Probing Microenvironment in Ionic Liquids by Time-Resolved EPR of

Sep 30, 2015 - For one of these microenvironments TR EPR data is quite similar to those obtained in common frozen organic solvents (toluene, glycerol,...
0 downloads 0 Views 2MB Size
Subscriber access provided by NEW YORK MED COLL

Article

Probing Microenvironment in Ionic Liquids by Time-Resolved EPR of Photoexcited Triplets Michael Yu. Ivanov, Sergey L. Veber, Sergey A. Prikhod'ko, Nicolay Yu Adonin, Elena G. Bagryanskaya, and Matvey V. Fedin J. Phys. Chem. B, Just Accepted Manuscript • Publication Date (Web): 30 Sep 2015 Downloaded from http://pubs.acs.org on October 2, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

The Journal of Physical Chemistry B is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Probing Microenvironment in Ionic Liquids by Time-Resolved EPR of Photoexcited Triplets M. Yu. Ivanov,a,b S. L. Veber,a,b S. A. Prikhod’ko,c N. Yu. Adonin,c E. G. Bagryanskaya,a,b,d M. V. Fedina,b,* a

International Tomography Center SB RAS, 630090, Novosibirsk, Russia

b

Novosibirsk State University, 630090, Novosibirsk, Russia

c

Boreskov Institute of Catalysis SB RAS, 630090, Novosibirsk, Russia

d

N. N. Vorozhtsov Novosibirsk Institute of Organic Chemistry SB RAS, 630090, Novosibirsk,

Russia

Abstract Unusual physicochemical properties of ionic liquids (ILs) open vistas for a variety of new applications. Herewith, we investigate influence of microviscosity and nanostructuring of ILs on spin dynamics of the dissolved photoexcited molecules. We use two most common ILs [Bmim]PF6 and [Bmim]BF4 (with its close analogue [C10mim]BF4) as solvents and photoexcited Zntetraphenylporphyrin (ZnTPP) as a probe. Time-Resolved Electron Paramagnetic Resonance (TR EPR) is employed to investigate spectra and kinetics of spin-polarized triplet ZnTPP in the temperature range 100-270 K. TR EPR data clearly indicate the presence of two microenvironments of ZnTPP in frozen ILs at 100-200 K, being manifested in different spectral shapes and different spin relaxation rates. For one of these microenvironments TR EPR data is quite similar to those obtained in common frozen organic solvents (toluene, glycerol, N-methyl-2-pyrrolidone). However, the second one favors the remarkably slow relaxation of spin polarization, being much longer than in the case of common solvents. Additional experiments using continuous wave EPR and stable nitroxide as a probe confirmed the formation of heterogeneities upon freezing of ILs and complemented TR EPR results. Thus, TR EPR of photoexcited triplets can be effectively used for probing heterogeneities and nanostructuring in frozen ILs. In addition, the increase of polarization lifetime in frozen ILs is an interesting finding that might allow investigation of short-lived intermediates inaccessible otherwise.

                                                             *

 [email protected] 

1    ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

I. Introduction Ionic liquids (ILs) exhibit a number of unusual and advanced properties making them perspective media for many chemical processes of fundamental and applied significance, especially in field of Green Chemistry.1-4 ILs have nearly no vapor pressure, low toxicity, high thermal stability and ability of solving a wide range of reagents. In addition, physical properties of ILs, such as their high polarity, extraordinarily high viscosity and effects of nanostructuring with a formation of nanomesoscale polar and non-polar domains4-12 assume a variety of new applications. In particular, diffusion/mobility-controlled processes in ILs may proceed noticeably different compared to common organic solvents, ultimately leading to a change of the reaction pathways and the products formed.13,14 Heterogeneities and nanostructuring of ILs have been of great interest for both experimental and theoretical studies.5-12 It was found that charged cationic head groups and anions aggregate in polar nano-domains, whereas hydrophobic alkyl tails of cations aggregate in non-polar nanodomains. As the length of the alkyl chain increases, the non-polar domains become larger. The existence of nm-scale spatial heterogeneities due to the segregation of non-polar moieties dispersed in a polar network was evidenced both theoretically and experimentally.8,10 A good approach to investigate heterogeneities in liquids and soft matter is Electron Paramagnetic Resonance (EPR) of spin probes,15 where the mobility of dissolved/incorporated nitroxide radical reflects its interaction with microenvironment nearby. A number of EPR studies in ILs using spin probes have been accomplished, both in neat ILs and in IL/water mixtures.16-34 In particular, it was shown that nanostructuring of ILs in water can be monitored via polaritydependent hyperfine splitting. However, in general, the shape of the nitroxide spectrum was found to be not very sensitive to the effects of nanostructuring, therefore developing other types of probes for heterogeneities in ILs would be a topical task. At the same time, interactions between dissolved probes and anions/cations of IL may themselves create heterogeneities and form “cages” around the probes.13 Such effects are interesting as well, since localization of reagents in relatively persistent solvent cages may significantly alter the reaction pathways and physical processes behind. One of numerous processes strongly dependent on nature and properties of the solvent is the formation and decay of electron spin polarization in short-lived paramagnetic intermediates. Magnetic interactions are being modulated by diffusional rotation of the molecule, therefore the character of this motion has impact on both the polarization pattern and its relaxation, and new types of spin phenomena may be anticipated in ILs. Time-resolved (TR) EPR is the method of choice for studying spin-polarized intermediates in solutions.35-40 So far, only one paper was 2    ACS Paragon Plus Environment

Page 3 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

published reporting TR EPR study in ILs.41 Using photoexcited triplet molecule of Zntetraphenylporphyrin (ZnTPP) authors clearly demonstrated that high viscosity of ILs suppresses electron spin relaxation caused by diffusional rotation of ZnTPP. As a result, TR EPR spectra may be obtained even at room temperature, which was previously unreachable in less viscous common organic solvents. However, temperature dependence of TR EPR spectra and kinetics was not obtained and analyzed, and the potential of photoexcited triplet molecule to serve as a probe for nanostructuring in ILs was also not investigated. In this work we report a detailed study of TR EPR on ZnTPP in wide temperature range (100-270 K) in three different ILs. Contrary to most previous EPR studies in ILs, we focus on temperatures lower than ambient to reach extremely high values of microscopic viscosity (ηmicro) sensed by ZnTPP and to investigate effects of nanostructuring. The analysis of variable-temperature TR EPR spectra and kinetics indicates a formation of heterogeneities in frozen ILs that can be probed by photoexcited ZnTPP. We supplement TR EPR results by more common approach using continuous wave (CW) EPR of nitroxide spin probes and compare the sensitivities of two methods for probing heterogeneities in ILs. Finally, we compare TR EPR data in ILs with those in common organic solvents and draw conclusions on microenvironment around ZnTPP dissolved in IL.

2. Experimental Ionic

liquids

1-Butyl-3-methylimidazolium

methylimidazolium

hexafluorophosphate

tetrafluoroborate ([Bmim]PF6),

([Bmim]BF4),

1-Butyl-3-

1-decyl-3-methylimidazolium

tetrafluoroborate ([C10mim]BF4), Zn-tetraphenylporphyrin (ZnTPP) and stable nitroxide radical 2,2,6,6-tetramethylpyperidine-1-oxyl (TEMPO) were purchased from Sigma Aldrich; standard organic solvents were purified by distillation. ZnTPP was used as received; the powder in abundance was mixed with the corresponding IL and exposed to ultrasonic bath in the closed Eppendorf tube. In this way the maximum amount of ZnTPP was dissolved, and the remaining particles were precipitated by centrifuging. The dissolved fraction was isolated and left in the argon atmosphere at 10-2 Torr pressure for 48 hours in order to remove water from IL. Finally, solution was placed into the quartz EPR tube, evacuated (10-2 Torr pressure) and sealed off. Samples with TEMPO were prepared in the same way with the only difference that no ultrasonic treatment and centrifuging were required, and the concentrations of ~1 mM were easily obtained. Note that drying ILs from water prior to experiments was the crucial step. In this way we ensured that the concentration of water in ILs did not exceed 200 ppm (m/m), as was determined by coulometric Karl Fischer titration with Mettler Toledo DL39 instrument and reagent Hydronal 3    ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 29

Coulomat AD. Note that such concentrations of water as 200-300 ppm are typical for the studies of physical properties of ILs (e.g. Refs.25,33), therefore the amounts of water in our samples were comparable with those found in literature. The presence of water at higher concentrations leads to pronounced changes in TR EPR data obtained, attributed to the high sensitivity of microscopic viscosity in ILs to the presence of water. The glass transition temperatures (Tg) of the dried ionic liquids were estimated by monitoring the deformation of IL sample upon continuously decreasing temperature (with the rate ~1 K/min). When the glass transition temperature was reached, deformation ability vanished, indicating transition of IL into a glassy state. EPR measurements were done using a home-made continuous wave / TR EPR setup based on X-band Bruker EMX spectrometer (9 GHz) equipped with N2-cooled temperature control system (T~100-300 K). In all variable-temperature experiments the sample was first shock-frozen in liquid nitrogen and then transferred into the probe and equilibrated with its temperature for several minutes. Nd:YaG laser LOTIS-TII with the excitation wavelength of 532 nm was used. All simulations of CW and TR EPR spectra were done using EasySpin,42 whereas TR EPR kinetics were simulated using the numerical solution of Bloch equations. As will be shown below, TR EPR kinetics in frozen ILs becomes very slow and decays on the timescale of tens of microseconds. In such case, a possibility of experimental artifacts must be carefully avoided. First of all, very low amplitudes of detecting mw field should be used to avoid mw damping of the kinetics. This was ensured for each temperature by repeating experiments at several values of mw power and then choosing the power level keeping the shape of TR EPR kinetics unperturbed. Concentration of oxygen in the sample is another crucial factor for the rate of the kinetics decay, and it could not be precisely controlled. As was mentioned above, all samples were kept in argon atmosphere for 48 hours and then sealed off in the EPR sample tube, thus most of the oxygen should have been eliminated; still we observed noticeable (~30%) variation of the absolute value for TR EPR kinetics decay from sample to sample. We therefore do not base our conclusions on the absolute rates of kinetics, whose precise control is beyond our experimental capabilities, but rather observe and interpret general trends. With respect to TR EPR spectra, this experimental complication does not arise, and spectral shapes can be reliably interpreted.

3. Results 3.1. TR EPR spectra Time-resolved (TR) EPR is less common technique compared to Continuous Wave (CW) EPR. Typically, it uses laser excitation to create radical intermediates and transient detection to follow 4    ACS Paragon Plus Environment

Page 5 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

their evolution in time. Modulation of magnetic field is not used, and thus TR EPR spectra and kinetics are directly detected. Depending on mechanism of electron spin polarization, the absorptive and/or emissive lines can be observed. ZnTPP is a well-known molecule that forms triplet state upon photoexcitation (532 nm) with the lifetime on the order of milliseconds depending on temperature and microenvironment.43-46 In particular, photoexcited ZnTPP was previously studied by TR EPR in frozen organic solvents and liquid crystals.44,47,48 The spectra are strongly spin-polarized, and the lifetime of spin polarization is significantly (2-3 orders of magnitude) smaller than the lifetime of triplet state itself, therefore the influence of triplet lifetime on the observed spin phenomena is usually neglected. In order to investigate possible nanostructuring effects of ILs on spin-polarized ZnTPP we have selected two most common ionic liquids [Bmim]BF4 and [Bmim]PF6. The third IL [C10mim]BF4 was selected for comparison with [Bmim]BF4 and for investigation of influence of the length of alkyl chain. The glass transition temperatures (Tg) of selected ILs are ~190-200 K.49 Since TR EPR is applicable to ZnTPP in ILs even at room temperatures,41 the studies at 100-270 K cover all possible states of these ILs. Time-resolved EPR spectra of ZnTPP in ILs obtained by us here are principally similar to the typical spin-polarized spectra of ZnTPP in standard organic solvents (Fig.1). The shapes of these spectra are determined by zero-field splitting (ZFS) characterized by rhombic D-tensor with known components (D=32.33 mT, E=10.13 mT,48 see simulated spectrum in Fig.1h). TR EPR spectra in ILs and standard solvents show the same Absorption/Emission (A/E) polarization pattern due to the triplet polarization mechanism leading to predominant population of Tz state.48 A remarkable advantage of ILs is that, contrary to standard organic solvents, TR EPR measurements can be done at room temperature, as was first shown in ref.41. This owes to the high viscosity of IL leading to a slower rotation of ZnTPP and hence to suppression of the primary relaxation mechanism – the modulation of ZFS by rotational diffusion. TR EPR spectra in ILs and standard solvents can only be compared when standard solvents are frozen (Fig.1). Despite general similarities, there is a distinct difference found in ILs, which is most evident for [Bmim]BF4. The shape of the TR EPR spectrum in ILs is clearly contributed by two components, as is manifested by the changes in the spectrum shape vs. time delay after the laser flash (τDAF). The TR EPR spectrum of one component is characterized by relatively sharp peaks of canonical Z-orientations (B~310 and 380 mT) and fast relaxation, therefore its contribution is visible only at short τDAF (Fig.1a-d). Such type of spectrum is similar to TR EPR spectra of ZnTPP observed in standard solvents. The second component found in ILs yields TR EPR spectra having 5    ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 29

more isotropic shapes, analogous to spectra in fluid ILs, and relaxing much slower (Fig.1a-d, long τDAF). At short time delays τDAF TR EPR spectra in ILs are contributed by both fractions, whereas at long time delays only the slow-relaxing fraction remains. Remarkably, such trends are not found for frozen standard organic solvents (toluene, glycerol, N-methyl-2-pyrrolidone (NMP)), as one observes comparing Fig.1a-d with Fig.1e-g. Although spectral shape slightly changes as a function of τDAF, no “isotropic-like” slow-relaxing fraction is found. Therefore, the two components found in ILs indicate the occurrence of structuring phenomena, which are specific for ILs and absent in standard solvents.

Fig. 1. Normalized TR EPR spectra of ZnTPP in ionic liquids (a-d) and standard organic solvents (e-g) measured at different time delays τDAF after the laser flash (different colors; integration window 1 μs for all traces, the beginning of integration indicated). The corresponding solvent and temperature are indicated in the plots. Black line in (h) shows the simulated spectrum of ZnTPP triplet (sim) using the ZFS parameters of Ref.48.

The manifestation of two fractions in [Bmim]BF4 is found already at 100 K (Fig.1a), and becomes even more evident at T~160-200 K (Fig.1b). TR EPR spectra of [Bmim]PF6 at 100 K do not show any signs of two contributions; however, at T~180-200 K (Fig.1c) the same manifestations as those of [Bmim]BF4 appear. Likewise, for the third IL [C10mim]BF4 manifestations of two fractions of ZnTPP are observable at T~180-200 K (Fig.1d). 3.2. TR EPR kinetics If two fractions of ZnTPP coexist in frozen ILs and spin relaxation in these two cases is different, one reasonably assumes that TR EPR kinetics decay should be biexponential. However, the situation is more complicated, because TR EPR kinetics shows biexponential decay even in common frozen organic solvents. Such biexponential behavior was previously observed for parent 6    ACS Paragon Plus Environment

Page 7 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

compound CuTPP and explained by either two types of coordination or equilibrium between dimers and monomers coexisting in solution (for both components the primary relaxation mechanisms are the modulation of ZFS and coupling with phonon).50 In case of ZnTPP similar trend is readily observed in frozen toluene, glycerol, N-methyl-2-pyrrolidone (NMP) (Fig.2a). Remarkably, for three standard solvents the TR EPR spectrum shape does not noticeably change depending on τDAF (Fig.1e-g) and, in agreement with that, TR EPR kinetics is closely the same across the spectrum. In other words, both fractions in common organic solvents have the same spectral characteristics, yet different relaxation times. Contrary to that, at certain temperatures TR EPR kinetics of ZnTPP in ILs noticeably depends on the spectral position. This is exemplified using [Bmim]BF4 at 100 and 200 K in Fig. 2b. Such behavior is completely consistent with the changes of the TR EPR spectra vs. τDAF shown in Fig.1a,b. At the edges of the spectrum (canonical Z-orientations) the kinetics decays faster at both 100 and 200 K, and closer to the center of the spectrum this dependence disappears. At 200 K the dependence of kinetics decay on spectral position is more pronounced compared to 100 K, in agreement with the spectral shapes shown in Fig.1a,b. Figure 2a compares TR EPR kinetics measured at 100 K in three ILs and in three common organic solvents. As was mentioned in section 3.1, TR EPR spectra of [Bmim]PF6 and [C10mim]BF4 negligibly depend on τDAF at 100 K, as well as the spectra in common organic solvents, thus TR EPR kinetics for each case is the same in any position of the spectrum. At the same time, for [Bmim]BF4 some dependence on τDAF is found (Fig.1a), but such dependence of TR EPR kinetics on spectral position is not too strong (Fig.2b), facilitating uniform comparison with other solvents at 100 K. One observes that kinetics in ILs are longer than those measured in common organic solvents, and this is especially evident in [Bmim]BF4 (Fig. 2a). All kinetics were simulated with an account of two contributions weighted with coefficients k for the slow-relaxing component and (1-k) for the fast-relaxing component. Numerical solution of Bloch equations was performed with relaxation times T1,i and T2,i being adjustable parameters (more details in Supporting Information). Given that T2,i>T1,1, we consider the TR EPR spectrum at τDAF=τ0>>T1,1 being the pure spectrum of slow-relaxing (with T1,2) fraction. Then the intensity of this slow-relaxing spectrum can be recalculated back to the short time delays (multiplied by exp(τ0/T1,2)) corresponding to the maximum of the TR EPR kinetics and used to decompose experimental spectrum into two fractions, as is exemplified in Fig.4. Assuming that the amplitude of polarization arising due to the triplet mechanism is the same for both fractions, we can compare their integral intensities (Islow and Ifast) and obtain contribution of slow-relaxing fraction keff=Islow/(Islow+Ifast). If two fractions have the same shapes of TR EPR spectrum, as is the case for standard organic solvents whose spectral shapes are independent of τDAF (Fig.1e-g), then clearly keff=k (the value defined above and shown at 100 K in Table 1). Interestingly, we have found that keff values in all three studied ILs are approximately constant at T=100-200 K (Fig. S1) and are close to the corresponding k values listed in Table 1: keff([Bmim]BF4)≈0.6, keff([Bmim]PF6)≈0.4, keff([C10mim]BF4)≈0.2-0.3. At T>200 K the values T1,2 and T1,1 drastically shorten and become close to each other, therefore TR EPR kinetics is well described by monoexponential function and decomposition into two fractions is not applicable any longer.

Fig. 4. Decomposition of TR EPR spectrum of ZnTPP in [Bmim]PF6 at 200 K into two contributions of slowrelaxing and fast-relaxing fractions.

3.4. Auxiliary CW EPR studies In addition to studies by TR EPR, it is reasonable to employ also the more common and simple method, CW EPR of nitroxide spin probes, to investigate microviscosity and possible structuring effects in selected ILs. The mobility of stable nitroxide radical provides information on local viscosity and environment, and this approach is widely used in various fields of science, including 10    ACS Paragon Plus Environment

Page 11 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

studies of ILs.16-34 However, in most cases previous CW EPR data were obtained at temperatures close to ambient and well above melting temperatures of corresponding ILs, and only a few works addressed the low temperature regions between glass transition (Tg) and melting (Tm) temperatures. To compare the capabilities of CW EPR of nitroxide probes and TR EPR of photoexcited triplets in the low-temperature region, we investigate three selected ILs using CW EPR in more detail. Considering viscosity of ILs, one distinguishes between the macroscopic and microscopic viscosities, which might have different values. Macroscopic viscosity corresponds to the laminar flow of IL and is routinely measured by viscosimetry, whereas microviscosity (ηmicro) is the local viscosity experienced by small molecules in the volume of IL. Stable nitroxide radicals are known to have EPR spectral shapes being very sensitive to the rate of rotational diffusion, and thus allowing one to determine the diffusion coefficient and, knowing the hydrodynamic radius of nitroxide, the viscosity of the liquid. We employ the most common spin probe TEMPO and X-band continuous wave EPR to investigate three selected ILs at T=80-300 K (Figs. 5 and S2). First, the rotational correlation time τc is obtained from the simulation of experimental spectra, that is connected with the rotational diffusion coefficient D via τ c = 1 (6 D ) . Then, using the Stokes-Einstein-Debye relation

τ c = 4πr 3η 3kT , where r is the hydrodynamic radius of the molecule (~0.34 nm for TEMPO 51-53), η=ηmicro is the viscosity and kT is the thermal energy, we obtain the corresponding ηmicro(T) dependences. In the temperature region ~300-200 K all studied ILs demonstrate most pronounced changes in EPR spectral shapes associated with drastic increase of the viscosity, as is exemplified for [Bmim]BF4 in Fig.5a (similar data for two other ILs are given in Fig. S2). However, surprisingly, at T~160-240 K the EPR spectra in all three ILs are best described by two contributions corresponding to (i) slow rotating and (ii) completely immobile TEMPO. This is most evident from the shape of the central (Ms=0) line that is split into two components: their positions exactly correspond to those of slow-tumbling and immobile TEMPO, and their intensities interconvert with temperature (illustrated for [Bmim]BF4 in Fig.5b). Note, that such kind of spectral interconversion (Fig.5b) is not typical for common organic solvents (see e.g. variable-temperature spectra of TEMPO in toluene and glycerol in Fig.S3). Denoting the slow-tumbling fraction as α and immobile one as (1α), we find that the α value decreases from 1 to 0 in the range ~240-160 K (Fig.6a). Below 160 K the spectrum is again well described by a single (immobile) contribution. Since the microscopic freezing begins below Tm~240-200 K,22 such behavior is indeed assignable to the coexistence of liquid and frozen fractions in ILs. The biphasic behavior between Tg and Tm was already observed 11    ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 29

by EPR of spin probes previously, although spectral decomposition was not performed.22,30 At the same time, it is also possible that, as was shown previously,27 nitroxides are localized in two types of environments corresponding to the alkyl chains of cations or to the anionic groups, and transition into the frozen phase (or softening/melting of the frozen phase) occurs at slightly different temperatures in the first and the second type. We emphasize that for each measurement the sample was first shock-frozen at 77 K, quickly transferred into the cryostat maintaining the desired temperature and thermalized to equilibrium. It is remarkable that the two fractions coexist in a rather wide temperature range ~80 K, as the mobility of TEMPO allows one to observe. Having two microenvironments of TEMPO at ~160-240 K, we still can estimate local microviscosity for the mobile fraction T>200 K (Fig. 6b). At T260 K (e.g. for [Bmim]BF4 and [Bmim]PF6 34). In several previous works the hydrodynamic radii of nitroxides in ILs calculated from the obtained τc values and macroscopic viscosities were found to be noticeably smaller than the corresponding geometrical radii (~0.1 nm) due to microviscosity effects.28 Here, we instead take the known hydrodynamic radius of TEMPO (~0.34 nm 51-53) and use it to convert τc into ηmicro in the broad temperature range (Fig.6b). Such a big difference between macroscopic and microscopic viscosities should be assigned to the boundary conditions of the probe, i.e. to the effects of microviscosity, possibly originating from the structural organization of ILs around TEMPO. Note that the values ηmicro(T) shown in Fig.6b (T>200 K) correspond to apparent microviscosity and can be directly used in Stokes-Einstein-Debye expression in order to estimate τc values for small molecules. In particular, this might be useful for detailed study of spin relaxation in photoexcited triplets in ILs.

12    ACS Paragon Plus Environment

Page 13 of 29

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Fig. 5. (a) Temperature-dependent EPR spectra of TEMPO in [Bmim]BF4 (blue) with the corresponding simulations (red). The values of temperatures are indicated on the plot; the complete sets of parameters used in simulations are given in Supporting Information. (b) Magnified region of spectra where interconversion of two phases with temperature is best visible (characteristic peak positions are marked with dashed lines). All spectra are normalized to the maxima.

Thus, CW EPR study of TEMPO mobility in [Bmim]BF4, [Bmim]PF6 and [C10mim]BF4 yielded the corresponding values of microviscosity at T=200-300 K and revealed heterogeneity (coexistence of frozen/liquid phase) at T=160-240 K. In addition, Fig. 6a locates the glass transition (Tg) and melting (Tm) temperatures of the studied samples. The literature data on Tg and Tm covers a broad range around 200 K,49 therefore visualization of intermediate region between Tg and Tm provided for our samples by Fig.6a is advantageous for the following discussion.

13    ACS Paragon Plus Environment

The Journal of Physical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 29

Fig. 6. (a) Temperature dependence of the mobile fraction (α) of TEMPO probes in three studied ILs. (b) Obtained microviscosity ηmicro vs. temperature for three studied ILs. Accuracy of each value is ~30%.

4. Discussion 4.1. Comparison of TR and CW EPR results

We have implemented TR EPR of photoexcited ZnTPP to probe local microenvironment in three representative ILs. As an additional independent approach, we also applied CW EPR using TEMPO nitroxide. Most often ILs are investigated near room temperatures, as this is directly relevant to synthetic procedures of green chemistry. Some effects of nanostructuring have been observed at such conditions previously by several techniques.5-12 At the same time, it would also be interesting to investigate similar effects in ILs at lower temperatures. Possibilities of such effects of nanostructuring in supercooled ILs were evidenced previously.8 Both TR and CW EPR have demonstrated the presence of heterogeneities in studied ILs at T