Ionic Liquids - ACS Publications - American Chemical Society

---

Chapter 33

Photochemistry in Ionic Liquids 1

2

1

Charles M. Gordon , Andrew J. McLean , Mark J. Muldoon , and Ian R. Dunkin Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

1

1

Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow G1 1XL, Scotland, United Kingdom Department of Chemistry and Chemical Engineering, University of Paisley, Paisley PA1 2BE, Scotland, United Kingdom 2

This chapter describes our and other researchers' investigations into photochemistry in ionic liquids. The first part will describe our investigations into the nature of ionic liquids as solvents. The remainder of the chapter will describe our investigations into how ionic liquids can influence some fundamental reaction types.

Ionic liquids have attracted great interest in recent years as alternative reaction media to supplement or replace conventional organic solvents (1-3). This work has been driven in part by the idea that ionic liquids represent a "greener" alternative to such media, largely as a result of their negligible vapour pressures, resulting in greatly improved control over harmful emissions. Increasingly, however, the interest has arisen from the fact that ionic liquids can provide real improvements in performance over conventional systems. One area that has received little attention to date, however, is photochemistry in ionic liquids. This may be explained by the fact that many examples of synthetic photochemistry require relatively dilute reaction solutions,

428

© 2002 American Chemical Society

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

429 and hence quite large volumes of solvent. Nevertheless, photochemical reactions often result in different product distributions from those obtained in thermal reactions, and in some cases can provide access to alternative reaction products. Furthermore, the use of techniques such as laser flash photolysis has been instrumental in the determination of reaction mechanisms. Such information has provided important information regarding the nature of many reactive intermediates, and thus enhanced the understanding of many chemical processes. Equally neglected have been photophysical investigations involving ionic liquids. Many different probe molecules can provide important information about the manner in which a reaction medium will interact with different types of solute molecule. This chapter will describe efforts by us and other researchers to remedy this gap in the understanding of ionic liquids. The work described here will be largely confined to ionic liquids based on 1-alky 1-3-methylimidazolium cations ([Rmim] ). This is not due to any property inherent in such salts, but simply because [Rmim] salts have dominated the literature over the past ten years or so. As other types of ionic liquid become more commonly used, we hope to be able to expand the scope of our investigations. The main liquids employed are illustrated in Figure 1 along with the abbreviations used to identify particular cations or anions. +

+

R-t = C H , R = H: [bmimf R-i = C H R = CH : [bmmimf R-i = C H , R = H: [omimf 4

9

2

4

9l

2

8

17

3

2

Ri = C8H17, R = CH : [ommim3 2

R

2

3

+

X = PF BF CF3SO3 (TfO), (CF S0 ) N (Tf N) 6l

4l

3

2

2

2

FIGURE 1 : Structures of ionic liquids employed in these studies A major requirement was the preparation of "spectroscopic grade" ionic liquids, particularly with regard to their colour. Many ionic liquids have a slightly yellow colour i f not carefully prepared, indicating the presence of coloured impurities that could interfere with the processes under study. In order to prepare reliably colourless ionic liquids, a number of precautions were required which are outlined in the experimental section at the end of this chapter. The photochemical window of such liquids is quite wide, with effectively no absorbance at λ > 300 nm in "clean" liquids. A study of the literature reveals only three papers describing photochemistry in room temperature ionic liquids of this type. The first report, by Osteryoung and co-workers (4\ describes the photooxidation of iron(II) diimine complexes in acidic ionic liquids based on the 1-ethylpyridinium bromide/aluminium(III) chloride system. These reactions gave almost quantitative formation of iron(III) diimine complexes. It was suggested that the ethylpyridinium cation was acting

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

430 as the electron acceptor, and that the radical species thus formed then dimerised to form a highly coloured species displaying a new absorption band at 680 nm. More recently, Pagni and co-workers (5, 6) described the photoreactions of anthracene (An) and 9-methylanthracene (9-CH An) in 1-ethyl-3methylimidazolium (emim) and 1-butylpyridinium (bp)/ aluminium(III) chloride ionic liquids. Both acidic and basic systems were investigated, giving different results. Irradiation of An in basic emim-based ionic liquids gave exclusively the 4+4 dimer formed in conventional solvents like C H C N . Irradiation of 9-CH An on the other hand yielded the dimer as the major product, but also six minor products not observed in conventional solvents. These results suggested that 20% of the photoreaction in the ionic liquid occurred via pathways initiated by electron transfer from excited state 9-CH An to the solvent cation. A somewhat different product mixture was formed in the bp-based melts, an observation that was ascribed to the greater ease of reduction of [bp] than [emim] . It was also noted that reactions involving bimolecular processes occurred more slowly in the ionic liquids than in the less viscous C H C N . Overall, it appears that photochemical reactions can follow different pathways in ionic liquids and conventional solvents. The evidence suggests that the imidazolium and pyridinium cations can act as electron acceptors. 3

3

3

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

3

+

+

3

Solvent Properties O f Ionic L i q u i d s There have now been many reports of the use of ionic liquids for a wide range of chemical applications (7-3). In contrast however, there are remarkably little physical data reporting their solvent properties. The ground and excited states of many molecules are strongly influenced by interactions with their environment. For this reason, the photophysical behaviour of appropriate probe molecules can be used to investigate the manner in which solvents interact with different solutes. Such information is extremely important i f reactivity patterns are to be understood, and until recently there has been a dearth of such information for ionic liquids. One complicating factor is the chemical complexity of the liquids themselves. Clearly, a mixture of cations and anions potentially doubles the number of different interactions that must be considered. A wide range of photophysical probes now exist, but most only probe particular interactions with solvent molecules. The probes used must be chosen with care, therefore, before a "polarity" scale for ionic liquids may be prepared. Although a number of authors have reported the use of solvatochromic probe molecules to investigate the behaviour of molten salts based on alkylammonium cations, only a handful of such studies are reported for [R-mim] based salts. Bonhôte et al. showed that [emim][Tf N] displayed a polarity similar to that of ethanol based on the fluorescence spectrum of pyrene +

2

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

431 (7). They also reported that the apparent relative permittivity value for this solvent was below 10, based on the fluorescence spectrum of pyrenecarboxaldehyde. Finally, room temperature phosphorescence of 1bromonaphthalene was observed in this solvent. More recently, Carmichael and Seddon reported a study of the UV-visible absorption spectrum of the solvatochromic dye Nile Red in a range of [Rmim] based salts showing similar λ values to those obtained in short chain alcohols (8). Alteration of the length of the alkyl substituent had little effect on the position of λ . Finally, Brennecke co-workers have investigated the polarity of four [Rmim] and pyridinium based salts using the fluorescent probes 4-aminophthalimide and 4(N,N-dimethylamino)phthalimide (9). Their data indicated polarities lying between those of methanol and acetonitrile, depending on the cation present. +

π13Χ

ιη&χ

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

+

Reichardt's dye One of the most widely used polarity scales is the E scale, first reported by Reichardt based on the betaine dye 1 (commonly know as Reichardt's dye) (10, T

η).

This molecule displays a strong intramolecular charge-transfer absorption that is shifted by over 350 nm on moving from non-polar solvents such as diphenyl ether (λ = 810 nm) to polar solvents such as water ^ = 453 nm). The large negative solvatochromism arises from a difference in solvation of the electronic ground and excited states. Taft and Kamlet calculated that ca. 68% of the shift in transition energy could be assigned directly to a hydrogen bonding interaction with O" (12), thus indicating that the E scale is largely a measure of hydrogen bond donor strength for protic solvents. Compound 1 has been previously applied to investigations of a number of non-imidazolium based molten salts (13-16). We recorded the UV-visible spectrum of 1 in a range of imidazolium-based ionic liquids, investigating the effect of alteration of both the cation and anion (17). The values obtained are listed in Table I, along with some previously reported molten salt values, and some for conventional molecular solvents. The ηι&χ

m a x

T

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

432 values Ε (30) and E are parameters calculated from the value of λ ^ χ , representing the molar transition energy and a normalised polarity scale respectively. The limits of the £ scale are water ( £ = 1.0) and tetramethylsilane, T M S ( £ = 0.0) (7ft 77). Ύ

T

N

N

T

T

N

T

Table I. X , E (30) and Ej* values observed for 1, and X observed for 2 in a range of ionic liquids and other solvents at 298 Κ max

T

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

Solvent

max

*

C (l)7

'

Imidazolium-based ionic liquids [bmim][PF ] [omim][PF ] [bmim][Tf N] [omim][Tf N] [bmmim][Tf N] [ommim][Tf N] [bmim][TfO] [bmim][BF ] [ommim][BF ] Other molten salts (11) [C H NH ][N0 ] [(C H ) NH][N0 ] [(C H ) N][C H C0 ] "Conventional" solvents (11) Ethanol Propan-2-ol 1,2-Dichloroethane Acetone N,N-Dimethylformamide 6

6

2

2

2

2

4

4

2

5

3

4

9

6

13

3

3

3

4

6

5

2

N

nm

E (30)/ kcal mol"

546.5 558.0 555.5 559.0 588.0 599.5 547.0 545.0 592.0

52.3 51.2 51.5 51.1 48.6 47.7 52.3 52.5 48.3

0.667 0.633 0.642 0.629 0.552 0.525 0.667 0.673 0.543

516.5 516.5 546.0 548.5 547.5 549.5 601.5 -

464 504 651

61.6 56.7 43.9

0.954 0.803 0.407

-

-

551 591 692 678 662

51.9 48.4 41.3 42.2 43.2

0.654 0.546 0.327 0.355 0.386

585 591 500 569 602

x

E

values

T

T

1

^(2)/ nm

-

The data for the imidazolium salts fall into two distinct categories: the 1,3disubstituted cations give £ (30) values close to those obtained for short chain primary alcohols, while the 1,2,3-trisubstituted cations had noticeably lower E values, similar to those obtained for secondary alcohols. Thus, the methyl group at the 2-position on the imidazolium ring clearly results in weaker interactions with 1. These results suggest that the strongest influence on the value of £ ( 3 0 ) is the strength of the hydrogen bonding interaction between the imidazolium ring protons and the phenoxide oxygen on 1. Alteration of the anion seems to have little effect on the £ values for a particular cation, although [bmim][Tf N] T

T

T

T

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

2

433 displays a λ value approximately 10 nm higher than that observed for other salts containing the [bmim] cation. The reason for this anomalous value is not currently understood. A n increase in the length of one of the N-alkyl substituents results in only modest decrease in the E value. This concurs with observations made using the same probe molecule for primary alcohols of varying alkyl chain length, where increasing chain length was found to give a similar effect to that observed on addition of w-hexane to short chain alcohols (18). The non-imidazolium ionic liquids straddle our values; those containing acidic protons (i.e. ammonium salts) possess high E values, and those without (i.e., tetraalkylammonium salts) lower. When carrying out polarity measurements, two caveats must be highlighted. Although the [PF ]" and [Tf N]" based ionic liquids reported here are hydrophobic, this does not mean that they are entirely water free, unless specifically pre-dried. Even small quantities of water can have a significant effect on the position of X™ of 1. For example, in water saturated [bmim][PF ] ^max 526 nm, compared with a value of 546.5 nm in the dry liquid at 298 K . Thus, care must be taken to ensure that the liquids are dry when novel values are reported. The measurements reported in Table I were recorded using liquids containing 1 which were heated under vacuum at 70 °C until no further red shift in the C T band was observed. Despite these precautions, it is possible that traces of water remain, but we believe that these values are as accurate as may be obtained under the conditions generally employed for drying ionic liquids. Similarly, the presence of even small traces of other solvents of different polarities can also cause large changes in the position of A™ . A second potential problem can arise either in ionic liquids that are prepared from free acids, or in ones in which the formation of a protic acid might occur via decomposition. The presence of any significant quantity of free H ions can result in protonation of 1 at the oxygen atom, yielding a colourless species which is useless as a measure of relative polarity. ιγΐ3Χ

+

T

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

T

6

2

ax

6

=

ax

+

Basicity of ionic liquids The results listed above clearly suggest that the imidazolium-based ionic liquids can act as Η-bond donor solvents, and that the strength of the interaction is largely cation dependent, We have used the C u salt 2 to probe the basicity of a range of ionic liquids. 2+

[Χ]-

X = BPh or CI0 4

4

/ \

2

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

434 It has previously been shown that there is a correlation between the position of the lowest energy d-d band of this salt and the donor number of a solvent (7921). More recently, the same salt has been used to estimate the donor number of anions in solution (22). The solvatochromism arises from changes in the splitting of the d-orbitals of C u as the salt becomes five- or six-coordinated. The probe is used as a [BPh*]' salt as this anion is assumed to be entirely noncoordinating. The X values obtained for 2 in a range of ionic liquids are listed in Table I, along with those for some conventional solvents. The data clearly show that the basicity of the ionic liquids employed here is entirely anion dependent, the order being [PF ]* < [Tf N]' < [TfO]". Unfortunately, no consistent results could be obtained for [BF ]" salts. Based on previous data one would expect the value to lie between that of [PF ]' and [Tf N]", whereas different values in the range 550-600 nm were generally observed. We ascribe this to the presence of traces of CI" or Br' in the water miscible [BF ]" salts. The polarity data from 1 and 2 have recently been used to provide an explanation for the rate enhancements often observed when ionic liquids are used as solvents for biphasic catalysis (23). 2+

m a x

6

2

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

4

6

2

4

Laser flash photolysis investigations in ionic liquids Laser flash photolysis is an extremely versatile technique for investigating the behaviour of reactive intermediates. Photochemical processes may be studied by following the time dependence of either absorbance spectra or luminescence. We have applied laser flash photolysis with UV-visible detection to three different photochemical systems, notably benzophenone, [Ru(bpy) ] and Diels-Alder reactions of 0 . The first of these will be described in some detail, and results gained will be related to the other systems where appropriate. One of the most widely studied areas of photochemistry involves that of benzophenone and other organic ketones. One of the "classic" systems of this type is the benzophenone/naphthalene (or naphthalene derivative) couple, whose behaviour has been investigated in a wide range of different solvents (24). The first step of this process involves the photoexcitation of benzophenone (Bp), ultimately giving the ηπ* triplet excited state Bp*. In this lowest excited state, the C=0 bond is considerably weakened compared with the ground state, which means that B p * can also be considered as a biradical species, as indicated in Scheme 1 below (25). This species can then undergo a number of deactivation processes, the most important of which are phosphorescence, quenching via triplet energy transfer to an appropriate quencher (Q, e.g. naphthalene), or hydrogen atom abstraction from the solvent to form a ketyl radical (KR). The degree of quenching is dependent on the concentration of naphthalene present, while the contribution of hydrogen abstraction is solvent-dependent. 2+

3

1

2

3

3

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

435

hv

3Bp* Phosphorescence

3βρ*

Quenching

Bp + Q

3Bp* + Q

- KR+ R«

3Bp* + RH

H abstraction

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

Scheme 1

Hydrogen abstraction We have investigated both the Η-abstraction reaction and the quenching reaction with naphthalene in five different ionic liquids as well as a number of conventional organic solvents for comparison (26). Flash photolytic excitation ( λ = 355nm) of 5 m M solutions of Bp in the ionic liquids gave rise to a broad feature with a maximum around λ = 525 nm in all cases. This species decayed over several microseconds forming a second, long-lived species that exhibited a slight solvent dependence in its absorption maximum (X = 530-550 nm). A n example of the spectra obtained is shown in Figure 2. 6χ

max

0.0250.020 m 0.015<

0.0100.005 0.000460

480

500

520 540 560 Wavelength / nm

580

600 620

3

Figure 2. Transient spectra at 23.5 °C of B p * and K R in [omim][Tf N], λ = 355 nm. The spectra are recorded 0.2 μ$ (squares, solid line) and 16.0 μ$ (triangles, dotted line) after the laser pulse. The inset shows a typical kinetic trace recorded at λ = 530 nm. 2

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

ε χ

436 The initially formed, short-lived, component was quenched by naphthalene at a diffusion controlled rate, forming triplet excited state naphthalene (k = 425 nm). The decay rate of both components was strongly dependent on oxygen concentration. We are therefore very confident of the assignment of the short­ lived species as Bp*. K R is known to exhibit a X value around 540 nm (+/-10 nm), depending on the solvent (27). The long-lived component absorbing >15 μ$ after the laser pulse is therefore fully consistent with the presence of this species. In the ionic liquids, the yield of the long-lived species increased with increasing temperature, indicating an increasing quantum yield for the H abstraction process. In cyclohexane, 1-butanol and toluene, on the other hand, little change was observed over the temperature range investigated. In the absence of Np, the rate constant for the decay of B p * (k ) is given by the equation: max

3

max

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

3

obs

*obs

=

*H +

*τ

where k is the rate constant for the hydrogen abstraction process, and k the rate constant for deactivation via intersystem crossing. Based on the assumption that the absorbance measured directly after the laser pulse arises entirely from Bp*, while the absorbance after several μ$ arises from K R only, it is possible to calculate the values of & and k from the intensities of the transient absorption bands for B p * and K R . Using the values of £ recorded at a range of temperatures, the Arrhenius parameters listed in Table II were obtained for a range of ionic liquids, along with some conventional solvents in which hydrogen abstraction is also observed. The calculations assumed that the ratio of molar extinction coefficients for B p * and K R at 530 nm was 2.4, based on the assumption that all B p * was converted to K R at high temperatures in the [omim] salts. Clearly, this value may not hold for all ionic liquids, but the value of £ is unaffected by this ratio, only the value of the pre-exponential factor An, which must therefore be assumed to contain relatively large errors. H

T

3

H

T

3

H

3

3

+

H

a

3

Table II. Arrhenius parameters obtained for Η-abstraction by B p * from a range of ionic liquids and organic solvents E / kJmol Solvent AH/Sk /skobs / S [bmim][PF ] 3.6 χ 10 25.4 5.9 χ 10 2.0 χ 10 [omim][PF ] 4.8 χ 10 22.2 3.4 χ 10 2.7 χ ΙΟ [bmim][Tf N] 21.9 3.2 χ 10 9.8 χ ΙΟ 1.4 χ ΙΟ [bmmim][Tf N] 7.3 χ ΙΟ 1.7 χ ΙΟ 1.1 χ 10 27.3 [omim][Tf N] 1.3 χ ΙΟ 21.6 4.8 χ ΙΟ 7.3 χ ΙΟ Toluene 3.0 χ ΙΟ 14.5 9.8 χ ΙΟ 3.0 χ ΙΟ 4.9 χ ΙΟ 13.7 Cyclohexane 1.3 χ ΙΟ 4.9 χ ΙΟ 1 -Butanol 1.1 χ ΙΟ 13.3 1.1 χ ΙΟ 2.4 χ 10 At 23.5 °C l a

H

1

A

H

5

5

5

5

9

5

5

8

5

5

10

6

5

9

6

6

8

6

6

9

7

7

9

9

6

6

2

2

2

α

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

437 It is clear from Table II that: the activation energy for hydrogen abstraction is effectively identical in all ionic liquids investigated, and ca. 10 kJ mol" higher than that observed in the conventional solvents and, • the rate constant k for hydrogen abstraction is an order of magnitude lower in the ionic liquids than in the conventional solvents. The fact that the rate of abstraction increases with increasing alkyl chain length, along with an increase in the pre-exponential factor A , suggests that only the C - H bonds of the alkyl substituents are being broken. The reason for the lower degree of hydrogen abstraction observed in the ionic liquids is currently not clear. One obvious suggestion is that there is a specific interaction between the benzophenone carbonyl group and the hydrogen atoms on the imidazolium ring, resulting in a stabilisation of the ground state. This would be expected to result in shifts in the position of the ηπ* transition in the UV-visible spectrum. In fact, the UV-visible spectrum of Bp in all of the ionic liquids is effectively identical, with the X of the ηπ* absorption lying around 338 nm, between that observed for methanol and acetonitrile. Unfortunately the ππ* transition cannot be observed owing to the absorption of the solvent, but the position of the solvent edge suggests no significant shift compared with other solvents. Another possibility is that the C - H bonds in the ionic liquids are stronger than in the conventional organic solvents. The fact that the v(CH) regions in the infrared spectra of the ionic liquids are very similar to those observed for conventional alkyl groups would argue that this is also not the case, however. •

1

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

u

max

Thus we are left only with the possibility that interactions are occurring in the excited state that do not occur in the conventional solvents. It is possible that the dipole of Bp means that it is preferentially solvated by the dipolar imidazolium headgroup of the cation, and thus that the hydrogen abstraction reaction requires a large change in geometry. We currently have no spectroscopic or other information to support this hypothesis however. Nonetheless, it is clear that the ionic liquids are more resistant to radical abstraction reactions than many conventional media, and may suggest that such reactions might be carried out efficiently in these solvents. There is already evidence that ionic liquids represent excellent media in which to carry out radical polymerisation reactions (28).

Bimolecular processes in ionic liquids 3

The rate of quenching of the phosphorescence of B p * by Np (& ) was recorded over a range of temperatures and concentrations of Np, for each of the ionic liquids. A plot of k vs. [Np] allowed the calculation of a value for the bimolecular rate constant (k ) for the quenching reaction at each of the obs

obs

q

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

438 temperatures under study, which in turn allowed the construction of Arrhenius plots for each solvent. Using these plots it was possible to calculate the activation energy for the bimolecular triplet energy transfer reaction ( £ ) , and the entropy of activation AS*. The activation energies for viscous flow (Zs ) for all of the ionic liquids under study were also calculated using variable temperature viscometry. Over an extended temperature range, a plot of In η vs. 1/T is curved, but in the range 10-70 °C all of the liquids follow approximately Arrhenius behaviour (see Figure 3). R

a

v

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

a

FIGURE 3 Plot of In η (η measured in Pa.s) against 1/T for [bmim][PF ] (o), [omim][PF ] (x), [bmim][Tf N] (•), [omim][Tf N] (+), and [bmmim][Tf N] (O). Inset shows the same data over the temperature range 10-70 °C with linear fit lines used for calculation of Arrhenius parameters 6

6

2

2

2

A comparison of the values of £ and £ for five different ionic liquids is given in Figure 4. This clearly shows that, for a given ionic liquid, effectively identical values are obtained for the activation energies of the energy transfer process and for viscous flow. Thus, the triplet energy transfer process appears to be purely diffusion limited in all of the ionic liquids studied, and the values of k obtained should represent limiting values for the rate of reaction of two neutral organic molecules like Bp and Np. a

a

q

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

439

v

F I G U R E 4. Activation energies for viscous flow ( £ ) (front columns, light grey) compared with activation energies for quenching of B p * phosphorescence by Np ( £ ) in a range of ionic liquids and conventional solvents. a

3

R

a

We have used a similar approach to investigate the electron transfer reaction between photoexcited [Ru(bpy) ] (*R ) and the methylviologen dication ( M V ) in ionic liquids (Scheme 2), 2+

2+

3

2+

R2+ ^ *R2+ + MV2+ «.

hv hv'

*R2+ R3+ + MV+

Scheme 2 ]

and also the photo initiated Diels-Alder cyclisation reaction between 0 diphenylbenzofuran (DPBF, Scheme 3).

2

Ph DPBF

and

Ph

Scheme 3

The preliminary results of the former process in [bmim][PF ] have already been reported (29). The activation parameters for this process were obtained by 6

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

440 +

monitoring the fluorescence lifetime of *R as a function of temperature and [ M V ] . For the 0 / D P B F reaction, it was first necessary to generate 0 by irradiation of phenazine in an unsaturated ionic liquid. The rate of quenching of *0 was then monitored as a function of temperature and [DPBF]. The results gained to date are summarised in Table III. 2+

1

!

2

2

2

TABLE

III.

Solvent

A summary of kinetic parameters obtained for different Reaction

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

6

v

a

1

2

v

a

3

Bp*+Np *R + M V Ό + DPBF Bp*+Np *R + M V ' 0 + DPBF 2+

2 +

2

3

2+

2

E / kJ mol 37.7 41.8 35.5 23.9 26.2

q

3

[bmim][PF ] ( E = 37.6 kJ mol" ) [bmim][Tf N] ( E = 25.4 kJ mol"')

R

k (25°C)/ dm mol" s" 1.2 χ ΙΟ 2.9 χ ΙΟ 2.5 χ ΙΟ 3.8 χ ΙΟ 1.2 χ ΙΟ 5.1 χ ΙΟ

2 +

1

a

1

8

7

8

8

8

-1

-

8

AS*/ J K" mol 28 29 27 -9 -11 1

-1

-

Inspection of Table III shows that the activation energies for all three bimolecular processes are very close to those obtained for viscous flow in the two solvents, suggesting that the reactions are basically diffusion controlled in all cases. The largest discrepancy is observed for the reaction between * R and M V in [bmim][PF ], but even here E* is higher than £ by only 4.2 kJ mol" . This suggests that the ionic liquids provide a large degree of charge shielding when bringing together two charged species, with electrostatic repulsions contributing no more than 10 % of the overall activation energy. For a given solvent, the values of AS* are remarkably similar for all processes, even allowing for the large potential errors resulting from the extrapolation required to calculate this parameter. This suggests that solvent reorganisation plays a major role in formation of the transition state, as one might expect from a diffusion limited process. The largest differences between the different reactions are observed in the values of k , as one might expect. The electron transfer reaction between * R and M V is predictably the slowest, as it involves the movement of two large species. As these and other reactions are studied in more solvents, it should be possible to build a picture of how factors such as charge and solute size affect the way that ionic liquids solvate different types of molecule, and the effect of the solvation on the reaction dynamics. 2+

2 +

v

6

a

q

2+

2 +

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

1

441

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

Summary The area of photochemistry in ionic liquids remains very incompletely understood. Although the limited range of investigations using solvatochromic dyes are beginning to give a picture of the way in which these solvents interact with different solutes, the level of understanding remains incomplete. Further photophysical studies will hopefully provide more information. The laser flash photolysis experiments have shown that many bimolecular reactions become diffusion controlled in these solvents owing to their large viscosities, even when charge effects might be expected to have large effects. The bimolecular rate constants gained for the different types of reactions should therefore give an idea of the limiting rate constants for different types of chemical reaction. The observation that Η-abstraction by B p * is slower in the [Rmim] ionic liquids than in conventional solvents suggests that preparative reactions involving biradical intermediates may provide interesting results. It is clear that many more systems must be investigated, however, i f we are to fully understand the properties of these novel solvents. 3

+

Experimental details The ionic liquids were prepared largely following literature procedures, by quaternisation of 1-methylimidazole or 1,2-dimethylimidazole with the appropriate 1-chloro- or 1-bromoalkane, followed by anion exchange using H P F , Li(Tf N) or N H B F as appropriate (7, 30). A number of precautions were necessary, however, to ensure that the solvents were colourless and suitable for spectroscopic and photochemical studies. The 1-haloalkanes were first cleaned following a literature procedure (31). This involved washing with concentrated sulfuric acid until the washings were colourless, next with sodium hydrogen carbonate solution, and finally with deionised water to remove the remaining acid. They were then distilled from P 0 immediately prior to use. 1Methylimidazole and 1,2-dimethylimidazole were vacuum distilled from NaOH immediately prior to use. The quaternisation reaction was carried out either with neat reagents, or with a small amount of anhydrous ethyl acetate added. The reaction was heated at no more than 100 °C under N until the reaction had gone to completion (usually 24 hours for 1-bromoalkanes and 72 hours for 1chloroalkanes). The ethyl acetate layer was decanted off, and the salt was then dried under high vacuum at 70 °C for several hours. The water immiscible salts were then prepared simply by mixing aqueous solutions of the appropriate halide salt with a slight excess of H P F (care must be taken as this reaction is very exothermic and HF is generated) or L i T f N as appropriate. The ionic liquids separated as dense lower fractions. These were 6

2

4

4

2

5

2

6

2

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

442 then washed repeatedly with deionised water until the washings were neutral, and no trace of halide could be detected. The [BF ]' salts were prepared by mixing aqueous halide salt solutions with N H B F . The resulting [Rmim][BF ] was then separated by washing with CH C1 . Careful washing of the CH C1 solution with cold deionised water allowed removal of most remaining halide anions, as determined using A g N 0 solution. Unfortunately, the water wash generally also resulted in a decrease in the final yield of [Rmim][BF ] owing to the relatively high solubility of these salts in water. A l l ionic liquids were dried thoroughly by heating under high vacuum at 70 °C for several hours before use. In addition, the solutions used for the solvent probe measurements were dried again in this fashion after addition of the probe molecule. For the flash photolysis measurements, stock solutions of Bp, R and phenazine were prepared. Aliquots of the appropriate quenchers were then added to 3 ml samples of the stock solutions to prepare individual samples. The flash photolysis apparatus consisted of a N d - Y A G laser for excitation operating at 355 nm, and the probe beam was a pulsed Xe arc lamp. The sample was contained in the thermostatted fluorescence cell, and was degassed with dry N (or saturated with 0 in the case of the Diels-Alder reaction) prior to all measurements. The desired wavelength of fluorescence or absorbance signals were selected using a monochromator prior to reaching the detector. The data were then stored on an oscilloscope, and subsequently transferred to a computer for analysis. Viscometry measurements were carried out using a cone and plate viscometer, with samples again dried as described above. 4

4

2

4

4

2

2

2

3

4

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

2 +

2

2

References 1 2 3 4 5 6 7 8 9 10

Welton, T. Chem. Rev. 1999, 99, 2071. Holbrey, J.D.; Seddon, K.R. Clean Prod. Processes 1999, 1, 223. Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. Engl. 2000, 39, 3772. Chum, H.L.; Koran, D; Osteryoung, R.A. J. Am. Chem. Soc. 1978, 100, 310. Hondrogiannis, G.; Lee, C.W.; Pagni, R . M . ; Mamantov, G. J. Am. Chem. Soc. 1993, 115, 9828. Lee, C.; Winston, T.; Unni, A ; Pagni, R . M . ; Mamantov, G. 1996, 118, 4919. Bonhôte, P.; Dias, A.-P.; Papageorgiou, N.; Kalyanasundaram, K . ; Grätzel, M . Inorg. Chem. 1996, 35, 1168. Carmichael, A.J.; Seddon, K.R. J. Phys. Org, Chem. 2000, 13, 591. A k i , S.N.V.K.; Brennecke, J.F.; Samanta, A . J. Chem. Soc., Chem. Commun. 2001, 413. Reichardt, C. Chem. Soc. Rev. 1992, 147.

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.

443

Downloaded by UNIV OF BATH on October 28, 2014 | http://pubs.acs.org Publication Date: July 25, 2002 | doi: 10.1021/bk-2002-0818.ch033

11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Reichardt, C. Chem. Rev. 1994, 94, 2319. Taft, R.W.; Kamlet, M.J. J. Am. Chem. Soc. 1976, 98, 2886. Schroth, W.; Schädler, H.-D.; Andersch, J. Ζ Chem. 1989, 29, 129. Poole, S.K.; Shetty, P.H.; Poole, C.F. Anal. Chim. Acta, 1989 , 218, 241. Harrod, W.B.; Pienta, N.J. J. Phys. Org. Chem. 1990, 3, 534. Herfort, I.-M.; Schneider, H . Liebigs Ann. Chem. 1991, 27. Muldoon, M.J., Gordon C . M . , Dunkin, I.R. J. Chem. Soc., Perkin Trans. 2 2001, 433. Langhals, H . Nouv. J. Chim. 1982, 6, 265. Fukuda, Y.; Sone, K . Bull. Chem. Soc. Jpn. 1972, 45, 465. Soukup, R.W.; Sone, K . Bull. Chem. Soc. Jpn. 1987, 60, 2286. Persson, I. Pure Appl. Chem. 1986, 58, 1153. Linert, W.; Jameson, R.F.; Taha, A . J. Chem. Soc., Dalton Trans. 1993, 3181. Wasserscheid, P., Gordon, C . M . , Hilgers, C.; Muldoon, M.J.; Dunkin, I.R. Chem. Commun. 2001, 1186. See e.g. Shizuka, H . ; Yamaji, M. Bull. Chem. Soc. Japan 2000, 73, 267 and references therein. Suppan, P.J. J. Chem. Soc., Faraday Trans. 2 1986, 82, 2167. McLean, A.J., Gordon C . M . , Muldoon, M.J., Dunkin, I.R., J. Chem. Soc., Chem. Commun, submitted for publication. Land, E.J. Proc. Roy. Soc., 1968, 305, 457. Carmichael, A.J.; Haddleton, D . M . ; Bon, S.A.F.; Seddon, K . R . Chem. Commun. 2000, 1237. Gordon, C . M . ; McLean, A.J. J. Chem. Soc., Chem. Commun. 2000, 1395. Huddleston, J.G.; Willauer, H.D.; Swatloski, R.P.; Visser, A . E . ; Rogers, R.D. Chem. Commun. 1998, 1765. Amarego, W.L.F.; Perrin, D.D. The Purification of Laboratory Chemicals; Butterworth Heinemann Press: Oxford, 1996.

In Ionic Liquids; Rogers, R., et al.; ACS Symposium Series; American Chemical Society: Washington, DC, 2002.