from water adsorbed on ZrO

as determined by the energy deposited in the

0.0 0.5 1.0 1.5

100

125

150

175

ZrO

γ-rays

G(H

) (molecules/100 eV in water)

Water loading (wt%)

0.000

0.025

0.050

0.075

0.100

0.125

0.150

0.175

G(H

) (molecules/100 eV in total)

Figure 16.4 Yield of H

in the gamma ray radiolysis of the water adsorbed on ZrO

as a function of the

water weight percent: (

⚫

) yield relative to the energy deposited in the water alone; (◾) yield relative to the energy

deposited

in both the water and oxide.

Radiation Chemistry of Water with Ceramic Oxides 435

whole system (ceramic oxide and water) or deposited in the water alone (LaVerne and Tandon, 2002).

The amount of water in this example is equal to the physisorbed water as determined by the weight

gain following adsorption from constant humidity chambers. The yield of H

with respect to energy

deposited directly into the water is on the order of a few hundred molecules per 100eV of energy.

This yield is huge and it is too large to be due only to water decomposition by the energy directly

deposited in that phase. A yield of 100 molecules/100eV corresponds to 1eV for the formation of one

molecule. This amount of energy is much too small to account for bond breakage and atomic rear-

rangement and indicates that energy is being transferred from the solid phase to the water. A yield of

100 molecules/100eV would seem to imply the prolic production of H

, remember that the yield for

bulk water is 0.45 molecules/100 eV. However, the amount of water in these systems is extremely

small. On an absolute scale, the yield of H

is only about 0.08 molecules/100 eV when the dose is

determined with respect to the energy deposited in the whole system (ceramic oxide and water),

Figure 16.4. Dosimetry in the radiolysis of adsorbed water is commonly reported using both

methods for energy deposition, in the water alone or in the water and ceramic oxide, depending

on the point being made.

The yields of H

exhibit interesting trends as a function of the amount of water loading that can

be used to show the effects due to the type of ceramic oxide. Figure 16.5 gives the yields of H

from

water adsorbed on ZrO

and CeO

as a function of the water weight percent (LaVerne and Tandon,

2002). Dosimetry in this example is with respect to the total energy deposited in both the water and

the ceramic oxide. The results show a slight increase in H

yield with increasing water fraction, and

the yields with ZrO

are distinctly higher than that observed with CeO

. Another interesting feature

in this presentation of the data is the extrapolated yield of H

at zero water fraction. Obviously, no

can be produced without some form of water on the particle surface. The water in this example,

Figure 16.5, is determined from the weight gain of physisorbed water. A dissociated water layer and

an ice-like layer are strongly bound to the oxide surface beneath the physisorbed water layer. These lay-

ers of water near to the particle surface appear to be responsible for a large portion of the observed

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

γ-rays

CeO

ZrO

Surface effect

Yield from increasing water layers

Surface effect

G(H

) (molecules/100 eV in total)

Water loading (wt%)

Figure 16.5 Production of H

relative to the amount of total energy (water + oxide) deposited by gamma

rays as a function of the weight percent of physisorbed water on the oxides: (

⚫

) CeO

, (◾) ZrO

(LaVerne and

Tandon, 2002). The extrapolated H

yields at zero water loading represent the production due to dissociated

water

and ice-like water at or near the oxide surface.

436 Charged Particle and Photon Interactions with Matter

difference in the yields of H

. Water in the physisorbed layer is probably behaving like normal bulk

water when irradiated. Clearly, the water–oxide interface is playing an important role in water

decomposition

and in the production of H

Much of the literature on the radiolysis of adsorbed water has focused on the effects of the type

of oxide on the production of H

at the water–oxide interface. Following an extensive investigation

many types of oxides, the observation was made that oxides within a narrow band gap energy at

about 5eV can lead to a signicant increase in H

(Petrik et al., 1999, 2001). The dose in these stud-

ies was determined with respect to energy deposited in water alone and some of those results are

reproduced in Figure 16.6. The narrow band gap energy response for the large increase in H

yield

suggests that a resonance process is responsible. An exciton may possibly migrate to the particle

surface and resonantly couple to a specic state of a water molecule that is favorable toward the pro-

duction of H

. However, not all oxides within this band gap lead to an increase in H

yields, so the

band gap cannot be the sole characteristic for enhanced H

production. Furthermore, a resonance

process implies a coupling between the oxide and the water. Figure 16.6 shows the dissociative

electron absorption (DEA), cross section, and the differential dipole oscillator distribution (DIOS)

for water ice as functions of the absorbed energy (LaVerne and Mozumder, 1986; Klyachko et al.,

1997). The energy dependence of neither of these processes matches well with the proposed band

gap resonance energy. Special states of water may be formed when they are in the vicinity of certain

ceramic oxides. Of course, resonance and a variety of other processes may be contributing in vari-

ous magnitudes to the radiolytic formation of H

from adsorbed water. Further work is required to

understand

this important phenomenon.

The

mechanism for the transport of energy to the surface of the ceramic oxide is still in ques-

tion. The transport medium in the radiolysis of ZrO

has been suggested to be excitons, which are

electron-hole pairs (Petrik et al., 2001). One method for identifying the nature of the energy carrier

is by the addition of dopants that can trap the transient species within the particle. An increase in

the number of trapping sites should decrease the probability of excitons reaching the surface and

thereby decrease the observed H

production. The substitution of Nb

for Zr

sites in ZrO

at an

amount of only 0.1% by mass gives rise to a dramatic decrease in H

yields to essentially zero.

Substitution of a small amount of Li

or Rb

ions in ZrO

is found to increase the H

production

0 2 4 6 8 10

100

120

140

Relative yield or intensity

Energy (eV)

DEA

DIOS

Figure 16.6 The relative yield of H

(

⚫

) (Petrik et al., 2001) determined from the energy deposited in the

water alone as a function of the band gap energy for several oxides. The dashed lines are the relative cross

sections for dissociative electron attachment (Klyachko et al., 1997) and the differential dipole oscillator

strength distribution for amorphous ice at specic energies (LaVerne and Mozumder, 1986). (Reprinted

from Petrik, N.G. et al., J. Phys. Chem. B, 105, 5935, 2001. With permission.)

Radiation Chemistry of Water with Ceramic Oxides 437

by a factor of 2 (Petrik et al., 2001). These results suggest that the precursor to H

production is

due to an electronic excitation such as an exciton that diffuses through the solid to the interface.

The attenuation of secondary electrons would require much larger concentrations of dopants than

observed. However, these results have to be resolved with the radiolysis of suspensions of SiO

where there is strong evidence that nonthermal electrons directly escape the particle. A competition

must exist between exciton transport to the surface and trapping by defect sites within the particle.

Excitons are self-trapped in SiO

within 250fs so they obviously cannot travel a large distance

through the particle (Saeta and Greene, 1993; Zhang et al., 1997). The different energy carriers may

exist in most of the ceramic oxides, but they may have dissimilar degrees of importance and future

studies

will have to resolve the issue.

Studies

on the production of H

from SiO

of different surface areas suggest that the transport

is limited to the outer layer of 100nm thickness (Aleksandrov et al., 1991). Other studies with SiO

suggest

that particles smaller than 22

have complete transfer of energy to or through the water–

oxide

interface (Milosavljevic et al., 2004). Thomas and coworkers found a dramatic decrease in

yields on variation of SiO

particles size from 6 to 100nm (Zhang et al., 1997). Exciton migra-

tion distance is estimated to be only 5nm in ZrO

(Petrik et al., 2001). Nonthermal electrons may

travel relatively large distances in solid media, whereas excitons must be formed very near to the

surface in order to react with adsorbed molecules. The transport layer thickness may also depend

on the morphology of a given ceramic oxide. Differences in electronic transport are observed in

the monoclinic and tetragonal forms of ZrO

(Lemaignan, 2002). These differences may explain

the increased production of H

observed in tetragonal ZrO

as compared to the monoclinic form

(LaVerne, 2005). This latter dependence of yields on morphology is especially important in nuclear

reactor technology as radiation is known to change the microscopic crystalline structure of certain

ceramic oxides (Lemaignan, 2002). Radiation aging of reactor components could modify the chem-

istry at the water–oxide interface and thereby greatly increase the difculty in making long-term

predictions.

16.6 radiolysis oF ConFined water

Water can be in intimate contact with surfaces while conned to pores or cavities within the ceramic

oxide. This environment is different than that of adsorbed water in that the chemistry of the nor-

mally reactive species such as the hydrated electron, H atom, and OH radical can be modied due

to the conned space. Diffusion-controlled reactions of radicals can be restricted within a cavity

because of the small number of molecules. The passage of species in and out of cavities may be

limited by the diameter of the cavity or the size of the openings to the cavity. Reaction of species

between cavities can be prohibited or predetermined by some characteristic such as size or local

acidity. Many natural materials such as the zeolites have distinctive cavity sizes and shapes and

were among the rst to be studied using radiolysis (Zhang et al., 1998). The interest in these materi-

als followed naturally from their use as catalysts in a wide variety of circumstances especially in the

petroleum industry. Recent fabrication techniques have led to an array of porous materials speci-

cally

designed for research and industrial applications.

The

rst instances of the radiolysis of conned water made use of wet silica gel, alumina, and

quartz (Emmett et al., 1962; Weeks and Abraham, 1965). Gamma radiolysis of these materials

at 80K is relatively easy to perform and results in the trapping of the H atoms produced by the

water decomposition. Electron paramagnetic resonance (EPR) techniques give a readily identi-

able H atom signal because of their huge coupling constant. Interstitial sites have been identied as

Bronsted

or Lewis acid sites.

The

radiolytic production of the H atoms observed in the EPR experiments is derived from the

water incorporated in the cavities. Specically, the adsorbed water is in the form of siloxyl groups

(≡≡Si–OH). EPR studies on silica have provided extensive knowledge on the formation, migra-

tion, and reaction of H atoms (Shkrob and Trifunac, 1996, 1997; Chemerisov and Trifunac, 2001;

438 Charged Particle and Photon Interactions with Matter

Chemerisov et al., 2001). The solid phase makes up most of the mass of these systems so the initial

radiolysis process in silica is the production of the electron and hole, SiO

→ e

−

+ h

, followed by

−

+ h

→

exciton. Combination reactions of the electron and hole lead to the triplet exciton forma-

tion with a mean lifetime of about 150fs (Audebert et al., 1994). The excitons, electrons, and holes

are the main mobile species on the very short timescale. They will be trapped by interstitial sites or

diffuse to the surface where they may initiate chemical reactions with the conned water. As men-

tioned before, self-trapping of the exciton in SiO

occurs within 250fs, so excitons will not move

far (Saeta and Greene, 1993). On the other hand, many porous materials have very large surface-to-

volume

ratios so excitons can readily reach the water–oxide interface.

low temperatures, excitons produced by the radiation migrate to the silanol group giving the

atom and an oxygen hole center, ≡≡SiOH +

exciton → ≡≡SiO

•

+ H

•

. The H atoms are trapped at

temperatures below about 150K and are observable using EPR. As the temperature is raised, H atoms

become very mobile in silica and they decay mainly by recombining with oxygen hole centers or

polarons in silica without added water (Shkrob and Trifunac, 1996). H atoms may be produced by

reactions of thermal electrons with interstitial protons at higher temperatures, ≡≡SiOH

Si≡≡ +

−

→ ≡≡SiOSi≡≡ + H

•

, or ≡≡SiOH

+ e

−

→ ≡≡SiOH + H

•

. Even “dry” silicates produce a signicant

amount of H

due to reactions of H atoms produced from the hydroxyl surface groups (Rotureau

et al., 2005). H atoms and excitons can migrate to cavity surfaces and initiate chemical reactions

with conned water (Shkrob et al., 1999). Electrons and holes are more likely to recombine in the

bulk solid unless they are formed very near the surface. Obviously, energy directly deposited in the

conned water will also lead to product formation. Competition reactions leading to product forma-

tion seem to be controlled by migration of species between the phases as well as pore size, shape,

and

water fraction.

Further

evidence that free electrons are involved in the radiolysis of interstitial water is shown

by the production of N

from zeolites with N

O (Nakazato and Masuda, 1986; Aoki et al., 1988).

More importantly for water radiolysis, the yield of H

is found to decrease with the addition of

O and increase with the addition of ammonia. The N

O scavenges the electron precursors to H

while the ammonia scavenges the positive hole centers to decrease recombination reactions. Zeolite

slurries are observed to give H

yields of about 1–2 molecules of energy adsorbed in the system

(zeolite + water), which is a considerable increase from the 0.45 value for pure water (Cecal et al.,

2008). Smaller-sized cavities were found to give higher H

yields than larger ones, suggesting that

cavity size and not just surface area is important. Kinetic studies to determine the reactivity of the

electrons in the cavities suggest that the electrons are not hydrated (Nakazato and Masuda, 1986).

However, spectroscopic studies have identied hydrated electrons in the radiolysis of water in cer-

tain

types of zeolites (Liu et al., 1995).

recent series of papers have thoroughly examined the decomposition of water conned in a

variety of porous silica and mesoporous molecular sieves (Foley et al., 2005; Le Caer et al., 2005a,b;

Rotureau et al., 2005). The formation of H

in large cavities lled with water mainly occurs by the

same mechanisms as in bulk water radiolysis, that is, by reactions of H atoms and hydrated elec-

trons. The systems were examined as a function of water content within the cavities and also as a

function of cavity size. As water is depleted from the cavity, a void is formed in the center of the

cavity. The remaining water adsorbed on the walls is more stable due to the ice-like structure dis-

cussed in the previous section. H

yields were found to increase dramatically with decreasing water

content. Yields of H

are as high as 30 molecules/100 eV energy initially deposited in the water

phase (Rotureau et al., 2005). The production of H

from the adsorbed water appears to be much

more efcient than in bulk water, even bulk water conned to a cavity. Hot electrons or H atoms

may pass through the interface and produce H

by dissociative attachment reactions (Rotureau

etal., 2005). Excitons at the surface may also transfer energy to the conned water leading to

production by excited-state dissociation. A variety of mechanisms may occur, but no experi-

mental evidence exists on the relative importance of these reactions leading to H

production.

Interestingly, EPR and IR techniques show more defects in irradiated dry silica than when

Radiation Chemistry of Water with Ceramic Oxides 439

water is present (Le Caer et al., 2005a,b). Transfer of energy from the solid phase to the water is

preferred,

even to the point of depletion of the available water molecules.

normal bulk water radiolysis, OH radical scavengers such as bromide anion or alcohol are

added to prevent depletion of the H

yield. The addition of bromine anion, formate, or isopropanol

has no effect on the production of H

from water in silica cavities (Rotureau et al., 2005). Radicals

may be separated by the pores thereby decreasing combination reactions or the OH radicals may be

more reactive with something else in the cavity than with H

. OH radicals are produced in cavities

as evidence by their scavenging with coumarin (Foley et al., 2005). As with H

yields, OH radical

yields increase with decreasing water content of the cavities suggesting that adsorbed water is very

efcient in their production. In fact, the scavenger studies suggest that a signicant amount of OH

radicals are produced at the silica–water interface. Higher yields of OH radicals are observed with

increasing areas of the interface, inferring the transfer of energy from the silica to the interface is

responsible for at least part of the OH production. Some OH radicals may be scavenged by wall

sites, but a signicant fraction combines to form H

. The yield of H

has been determined using

scavenger techniques and found to agree well with H

yields (Le Caer et al., 2007). The results sug-

gest that the complementary reducing and oxidizing stable species of H

and H

are formed in

equal yields in the radiolysis of conned water. If so, then both oxidizing and reducing species must

transported through the silica–water interface, leading to the decomposition of water.

16.7 ConClusions

The radiolysis of water in association with interfaces is becoming one of the new challenges in asso-

ciation with nuclear technology. Understanding the radiolytic decomposition of water adsorbed on

or very near to solid surfaces can be even more demanding than bulk water because of the transport

of charge and energy to and through the water–solid interface. Many studies have shown that the

radiolysis of water is different at a solid interface than in the bulk liquid, although the mechanisms

are still not resolved. The nature of the energy carrier and the mode of transport in the solid are still

in question. Of particular importance is to identify the specic properties of the solid responsible

for the transport of energy and charge. The nature of the water at or near solids and the effect of the

interface

on its radiolytic decomposition still require investigation.

aCknowledgments

This contribution is NDRL-4810 from the Notre Dame Radiation Laboratory, which is supported by

the

Ofce of Basic Energy Sciences of the U.S. Department of Energy.

reFerenCes

Al-Abadleh, H. A. and Grassian, V. H. 2003. FT-IR study of water adsorption on aluminum oxide surfaces.

Langmuir

19: 341–347.

Aleksandrov,

A. B., Byakov, A. Y., Vall, A. I., Petrik, N. G., and Sedov, V. M. 1991. Radiolysis of adsorbed

substances

on oxide surfaces. Russ. J. Phys. Chem. 65: 847–849.

Alig,

R. C. and Bloom, S. 1975. Electron-hole-pair creation energies in semiconductors. Phys. Rev. Lett. 35:

1522–1525.

Allen, A.

O. 1961. The Radiation Chemistry of Water and Aqueous Solutions. New

York: Van

Nostrand.

Aoki,

M., Nakazato, C., and Masuda, T. 1988. Hydrogen formation from water adsorbed on zeolite during

gamma-irradiation.

Bull. Chem. Soc. Jpn. 61: 1899–1902.

Audebert,

P., Daguzan, P., Dos Santos, A., Gauthier, J. C., Geindre, J. P., Guizard, S., Hamoniaux, G., Krastev, K.,

Martin, P., Petite, G., and Antonetti, A. 1994. Space-time observation of an electron gas in SiO

. Phys.

Rev. Lett.

73: 1990–1993.

Behar,

D. and Rabani, J. 2006. Kinetics of hydrogen production upon reduction of aqueous TiO

nanoparticles

catalyzed by Pd

, Pt

, or Au

coating and an unusual hydrogen abstraction; steady state and pulse radiolysis

study.

J. Phys. Chem. B 110: 8750–8755.