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

Подождите немного. Документ загружается.

270 Charged Particle and Photon Interactions with Matter

Figure 11.4 shows the transient absorption spectrum of a

TMPA N(SO CF )

2 3 2

−

solution of KI

recorded just after excitation by a 230nm light pulse. The spectrum is similar to that obtained pre-

viously by the pulse radiolysis technique, indicating that an electron can be ejected from iodide into

TMPA N(SO CF )

2 3 2

−

by CTTS band excitation.

As mentioned above, solvated electrons have not been observed in nanosecond pulse radiolysis mea-

surements on ionic liquids based on the imidazolium cation, instead forming radical species by attach-

ment to an imidazolium cation (Behar etal., 2001; Marcinek etal., 2001), which could potentially lead to

permanent degradation (Shkrob etal., 2007; Shkrob and Wishart, 2009). For this reason, neat imidazo-

lium cation–based ionic liquids might not be suitable for the nuclear fuel cycle process application, but

they could be used in mixtures. Therefore, we investigated the reaction of solvated electrons with Bmim

cations in TMPA

N(SO CF )

2 3 2

−

in the hope that the rate constant of the reaction would give us a strategy

for designing ionic liquid systems that are stable under irradiation. The rate constant was determined to

be 5.3 × 10

−1

, indicating that the half-life of solvated electrons in neat Bmim

N(SO CF )

2 3 2

−

would

be 350ps. In reality, the excess electrons may be scavenged even before they become solvated—femto-

second experiments suggest that the electrons may be scavenged in less than 5ps in imidazolium iodide

ionic liquids (Chandrasekhar etal., 2008). More detailed studies are warranted.

11.4 solvation dynamiCs and reaCtions

F p

re-solvated

(

dry)

leCtrons

The sections above have discussed radiation- and photon-induced ionization of ionic liquids and the

reactivity of solvated electrons; however, from the radiation chemistry standpoint, one of the major

differences between ionic liquids and conventional solvents is the fact that dynamic timescales are hun-

dreds to thousands of times longer in ionic liquids (Giraud etal., 2003; Holbrey etal., 2004; Castriota

etal., 2005; Shirota and Castner, 2005; Shirota etal., 2005, 2007; Arzhantsev etal., 2007; Jin etal.,

2007). Moreover, the response of an ionic liquid to the introduction of an excess electron is subject to

static and dynamic heterogeneities that depend on its composition. Although the initial stages of solvent

response involve inertial motions that operate on very short timescales, even in ionic liquids, the slower

response components scale with the viscosity of the medium. Consequently, it takes much longer to sol-

vate an excess electron in an ionic liquid as compared to normal liquids, and the mobility and reactivity

of pre-solvated electron states become much larger determinants of radiolysis product yields, as a result.

400 600 800

230 nm exc.

2.1 mM

Kl in TMPA-TFSI

1000

Wavelength (nm)

Absorbance (10

–3

)

1200 1400 1600

Figure 11.4 Transient absorption spectrum of the solvated electron in

TMPA N(SO CF )

2 3 2

−

by electron

photodetachment.

(Reproduced from Katoh, R. etal., J. Phys. Chem. B, 111, 4770, 2007. With permission.)

Radiation Chemistry and Photochemistry of Ionic Liquids 271

An extreme example of the slow response of ionic liquids is depicted in Figure 11.5, where

the transient spectra of the excess electron in the highly viscous alcohol-functionalized ionic liq-

uid

SH8 (N(SO CF ) )

2 3 2 2

−

at different times following the electron pulse are shown (Wishart etal.,

2005). The molecular structures of the alcohol-functionalized ionic liquids in question are shown

in Figure 11.6. The spectrum at 2ns has a shoulder in the range of 1200nm characteristic of an

electron solvated in an alkane ionic liquid environment. Over the course of tens of nanoseconds,

the shoulder disappears and the remaining spectrum peaks at 650nm, characteristic of an electron

solvated by alcohol functionalities. The average electron solvation time constants in methanol, etha-

nol, and ethylene glycol are about 2–9 ps and that of viscous 1,2,6-trihydroxyhexane (2490 mPa s)

is ∼1 ns, while in the viscous ionic liquid alcohols

SH6 (N(S O CF ) )

2 3 2 2

−

and SH8 (N(SO C F ) )

2 3 2 2

−

(viscosities 6860 and 4570mPa s, respectively), the observed time constants for absorption decay

are 25ns (SH8

) to 40ns (SH6

). In the case of these ionic liquid

alcohols, dynamic hindrance of hydroxypropyl side chain reorienta-

tion imposed by the electrostatic constraints of the ionic lattice plays

the

greatest role in slowing the electron solvation process.

more typical example involves the popular ionic liquid N-butyl-

N-methylpyrrolidiniumbis(triuoromethanesulfonyl)imide(bmpyrr

N(SO CF )

2 3 2

−

), which has a relatively low viscosity (94mPa s at

20°C) and a wide electrochemical window. Picosecond pulse-probe

transient absorption spectroscopy was performed at Brookhaven

National Laboratory’s (BNL) Laser-Electron Accelerator Facility

(Wishart etal., 2004) to observe the temporal blueshift of the excess

electron spectrum associated with the solvation process in bmpyrr

N(SO CF )

2 3 2

−

. The spectral data, measured from 600 to 1600nm

with a time resolution of 15ps, extending to 10 ns, showed a strong

absorption around 1400 nm at short time intervals that shifted to the

equilibrated spectrum of the solvated electron peaking at 1100nm.

Factor analysis of the data showed a multiexponential or distributed

400 600 800 1000

2 ns

SH8 (NTf

)

15 ns

50 ns

200 ns

400 ns

800 ns

Wavelength (nm)

Absorbance×10

1200 1400 1600

Figure 11.5 Transient spectra of solvated electron in an alcohol-functionalized ionic liquid

SH8 (N(SO CF ) )

2 3 2 2

−

. (Reproduced from Wishart, J.F. et al., Radiat. Phys. Chem., 72, 99, 2005. With

permission.)

SAn: R = CH

SEn: R = CH

OCH

SHn: R = CH

Figure 11.6 Molecularstruc-

ture of alcohol-functionalized

ionic liquids. (Reproduced from

Wishart, J.F. et al., Radiat.

Phys. Chem., 72, 99, 2005. With

permission.)

272 Charged Particle and Photon Interactions with Matter

solvation response with an average solvation time of 260ps (Wishart etal., manuscript in prepara-

tion). In comparison, the average solvation time for the uorescent solvation probe counarin-153 is

350ps in the same ionic liquid (Funston etal., 2007). Electron solvation dynamics in ionic liquids

is often difcult to measure because of the need to exchange (ow) the sample to avoid the accu-

mulation of radiolytic damage. Solvatochromic dye solvation measurements, which are much more

extensively studied in ionic liquids (Arzhantsev etal., 2007; Funston etal., 2007; Jin etal., 2007),

are a useful proxy for estimating electron solvation times.

Since electron solvation processes in normal liquids occur on the time scale of picoseconds, excess

electrons in ionic liquids can persist hundreds to thousands of times longer in mobile, pre-solvated

states with reactivity proles that differ from fully solvated states. As already shown in Figure 11.3,

the solvated electron yield near time zero decreased with the concentration of added electron scav-

enger. In water and most other molecular solvents, such decreases in the “initial” yield cannot be

observed in nanosecond pulse radiolysis measurements. This decreased yield is due to scavenging

of the pre-solvated electron by the solute (Lam and Hunt, 1975; Jonah etal., 1977, 1989; Glezen et al.,

1992; Lewis and Jonah, 1986). For each solvent and solute, the efciency of pre-solvated electron

scavenging can be quantied by the constant C

dened as follows:

exp

−













(11.1)

where

is the yield of solvated electrons at a given scavenger concentration c

is the yield in the absence of scavenger

is the concentration where only 1/e (37%) of the electrons survive to be solvated

Examples of C

values and reaction rate constants of solvated electrons with scavengers

k( )

−

are

listed in Table 11.1. In

MeBu N N(SO CF )

2 3 2

−

, the values of C

for benzophenone, pyrene, and

phenanthrene are 0.062, 0.063, and 0.084M, respectively. The C

value for Bmim

is 0.057M in

TMPA N(SO CF )

2 3 2

−

, suggesting that the pre-solvated electron reacts with Bmim

at least as

efciently

as with the aromatics.

is reported that C2-alkylation of imidazolium cations is effective in extending the electro-

chemical redox windows of the imidazolium-based ionic liquids (Hayashi et al., 2005), suggest-

ing that C2-alkylation may reduce the reaction rate with solvated electrons. The solvated electron

table 11.1

rate

Constants and C

values for the reactions

of

olvated

and p

re-solvated

electrons

with

everal

cavengers

ionic liquid scavenger

k( )

−

−1

) C

(m)

MeBu

N(SO CF )

2 3 2

−

Benzophenone (1.6 ± 0.1) × 10

0.062

MeBu

N(SO CF

2 3

)

−

Pyrene (1.7 ± 0.1) × 10

0.063

MeBu

N(SO CF )

2 3 2

−

Phenanthrene (1.3 ± 0.1) × 10

0.084

MeBu

N(SO CF )

2 3 2

−

Indole (4.3 ± 0.3) × 10

0.22

TMPA

N(SO CF )

2 3 2

−

Bmim

5.3 × 10

0.057

TMPA

N(SO CF )

2 3 2

−

EDmim

3.8 × 10

0.081

TMPA

N(SO CF )

2 3 2

−

BDmim

3.0 × 10

0.13

TMPA

(N(SO CF )

2 3 2

−

HDmim

3.0 × 10

0.12

Radiation Chemistry and Photochemistry of Ionic Liquids 273

capture rate constants and C

values for N-ethyl-2,3-dimethylimidazolium (EDmim

), N-butyl-2,3-

dimethylimidazolium (BDmim

), and N-hexyl-2,3-dimethylimidazolium (HDmim

) are summarized

in Table 11.1 (Takahashi etal., 2008). For each of the C2-alkylated imidazolium cations, there is only

a slight reduction of the rate constant compared to Bmim

. This is quite reasonable because the rate

constants are considered to be diffusion limited. On the other hand, the C

values for BDmim

and

HDmim

are signicantly larger than that for non-alkylated imidazolium Bmim

. These results have

interesting implications for the reaction mechanisms of pre-solvated and solvated electron capture and

the nature of the electron–cation adduct, which may suggest a strategy for designing ionic liquids that

are more stable under irradiation.

11.5 photoexCitation oF solvated eleCtrons

There have been several studies for dielectrons in molten salts and solid salts, such as KCl (Lynch

and Robinson, 1968). A recent ab initio simulation for electrons in LiF (Zhang etal., 2008) also

suggested the presence of two types of dielectrons, namely, the singlet dielectron and the triplet

dielectron. The stabilization of two electrons in one trap site is considered to be due to not only

electrostatic interactions but also through hole–orbital coupling among solvent molecules (Zhang

et al., 2008). Because ionic liquids can be considered to have similar characteristics as molten

salts, we can postulate the presence of the dielectron in ionic liquids. To examine this possibility,

an experiment was performed in which the solvated electron was excited by light pulses at 532 and

1064 nm. If the dielectron exists, photoexcitation might produce two solvated electrons, which could

be detected as an increase in the absorption signal in the near-infrared wavelength region. The sol-

vated electrons were produced by electron photodetachment from iodide in an ionic liquid solution

by a 248nm KrF excimer laser pulse, as described in Section 11.3. After a variable delay, a 532nm

Nd-YAG

laser pulse irradiated the sample (Takahashi etal., 2009b).

Figure

11.7 shows a transient absorption signal observed at 1000nm. When a 532nm laser pulse

was used to irradiate the sample during the lifetime of the solvated electron, the transient absorp-

tion signal displayed an instrument-limited decrease, and the decrease of absorption did not recover

within the lifetime of the solvated electron. Therefore, a fraction of the solvated electron population

was permanently removed by a 532 nm excitation. Transient absorption signals were monitored at

several different wavelengths, and a similar permanent bleaching was observed at each wavelength.

0.3

0.2

0.1

0.0

6004002000–200

Time (ns)

1000 nm

Figure 11.7 Transient absorption signals observed in

TMPA N(SO CF )

2 3 2

−

at 1000nm with and without a

532nm pulse. (Reproduced from Takahashi, K. etal., Radiat. Phys. Chem., 78, 1129, 2009b. With permission.)

274 Charged Particle and Photon Interactions with Matter

The above results suggest that a portion of the solvated electrons react and disappear during the

532nm laser pulse. One possible explanation may be the reaction of the dry electron with its parent

iodine

atom via the following reaction scheme:

I I

− −

 → +

hν( )

248

(11.2)

e e

p s

− −

 →

(11.3)

e (e )*

( )

− −

 →

hν 532

(11.4)

(e )*

I I

− −

+  →

(11.5)

I I I

− −

+  →

(11.6)

I I I I

2 2 3

− − − −

+  → +

(11.7)

where

−

is the pre-solvated electron

−

is the solvated electron

−

indicates the electron after excitation with a 532nm light pulse

If permanent bleaching of the solvated electron takes place due to the above reaction scheme, it may

be possible to observe a decrease in

−

concentration, because the iodine atoms would disappear

on reacting with

( )e *

−

. Therefore, the transient absorption of

−

was measured with a wide time

window. The absorption spectrum of

−

in TMPA-NTf

is similar to that in water, with absorption

maxima located at 400 and 740 nm. The evolution of the absorption signal from

−

was monitored at

830nm. At this wavelength, both the

−

species and also the reactivity of solvated electrons can be

observed. The results are shown in Figure 11.8. In the early time regime (t < 500ns), the absorption

signal occurs mostly due to the solvated electron. After the decay of the solvated electron, a slow

buildup of

−

was seen. As shown in Figure 11.8, when the sample was irradiated by a 532nm laser

pulse approximately 200 ns after the production of the solvated electron, a permanent bleaching was

observed; however, in the microsecond regime, where the absorption is mainly due to

−

, there was

no signicant change in the absorption. It can be concluded that the iodine atom is not a reaction

partner of the excited electrons, since the yield of

−

at longer time periods remains unchanged.

Although at the present time we cannot identify the reaction mechanism(s) for the photo-induced

bleaching of solvated electrons with a 532nm light pulse, we can gain some insight from the previ-

ous work done on water and alcohols (Silva etal., 1998; Yokoyama etal., 1998; Kambhampati etal.,

2001). In these studies, pump-probe spectroscopy of the solvated electron in water, methanol, and

ethanol was studied with a 300fs time resolution. At low pump power, the observed dynamics were

assigned to s → p excitation and subsequent relaxation of the localized solvated electron. In contrast,

at high pump power, two-photon absorption produces mobile conduction band electrons, which are

subsequently trapped and relax at a resonant site far away from the initial equilibrated electron site.

Interestingly, in the two-photon excitation, they observed an ultrafast proton-transfer reaction from

the alcohol molecules, resulting in a permanent bleach. Although we cannot conclusively say that we

are observing the same process, we tentatively conclude that the excitation of the solvated electron

results in the generation of a reactive dry electron or pre-solvated electron in the ionic liquid.

Radiation Chemistry and Photochemistry of Ionic Liquids 275

11.6 radiCal reaCtions and kinetiC salt eFFeCts

11.6.1 reactionS between diiodide anion radicalS in ionic liQuidS

Ionic liquids are nding increasing uses in electrochemical devices, such as electric double-layer

capacitors, lithium batteries, fuel cells, and dye-sensitized solar cells (DSSCs). In DSSCs, a photo-

oxidized dye molecule must be reduced by a redox couple (Papageorgiou etal., 1996; Wang etal.,

2004). The

I /I

− −

redox couple has been used for this purpose and plays an important role in charge

(hole) transport in DSSCs. The diffusion of ions in ionic liquids is generally slow because of the high

viscosity of ionic liquids, and this fact is a disadvantage when an ionic liquid is used as an electro-

lyte. Recently, using an ultramicroelectrode technique, it has been proposed that in the

I /I

− −

redox

couple system, a Grotthuss-like mechanism involving the exchange of I

between

−

and I

−

is respon-

sible for efcient hole transport in ionic liquid electrolytes for DSSCs (Kawano and Watanabe, 2003,

2005). The exchange reaction would occur by the following process:

I I I I I I I

− − − − − −

+ → → +

3 3

 

(11.8)

To have this type of exchange take place, I

−

and

−

must be in close proximity. However, the close

approach of I

−

and

−

is hindered in molecular solvents because of Coulombic repulsion between the

two ions. On the other hand, the ions of an ionic liquid may weaken the electrostatic repulsion by

Coulombic shielding, allowing I

−

and

−

to approach each other without a large Coulombic energy

barrier. The effect of Coulombic shielding was examined through the reaction between two diiodide

(

−

) anion radicals. The

−

species can be formed by the 248nm laser photolysis of iodide through

the

following reactions:

I I

− −

+ → +hν e

(11.9)

0.16

0.14

0.12

0.10

0.08

0.06

Absorbance

0.04

0.02

0.00

–9

–8

–7

–6

Time (s)

–5

–4

–3

Figure 11.8 Time courses of the transient absorbance at 830nm after 248nm excitation of KI in

TMPA N(SO CF )

2 3 2

−

with (cross symbols) and without (dots) a subsequent 532nm pulse. The 532 nm pulse

was delayed by 200ns after the 248 nm pulse. (Reproduced from Takahashi, K. etal., Radiat. Phys. Chem.,

78,

1129, 2009b. With permission.)

276 Charged Particle and Photon Interactions with Matter

I I I+  →

− −

(11.10)

I I I I

2 2 3

− − − −

+  → +

(11.11)

In molecular solvents, the diffusion-controlled rate for an ionic reaction can be calculated by the

Debye–Smoluchowski

equation (Debye, 1942):

k R D f

diff ab ab

= 4π ε( ) (11.12)

r R

( )

exp /

ε =

( )

−

c ab

(11.13)

Z Z e

k T

a b

4πε ε

(11.14)

where

is the reaction distance of the closest approach

is the relative diffusion coefcient of reactants

f(ε)

is the so-called Debye correction factor

e is the charge on the ions

is Boltzmann’s constant

is the permittivity of free space

is the dielectric constant

T is the absolute temperature

the diffusion coefcient can be expressed by a simple hydrodynamic model, that is, by the

Stokes–Einstein equation, D

= k

T/(6πηR

), the diffusion-limited rate constant (Equation 11.12)

may

be written as follows:

RTf

diff

8000

( )ε

(11.15)

where

is the viscosity (in Pa s)

is the gas constant (8.3144

−1

mol

−1

)

The Debye correction factors, f(ε), for H

O, MeOH, and EtOH were calculated from the dielectric

constants of the solvents, and are listed in Table 11.2. The Debye correction factor for an aqueous

solution was calculated to be 0.52, whereas the values for methanol and ethanol were calculated

to be 0.18 and 0.091, respectively, indicating that the Coulombic repulsion between the

−

radical

anions in the alcohol solvents is greater than that in water. These calculated Debye correction fac-

tors predict that the reaction between

−

anion radicals in the alcohol solvents is much slower than

that in water. As indicated in Table 11.2, the reaction rate constants in MeOH and EtOH are four to

ve

times lower than in water.

The

decay rate constants, k

, for reaction 11.11 in the molecular solvents and the ionic liquids

are plotted in Figure 11.9 as a function of the reciprocal of viscosity. The solid line in Figure 11.9

isthe plot for the diffusion-limited rate constants calculated from Equation 11.15 with f(ε) = 1,

that is, the calculations for reaction between neutral molecules (denoted as k

). For the molecular

solvents, the k

ratios for H

O, MeOH, and EtOH are 0.37, 0.064, and 0.086, respectively, reect-

ing the importance of electrostatic repulsion between the diiodide anion radicals. In contrast, the

reaction rate constants for the ionic liquids are very close to the solid line. The k

ratios for the

Radiation Chemistry and Photochemistry of Ionic Liquids 277

table 11.2

solvent

roperties

and r

ate

Constants for the b

imolecular

eactionbetween

−

anion radicals

no. solvent η

f(Ɛ)

1 H

O 0.89 78.4 7.4 × 10

(2.8 ± 0.2) × 10

0.52 0.37

2 MeOH 0.55 32.7 1.2

× 10

(7.7 ± 0.5) × 10

0.18 0.064

3 EtOH 1.01 24.6 6.1

× 10

(5.3 ± 0.7) × 10

0.091 0.086

4 40%

Gly

3.1 68 2.1 × 10

(9.1 ± 0.8) × 10

0.47 0.43

5 60%

Gly

8.6 61 7.7 × 10

(3.0 ± 0.2) × 10

0.43 0.37

6 70%

Gly

17 56 3.8 × 10

(1.7 ± 0.3) × 10

0.39 0.44

7 80%

Gly

43 51 1.5 × 10

(6.8 ± 0.9) × 10

0.35 0.45

8 TMPA

TFSI

−f

69 9.6 × 10

(1.1 ± 0.2) × 10

1.15

9 P13

TFSI

−f

55 1.2 × 10

(1.3 ± 0.2) × 10

1.08

10 P14

TFSI

−f

69 9.6 × 10

(9.1 ± 0.5) × 10

0.95

11 PP13

TFSI

−f

135 4.9 × 10

(6.6 ± 0.4) × 10

1.35

12 DEME

TFSI

−f

63 1.0 × 10

(1.2 ± 0.5) × 10

1.14

DEME

−

300 2.2 × 10

(4.0 ± 0.7) × 10

1.82

Source: Data

from

Takahashi,

K. etal., J. Phys. Chem. B, 111, 4807, 2007.

Viscosity, in mPa s.

Calculated diffusion-limited rate constant with f(ε) = 1, in M

−1

Experimental rate constant for reaction 3, in M

−1

Debye correction factor.

Glycerol–H

O mixture.

TFSI

−

N(SO CF )

2 3 2

−

0.001 0.01

1/viscosity (mPa

–1

)

k (M

–1

)

0.1 1

Figure 11.9 Plots of rate constants for reaction between

−

anion radicals in molecular solvents and ionic

liquids as a function of the inverse of viscosity: (◻) H

O, (▴) alcohols, (⚬) glycerol–H

O mixtures, (⦁) ionic

liquids. The numbers correspond to the numbers in Table 11.2. The lled square is data for reaction between

−

anion radicals in an ionic liquid. The line plots the calculated diffusion-limited rate constants for reaction

between neutral molecules. (Reproduced from Takahashi, K. etal., J. Phys. Chem. B, 111, 4807, 2007. With

permission.)

278 Charged Particle and Photon Interactions with Matter

ionic liquids (except

PP13 N(SO CF )

2 3 2

−

and

DEME BF

−

) are almost unity (Table 11.2), indicating

that the Debye factors, f(ε), for the ionic liquids are close to unity and that the electrostatic repulsion

between

the diiodide anion radicals is unimportant.

The

coincidence of the experimental rate constants of

−

in the ionic liquids with those calcu-

lated based on the viscosity was, to a certain extent, unusual. In previous pulse radiolysis experi-

ments on the reaction of butylpyridinyl radical with duroquinone (Skrzypczak and Neta, 2003), the

experimental reaction rates deviated signicantly from calculated values based on viscosities with

the simple hydrodynamic assumption. In that work, the experimental rates were almost an order of

magnitude higher than the values estimated from the viscosities of the ionic liquids. Skrzypczak

and Neta (2003) suggested that this difference was due to the voids that exist in ionic liquids and the

possibility that reacting species diffuse through the movement of small groups of ions, whereas vis-

cosity is related to simultaneous movement of all ions. Furthermore, Maroncelli etal. have pointed

out the limitation of the continuum approximation of solvents and the importance of considering

the molecularity of solvents in solvation dynamics (Maroncelli and Fleming, 1987; Papazyan and

Maroncelli, 1995; Ito etal., 2004). The reason the simple hydrodynamic model can predict the pres-

ent results for the reaction between

−

anion radicals could be due to the relatively simple molecular

shape of the diiodide anion radicals. However, for the reaction of

−

PP13 N(SO CF )

2 3 2

−

and

DEME BF

−

, the agreement between the measured reaction rate and the calculated rate was not

good. These results could reect a limitation of the Stokes–Einstein approximation and indicate the

importance

of considering the molecularity of ionic liquids.

The

diffusion coefcients of

−

anion radicals in ionic liquids have been measured recently by

Kimura and Nishiyama by the transient grating method (Nishiyama etal., 2009). The diffusion

coefcients of

−

, calculated rate constants by the Smoluchowski and Debye–Smoluchowski treat-

ments, and the experimental rate constants for comparison are listed in Table 11.3. An important

point is that the calculated values by the Smoluchowski equation, k

(Smoluchowski)

, are not consistent

with the experimental rate constants in the molecular solvents, while the calculated values by the

Smoluchowski equation are nearly consistent with the experimental rate constants in ionic liquids.

These results suggest that the Coulombic screening effect efciently shields the electrostatic inter-

action

between the

−

anion radicals in ionic liquids.

Kimura and Nishiyama also compared the measured diffusion coefcients of

−

with the calcu-

lated values from the Stokes–Einstein relation (Nishiyama etal., 2009). The calculated diffusion

coefcients in

TMPA N(SO CF )

2 3 2

−

and

PP13 N(SO CF )

2 3 2

−

are 1.2 and 0.62 × 10

−1

, respec-

tively, whereas the measured values are 2.5 and 1.17 × 10

−1

, respectively. Therefore, the hydro-

dynamics approximation predicts a smaller value by a factor of 1/2 than the measured value in

ionic liquids. Hence, we can say again that the viscosity is not a measure for the diffusion rate of

molecules

in ionic liquids.

table 11.3

diffusion

Coefcients of

−

and Calculated rate Constants in

molecular

olvents

and i

onic

iquids

solvent

(smoluchowski)

(debye–smoluchowski)

(exp)

Methanol 120 550 100 77

Ethanol 79 380 35 53

TMPA

N(SO CF )

2 3 2

−

2.5 12 0.036 11

PP13

N(SO CF )

2 3 2

−

1.17 5.6 0.017 6.6

Source: Data

from Nishiyama,

etal., J. Phys. Chem. B, 113, 5188, 2009.

The unit is 10

–11

–1

The unit is 10

–1

Radiation Chemistry and Photochemistry of Ionic Liquids 279

11.6.2 kinetic Salt effectS in ionic liQuid/Methanol MixtureS

It is well known that the addition of salt increases or decreases the rates of ionic reactions,

depending on whether the reactant charges are like or unlike each other. These effects have been

treated with Debye–Hückel theory, although the applicable ion concentration range is limited to

about 0.01mol L

−1

. In contrast, the concentrations of ions in ionic liquids may be in the range of

2–5molL

−1

. Therefore, ionic reactions under such high ionic strength conditions are of interest. In

Section 11.6.1, the disproportionation of diiodide anion radicals

−

in ionic liquids was discussed.

In this section, we examine the kinetic salt effect on the reaction between diiodide anion radicals in

methanol using a wide concentration of ionic liquid salts. To extract the specic salt effect of ionic

liquids in organic solvents, we also used the inorganic salt lithium bis(triuoromethylsulfonyl)

imide (

Li N(SO CF )

2 3 2

−

) for comparison. Using ionic liquids, it is possible to examine the salt effect

under

very high mole fraction conditions (Takahashi etal., 2009a).

Figure

11.10 shows examples of transient absorption signals at 700nm with different concen-

trations of

TMPA N(SO CF )

2 3 2

−

in methanol. These decay signals correspond to the dispropor-

tion reaction of

−

. As the concentration of TMPA N(SO CF )

2 3 2

−

increases, the decay rate becomes

faster, indicating that the addition of the

TMPA N(SO CF )

2 3 2

−

salt accelerates the reaction rate

between

−

anion radicals. Because the

−

anions are expected to be surrounded by TMPA

cations,

the Coulombic repulsion between the

−

anion radicals is screened. In Figure 11.11, log(k/k

) was

plotted against the square root of the ionic strength, I, for three different salts, where k and k

are

rate constants with and without salt, respectively. At low ionic strength, the values of log(k/k

) are

proportional to

, indicating that the Debye–Hückel limiting law could be applicable. However,

as the ionic strength increases, the plots deviate from a straight line. As can be seen in Figure 11.11,

the effect of ionic strength on the rate constants depends on the kind of salts used. Because these

salts are composed of the common anion

NTf

−

, the difference in the effect of ionic strength can

be attributed to specic effects of the cations. At low ionic strength, the TMPA

cation increases

the reaction rate most effectively, while the Li cation is less effective. There are a few possibilities

to explain this result, such as screening effects, reduction of

−

anion mobility, and electrolyte

ion association. At high salt concentrations, it can be expected that the degree of dissociation of

Li N(SO CF )

2 3 2

−

is lower than for ionic liquids; hence, the kinetic salt effect may be less effec-

tive in

Li N(SO CF )

2 3 2

−

than in ionic liquids. The screening effect on the electrostatic repulsion

between

−

anions may depend on the size of the cations. A larger cation could effectively screen

40×10

–3

0 40

(a)

(b)

(c)

(d)

Time (μs)

Absorbance

120

Figure 11.10 Transient absorption signals at 700nm with different concentrations of

TMPA N(SO CF )

2 3 2

−

methanol. (Reproduced from Takahashi, K. etal., Chem. Lett., 38, 236, 2009a. With permission.)