Multiprobe Spectroscopic Investigation of Molecular-level Behavior within Aqueous 1-Butyl-3-methylimidazolium Tetrafluoroborate Abhra Sarkar, † Maroof Ali, † Gary A. Baker,* ,‡ Sergey Y. Tetin, § Qiaoqiao Ruan, § and Siddharth Pandey* ,† Department of Chemistry, Indian Institute of Technology Delhi, Hauz Khas, New Delhi - 110016, India, Chemical Sciences DiVision, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA, and Department of Biotechnology, Core Research and DeVelopment, Abbott Diagnostics DiVision, Abbott Laboratories, Abbott Park, Illinois 60064, USA ReceiVed: NoVember 7, 2008; ReVised Manuscript ReceiVed: December 23, 2008 In this work, an array of molecular-level solvent featuressincluding solute-solvent/solvent-solvent interactions, dipolarity, heterogeneity, dynamics, probe accessibility, and diffusionswere investigated across the entire composition of ambient mixtures containing the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate, [bmim][BF 4 ], and pH 7.0 phosphate buffer, based on results assembled for nine different molecular probes utilized in a range of spectroscopic modes. These studies uncovered interesting and unusual solvatochromic probe behavior within this benchmark mixture. Solvatochromic absorbance probessa water- soluble betaine dye (betaine dye 33), N,N-diethyl-4-nitroaniline, and 4-nitroanilineswere employed to determine E T (a blend of dipolarity/polarizability and hydrogen bond donor contributions) and the Kamlet-Taft indices π* (dipolarity/polarizability), R (hydrogen bond donor acidity), and (hydrogen bond acceptor basicity) characterizing the [bmim][BF 4 ] + phosphate buffer system. These parameters each showed a marked deviation from ideality, suggesting selective solvation of the individual probe solutes by [bmim][BF 4 ]. Similar conclusions were derived from the responses of the fluorescent polarity-sensitive probes pyrene and pyrene-1-car- boxaldehyde. Importantly, the fluorescent microfluidity probe 1,3-bis(1-pyrenyl)propane senses a microviscosity within the mixture that significantly exceeds expectations derived from simple interpolation of the behavior in the neat solvents. On the basis of results from this probe, a correlation between microviscosity and bulk viscosity was established; pronounced solvent-solvent hydrogen-bonding interactions were implicit in this behavior. The greatest deviation from ideal additive behavior for the probes studied herein was consistently observed to occur in the buffer-rich regime. Nitromethane-based fluorescence quenching of pyrene within the [bmim][BF 4 ] + phosphate buffer system showed unusual compliance with a “sphere-of-action” quenching model, a further manifestation of the microheterogeneity of the system. Fluorescence correlation spectroscopic results for both small (BODIPY FL) and macromolecular (Texas Red-10 kDa dextran conjugate) diffusional probes provide additional evidence in support of microphase segregation inherent to aqueous [bmim][BF 4 ]. Introduction The term “ionic liquid” (IL) is currently applied to substances composed exclusively of ions that remain in the liquid state below the boiling point of water, preferably down to room temperature. The popularity of ILs is on the rise, with a number of properties making them appealing as electrolytes and solvents for a multitude of applications: low or nonexistent freezing points; negligible vapor pressures; wide tunability of many of their properties, including polarity/hydrophobicity, solvent miscibility, thermal stability, and redox characteristics. 1 How- ever, key challenges, limitations, and misconceptions still remain. 1,2 A number of research groups have clearly recognized the importance of fundamental knowledge in addressing the challenges offered by this emerging field. It has now become evident that numerous phenomena are not yet fully understood and that a deeper, molecular-level understanding of ILs and their mixtures is required in order to fully implement these fascinating fluids in a wealth of areas, and to foster additional opportunities for alternative approaches. One of the fundamental obstacles we face is the rather limited solubility observed for a number of solutes in usual ILs. Accordingly, modifying the physicochemical properties of ILs is a hotbed of current research. In one approach, researchers have begun to focus on the implementation of ILs mixed with molecular solvents, 3 an important subclass of which is the aqueous IL system. 3a-e In particular, due to the propensity for strong intermolecular hydrogen-bonding interactions between water and ILs, 4 the addition of water has the potential to significantly alter IL physicochemical properties in a straight- forward fashion. In this context, although tracking the structural features of the solution 5 as well as measurement of the bulk physicochemical properties 6 of aqueous IL systems are certainly important, a molecular-level view of local solvation surrounding spectroscopic probes may be particularly telling by furnishing otherwise inaccessible information on local solute-solvent and solvent-solvent interactions and the underlying dynamics and microheterogeneity present within solvent mixtures. This in- formation is currently lacking, despite its plain relevance to * To whom correspondence should be addressed. E-mail: (S.P) [email protected]; (G.A.B.) [email protected]; phone: +91- 11-26596503; fax: +91-11-26581102. † Indian Institute of Technology Delhi. ‡ Oak Ridge National Laboratory. § Abbott Laboratories. J. Phys. Chem. B 2009, 113, 3088–3098 3088 10.1021/jp8098297 CCC: $40.75 2009 American Chemical Society Published on Web 02/13/2009
Transcript
Multiprobe Spectroscopic Investigation of Molecular-level Behavior within Aqueous1-Butyl-3-methylimidazolium Tetrafluoroborate
Abhra Sarkar,† Maroof Ali,† Gary A. Baker,*,‡ Sergey Y. Tetin,§ Qiaoqiao Ruan,§ andSiddharth Pandey*,†
Department of Chemistry, Indian Institute of Technology Delhi, Hauz Khas, New Delhi - 110016, India,Chemical Sciences DiVision, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA, andDepartment of Biotechnology, Core Research and DeVelopment, Abbott Diagnostics DiVision,Abbott Laboratories, Abbott Park, Illinois 60064, USA
ReceiVed: NoVember 7, 2008; ReVised Manuscript ReceiVed: December 23, 2008
In this work, an array of molecular-level solvent featuressincluding solute-solvent/solvent-solventinteractions, dipolarity, heterogeneity, dynamics, probe accessibility, and diffusionswere investigated acrossthe entire composition of ambient mixtures containing the ionic liquid 1-butyl-3-methylimidazoliumtetrafluoroborate, [bmim][BF4], and pH 7.0 phosphate buffer, based on results assembled for nine differentmolecular probes utilized in a range of spectroscopic modes. These studies uncovered interesting and unusualsolvatochromic probe behavior within this benchmark mixture. Solvatochromic absorbance probessa water-soluble betaine dye (betaine dye 33), N,N-diethyl-4-nitroaniline, and 4-nitroanilineswere employed to determineET (a blend of dipolarity/polarizability and hydrogen bond donor contributions) and the Kamlet-Taft indicesπ* (dipolarity/polarizability), R (hydrogen bond donor acidity), and (hydrogen bond acceptor basicity)characterizing the [bmim][BF4] + phosphate buffer system. These parameters each showed a marked deviationfrom ideality, suggesting selective solvation of the individual probe solutes by [bmim][BF4]. Similar conclusionswere derived from the responses of the fluorescent polarity-sensitive probes pyrene and pyrene-1-car-boxaldehyde. Importantly, the fluorescent microfluidity probe 1,3-bis(1-pyrenyl)propane senses a microviscositywithin the mixture that significantly exceeds expectations derived from simple interpolation of the behaviorin the neat solvents. On the basis of results from this probe, a correlation between microviscosity and bulkviscosity was established; pronounced solvent-solvent hydrogen-bonding interactions were implicit in thisbehavior. The greatest deviation from ideal additive behavior for the probes studied herein was consistentlyobserved to occur in the buffer-rich regime. Nitromethane-based fluorescence quenching of pyrene withinthe [bmim][BF4] + phosphate buffer system showed unusual compliance with a “sphere-of-action” quenchingmodel, a further manifestation of the microheterogeneity of the system. Fluorescence correlation spectroscopicresults for both small (BODIPY FL) and macromolecular (Texas Red-10 kDa dextran conjugate) diffusionalprobes provide additional evidence in support of microphase segregation inherent to aqueous [bmim][BF4].
Introduction
The term “ionic liquid” (IL) is currently applied to substancescomposed exclusively of ions that remain in the liquid statebelow the boiling point of water, preferably down to roomtemperature. The popularity of ILs is on the rise, with a numberof properties making them appealing as electrolytes and solventsfor a multitude of applications: low or nonexistent freezingpoints; negligible vapor pressures; wide tunability of many oftheir properties, including polarity/hydrophobicity, solventmiscibility, thermal stability, and redox characteristics.1 How-ever, key challenges, limitations, and misconceptions stillremain.1,2 A number of research groups have clearly recognizedthe importance of fundamental knowledge in addressing thechallenges offered by this emerging field. It has now becomeevident that numerous phenomena are not yet fully understoodand that a deeper, molecular-level understanding of ILs and their
mixtures is required in order to fully implement these fascinatingfluids in a wealth of areas, and to foster additional opportunitiesfor alternative approaches.
One of the fundamental obstacles we face is the rather limitedsolubility observed for a number of solutes in usual ILs.Accordingly, modifying the physicochemical properties of ILsis a hotbed of current research. In one approach, researchershave begun to focus on the implementation of ILs mixed withmolecular solvents,3 an important subclass of which is theaqueous IL system.3a-e In particular, due to the propensity forstrong intermolecular hydrogen-bonding interactions betweenwater and ILs,4 the addition of water has the potential tosignificantly alter IL physicochemical properties in a straight-forward fashion. In this context, although tracking the structuralfeatures of the solution5 as well as measurement of the bulkphysicochemical properties6 of aqueous IL systems are certainlyimportant, a molecular-level view of local solvation surroundingspectroscopic probes may be particularly telling by furnishingotherwise inaccessible information on local solute-solvent andsolvent-solvent interactions and the underlying dynamics andmicroheterogeneity present within solvent mixtures. This in-formation is currently lacking, despite its plain relevance to
* To whom correspondence should be addressed. E-mail: (S.P)[email protected]; (G.A.B.) [email protected]; phone: +91-11-26596503; fax: +91-11-26581102.
† Indian Institute of Technology Delhi.‡ Oak Ridge National Laboratory.§ Abbott Laboratories.
J. Phys. Chem. B 2009, 113, 3088–30983088
10.1021/jp8098297 CCC: $40.75 2009 American Chemical SocietyPublished on Web 02/13/2009
efficiently harnessing IL-solvent mixtures in numerous ways.Within this framework, our groups have previously publishedpreliminary studies reporting on the influence of water over thebehavior of select spectroscopic probes dissolved within arche-type “hydrophobic” (1-butyl-3-methylimidazolium hexafluoro-phosphate, [bmim][PF6])3a,g,4k,7 and “hydrophilic” (1-butyl-3-methylimidazolium tetrafluoroborate, [bmim][BF4])3c ILs. Weshould point out that the miscibility of water in [bmim][PF6] islimited to roughly 2 wt % under ambient conditions (in theabsence of additional cosolvent),3e imposing a distinct restrictionon the ability to modulate the properties of [bmim][PF6] viahumidification. The incomplete solubilities of other nonaqueouscosolvents in ILs can introduce similar restrictions, of course.Prior investigation of solvation within aqueous [bmim][BF4] wasconfined to electronic absorption probes yielding basic informa-tion on dipolarity/polarizability, hydrogen-bond donor (HBD)acidity, and hydrogen-bond acceptor (HBA) basicity.3c Althoughoffering important initial insight on preferential solvation andstatic solvation, these studies revealed nothing regarding thedynamics, microviscosity, and diffusion within aqueous[bmim][BF4], which is information that is readily available froma number of fluorescent probe studies. Moreover, this earlierwork did not fully account for possible anion hydrolysis, knownto be a potential issue for unbuffered [bmim][BF4].7b,8
To obtain a more detailed and definitive understanding ofthe solute-solvent and solvent–solvent interactions occurringwithin the aqueous [bmim][BF4] system, we have utilized anintegrated approach in this study, collating information gatheredfrom nine different molecular probes employed in a host ofspectroscopic approaches, including excited-state intensity decaykinetics measurements, fluorescence correlation spectroscopy,and molecular quenching studies. In order to obviate potentialbias arising from the hydrolytic instability of [bmim][BF4],phosphate buffer (pH 7.0, 10 mM) was used in lieu of waterfor mixture preparation throughout. Results from electronicabsorption probes reporting on dipolarity/polarizability, HBDacidity, and HBA basicity were utilized in concert withfluorescent dipolarity, microfluidity, and lifetime probes toobtain a molecular-level understanding of solute-solvent andsolvent–solvent interactions arising within [bmim][BF4] +aqueous buffer mixtures. Additionally, luminescence quenchinganalysis using nitromethane as a collisional quencher andfluorescence correlation spectroscopic studies of disparatelysized tracers were used to collectively assess solute diffusionalbehavior within the aqueous [bmim][BF4] system. Taken in sum,our multiprobe approach provides the most definitive andintegrated view of solvent interactions currently available forany aqueous IL mixture, suggesting a promising path forwardfor exploring IL mixtures with other molecular solvents ofinterest.
Experimental Section
Reagents and Supplies. 2,6-Diphenyl-4-(2,4,6-triphenyl-N-pyridino)phenolate (Reichardt’s dye 30), 2,6-dichloro-4-(2,4,6-triphenyl-N-pyridino)phenolate (Reichardt’s dye 33), 4-nitroa-niline, and N,N-diethyl-4-nitroaniline were purchased in thehighest available purity from Aldrich Chemical Co., Fluka,Spectrochem Co. Ltd., and Frinton Laboratories, respectively.Ultrapure grade [bmim][BF4], containing halide and waterimpurities at levels below 10 ppm, was obtained from Merckand stored under dry argon. Bidistilled deionized water with>18.0 MΩ cm resistivity was obtained from a Millipore, Milli-QAcademic water purification system. The following fluorophoreswere used as received from the specified vendors: pyrene,
1-pyrenecarboxaldehyde, and Rhodamine 110 were purchasedfrom Sigma-Aldrich; 1,3-bis(1-pyrenyl)propane (BPP; B-311),4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-pro-pionic acid (BODIPY FL; D-2183), and Texas Red-labeled10 000 MW dextran (D-1828) were from Invitrogen. Ethanol(99.9%) was obtained from SD Fine-Chem Ltd., sodiumdihydrogen orthophosphate and disodium hydrogen orthophos-phate were from Qualigens, and nitromethane was an Acrosproduct.
Methods. All absorbance and fluorescence probe stocksolutions were prepared in absolute ethanol and promptly storedunder refrigeration in sealed amber glass vials. For samplepreparation, desired volumes of an ethanol probe stock weretransferred to clean quartz cells, followed by ethanol removalusing a controlled flow of high-purity dry, filtered nitrogen gas.[bmim][BF4] and 10 mM pH 7.0 phosphate buffer (PB) solutionsprepared by mass using a Mettler Toledo AB104-S balance(precision ) (0.0001 g) were then added to each cuvette toachieve the desired probe concentration. To ensure completedissolution of the probe molecule within the [bmim][BF4] +PB mixtures, the solutions were thoroughly mixed by acombination of magnetic stirring, intermittent vortexing, andgentle heating. All solutions were prepared under a blanket ofdry argon and quickly sealed with parafilm to minimize sorptionof environmental moisture.
A Perkin-Elmer LambdaBio 20 double-beam spectrophotom-eter with variable bandwidth was used for the acquisition ofUV-vis electronic absorption spectra. Stationary-state fluores-cence spectra were acquired on a Jobin-Yvon FluoroLog-3(model FL-3-11) modular spectrofluorimeter equipped withsingle-grating Czerny-Turner monochromators as wavelength-selection devices, a 450 W Xe arc lamp as the excitation source,and a photomultiplier tube as the detector. All absorbance andfluorescence data were acquired using 1 cm2 quartz cuvettes.Fluorescence spectra of the probes were collected with 2 nmexcitation and emission slit widths, with the exception of pyrene-containing probes for which a 1 nm emission slit width wasused. All spectroscopic measurements were performed intriplicate with samples prepared individually and the resultsaveraged. All spectra were duly corrected by measuring thespectral responses from suitable blanks prior to data analysisand statistical treatment. Pyrene fluorescence lifetimes weremeasured based on time-correlated single-photon counting(TCSPC)9 using a time-resolved spectrofluorimeter (Model FL900CDT, Edinburgh Analytical Instrument, UK) describedpreviously.10 A hydrogen gas-filled 40 kHz flash lamp (0.4 bar)was used as the excitation source (λex,peak ≈ 340 nm). Theinstrument response function (IRF) was obtained from acolloidal silica scattering solution (Ludox TM-50, Aldrich).Fluorescence correlation spectroscopy (FCS) measurements werecarried out on a custom-built inverted two-photon excitationfluorescence microscope system described previously.11
Results
Water-soluble Betaine Dye Solvatochromism and Kamlet-Taft Parameters. As detailed in earlier studies,3c we haveutilized the absorption-based solvatochromism of the water-soluble betaine dye 2,6-dichloro-4-(2,4,6-triphenyl-N-pyridino)-phenolate (betaine dye 33) to assess the general dipolarity/polarizability and hydrogen-bond donor (HBD) traits of mixturesof [bmim][BF4] and 10 mM pH 7.0 PB. ET(33), defined as themolar transition energy in kcal mol-1 of Reichardt’s dye 33 atroom temperature and normal pressure, is obtained using thelowest-energy intramolecular charge-transfer band maximum
Interesting and Unusual Solvatochromic Probe Behavior J. Phys. Chem. B, Vol. 113, No. 10, 2009 3089
(λmax/nm) according to ET(33) ) 28591.5/λmax (nm). Anestablished empirical relationship was used to convert ET(33)values to the more commonly employed ET(30),12 which wassubsequently used to arrive at the corresponding ET
N values usingeq 1.
In this equation, TMS represents tetramethylsilane andET(30)WATER and ET(30)TMS reference values are 63.1 and 30.7kcal mol-1, respectively. Because ET
N is normalized and dimen-sionless, interpretation is simplified, with values typicallyvarying between 0.0 for TMS (highly nonpolar) and 1.0 forwater (very polar). The ET
N values determined within[bmim][BF4] + PB (10 mM, pH 7.0) mixtures of differentcompositions are shown in the inset of Figure 1. The overalltrend in ET
N is in excellent agreement with our earlier observa-tions for aqueous (unbuffered) [bmim][BF4].3c It is obvious fromthese data that the ET
Nof PB is much higher than that of neat[bmim][BF4], indicating a considerably higher dipolarity/po-larizability and/or HBD acidity within the former. As iscommonly observed, the ET
N values do not exhibit a simpledependence upon solvent mixing, as signified by the digressionfrom the dashed profile that denotes the ideal behavior calculatedon a mole fraction basis according to
where SPi is the solvatochromic parameter of interest (in thiscase, ET
N) measured in solvent i present at mole fraction xi.Indeed, the experimental ET
N values fall well below predictions
based upon mole fraction additive spectral responses from theneat solvent components, with a similarly large deviationoccurring across much of the PB-rich phase (i.e., 0.9 g xPB g0.6).
As reported by our groups3 and others, information gainedfrom the use of suitable solvatochromic probes such as N,N-diethyl-4-nitroaniline (DENA) and 4-nitroaniline (NA) can becombined with ET(30) results to yield empirical Kamlet-Taftparameters for ILs and their mixtures, notably dipolarity/polarizability (π*), HBD acidity (R) and HBA basicity ().Experimental Kamlet-Taft parameters π*, R, and for[bmim][BF4] + PB mixtures were determined in 0.1 molefraction steps and are summarized in Figure 1. The overallbehavior is similar to that reported previously for aqueous[bmim][BF4],3c suggesting that inclusion of phosphate salt inthe aqueous component exerts minimal influence over themeasured Kamlet-Taft polarities in this system. This is notaltogether surprising given the similar behavior for the ET
N
parameter just discussed and the low ionic strength of PB used.There are two significant outcomes of the results presented inFigure 1. First, for each Kamlet-Taft parameter investigated,there is clear departure from the behavior predicted using eq 2.Second, the maximum deviation from mixing ideality consis-tently occurs near an xPB of 0.9. Thus, it appears that the additionof fractional amounts of [bmim][BF4] to PB results in strikingchanges in the cybotactic region surrounding the various probesolutes. The most expedient explanation for this outcomeconsistent with these data is that preferential solvation of theprobes by [bmim][BF4] takes place. To be sure, once xPB dropsto about 0.8, the hydrogen bond interactions (R, ) observedare similar to those for bulk [bmim][BF4], wherein xPB is 0.However, at this point, contributions due to structural alterationsarising from IL/water interactions cannot be ignored.
To further describe our obtained results, we applied thecombined nearly ideal binary solvent/Redlich-Kister (CNIBS/R–K) equation12,13 to our experimental data for the various SPs.According to CNIBS/R-K theory, the empirical SPs observedwithin a binary solvent mixture at constant temperature can beexpressed by eq 3.
In eq 3, SPm, SP[bmim][BF4], and SPPB are the SPs measuredfor the mixture, for neat [bmim][BF4], and for PB, respectively;Aj and j are the equation coefficients and the degree ofpolynomial expansion. In practice, the numerical values of j
Figure 1. Composition-dependent variation in the empirical Kamlet-Taftparameters characterizing the dipolarity/polarizability (π*), HBD acidity(R), HBA basicity (), and ET
N polarity (provided in inset) parametersdetermined for the [bmim][BF4] + PB (10 mM, pH 7.0) mixture underambient conditions. Dashed lines represent mole-fraction weighted idealadditive values, and the solid curves denote best-fits to Redlich-Kistermodels as per eq 3; fitting parameters are assembled in Table 1.
are varied to find an accurate mathematical representation ofthe experimental data. The results of regression analysisperformed to fit such a polynomial to our experimental dataare compiled in Table 1. In all cases, excellent fits toexperimental SPs were achieved for j ) 3. Calculated SPs usingthese recovered Aj values are presented as solid curves for allfour parameters shown in Figure 1. As can be seen, the observedresults are accurately described by CNIBS/R-K model predic-tions across the full composition range, and no proportional orsystematic errors are evident.
Static Fluorescence Dipolarity Probes. Molecular fluores-cence from an appropriate fluorophore is well-suited to furnishinformation regarding complex systems owing to the highsensitivity and orthogonality of information inherent to fluo-rescence techniques.14 In some previous investigations, we havedemonstrated the effectiveness of using a variety of fluorescenceprobes to obtain key insight into solute-solvent and solvent-solvent interactions within IL-based solvent mixtures.1g,2,3 Tofurther interrogate the dipolarity within [bmim][BF4] + PBmixtures, we have utilized two valuable fluorescence dipolarityprobes, pyrene and pyrene-1-carboxaldehyde.
Pyrene is one of the most widely used neutral fluorescenceprobes for polarity studies.2,3,15 The pyrene solvent polarity scale(Py I1/I3) is defined by its I1/I3 emission intensity ratio, whereI1 is the intensity of the solvent-sensitive band arising fromthe S1(V ) 0) f S0(V ) 0) transition and I3 corresponds to thesolvent-insensitive S1(V ) 0) f S0(V ) 1) transition.14f-k TheI1/I3 ratio increases with increasing solvent dipolarity and is afunction of both the solvent dielectric (ε) and the refractive index(n) via the dielectric cross term, f(ε,n2). Measured values forthe Py I1/I3 ratio in the [bmim][BF4] + PB system aresummarized in Figure 2A. Contrary to observations for the ET
N,π*, and R parameters, the Py I1/I3 indices point toward a higherdipolarity within [bmim][BF4] than in neat 10 mM pH 7.0 PB(1.95 ( 0.02 vs 1.79 ( 0.04). Even more interesting, as PB isinitially added to [bmim][BF4], instead of decreasing asexpected, I1/I3 actually rises and increases to values surpassing
the ratio determined in pure [bmim][BF4], hinting towardprominent solute-solvent and/or solvent-solvent interactionsand possible microsegregation. A similar “hyperpolarity” wasrecently reported for ET(33), π*, and R values within[bmim][PF6] mixtures with tetraethylene glycol (TEG).15p Tomore carefully assess the extent of solute-solvent and/orsolvent-solvent interactions present within the mixed system,an idealized I1/I3 value was calculated from measurements madeon pyrene in the neat solvents, following the methods of Acreeand co-workers.16
To utilize this expression, pyrene I1 and I3 intensities weremeasured in neat [bmim][BF4] + PB (10 mM, pH 7.0) underidentical conditions, and (I1/I3)calc for the mixtures calculatedusing the bulk mole fraction, xi, for each component. The brokenline in Figure 2A represents the mole fraction weighted responseaccording to eq 4, and the solid curve passing through theexperimental results (open circles) presents the nonlinearregression result according to the CNIBS/R-K model (Table1, vide supra). The experimental I1/I3 values deviate significantlyfrom those predicted assuming simple mixing, and yet againdivergence is greatest in the buffer-rich region with themaximum difference displayed at xPB ) 0.8.
Our results for pyrene behavior within the [bmim][BF4] +PB system are complemented by the behavior of pyrene-1-carboxaldehyde (PyCHO), a pyrene probe analog containing analdehyde functionality.14a,15 PyCHO has two types of closelylying excited singlet states (n-π* and π-π*), both of whichshow emission in fluid solution. In nonpolar solvents, theemission from PyCHO is highly structured and weak (φF < 10-3
in hexane), arising exclusively from the n-π*state. On increas-ing the polarity of the medium, however, the π-π* state isbrought below the n-π* state via solvent relaxation to becomethe emitting state, manifested by broad, moderately intenseemission (φF ≈ 0.15 in MeOH) that red-shifts with increasingsolvent dielectric.3a,d,f,14a,b,15p Figure 2B presents experimentalPyCHO λmax,fl values in [bmim][BF4] + PB mixtures at ambientconditions. It is noteworthy that PyCHO λmax,fl is higher in neatPB in comparison to [bmim][BF4], an observation in agreementwith the known static dielectric constant for water (78.4 at 298K) and the one estimated for [bmim][BF4] (11.7).17 The ideal,additive behavior calculated according to eq 2 using (λmax,fl)-1
as the SP is indicated by the dashed line. The CNIBS/R-K fitis also provided. One finds that the PyCHO behavior parallelsthe results observed earlier for ET
N and π*. Again, the greatestdeviation from ideality for PyCHO λmax,fl in [bmim][BF4] + PBmixtures appears in the buffer-rich region, peaking at xPB )0.9. In the present case, measured values fall close to theprediction line for xPB below 0.4. We note that the somewhatpuzzling observation that the Py I1/I3 and PyCHO λmax,fl indicessuggest a different pure phase as the more polar one (i.e., [bmim]-[BF4] and PB, respectively), this situation reflects the fact thatthe primary influence on PyCHO λmax,fl is the static dielectricconstant. A similar observation was made regarding these probesin the [bmim][PF6] + TEG system.15p
Fluorescence Excimer-based Microfluidity Probe. Varia-tions in the viscosity/fluidity of the solvent microenvironmentimmediately surrounding the fluorescent probe 1,3-bis-(1-pyr-enyl)propane (BPP) are sensitively reflected in its steady-stateemission spectrum.1g,2,3,14m,15a,b,l-p,18 Using BPP, it is well
Figure 2. Variation in Py I1/I3 (A) and PyCHO λmax,fl (B) values withxPB in [bmim][BF4] + PB (10 mM, pH 7.0) mixtures under ambientconditions. The dashed curves represent mole fraction weightedpredictions, and the solid curves are Redlich-Kister fits; the recoveredR-K parameters can be found in Table 1.
established that in addition to the vibronic emission commonto polyaromatic hydrocarbons (PAHs), a broad and featurelessband with a maximum intensity near 470 nm arises due toemission from an intramolecular excited-state dimer (excimer).Explicitly, in a low-viscosity medium, the terminal pyrene unitscan easily fold upon one another following photoexcitation toform an intramolecular excimer, yielding blue-green emission.With increasing local viscosity, however, the efficiency ofcyclization drops, with a corresponding reduction in the intensityof the excimer band normalized to the intensity of the monomerband (i.e., IE/IM).
Figure 3 shows experimental BPP excimer-to-monomeremission intensity ratios (IE/IM) measured in the [bmim][BF4]+ PB system. In this figure, the solid curve intersecting theIE/IM data represents the CNIBS/R-K modeling result. Therelative values for IE/IM in [bmim][BF4] + PB make sensequalitatively based upon the corresponding bulk viscosities ofthe pure phases (154 and 0.89 cP, respectively).17a Once again,it is instructive to contrast against expected values computedfor IE/IM. In this case, the ideal behavior was calculatedaccording to eq 4 by simply substituting IE and IM in place ofI1 and I3, respectively. The results of this exercise, shown asthe short-dashed profile in Figure 3, make conspicuous the factthat the viscosity experienced by the probe (microviscosity) atall intermediate compositions is significantly higher thanexpectations, particularly for PB-dominant mixtures. This is inaccordance with our observations discussed above, with theuppermost discrepancy between calculated and measured IE/IM
values occurring at an xPB value of 0.9.The reciprocal of the bulk viscosity (η-1 or fluidity) for the
[bmim][BF4] + PB system is also plotted in Figure 3, revealingthe parallel trends for IE/IM and η-1 with xPB. Quite remarkably,despite being essentially a measure of the microviscosity, wealso discovered that the BPP IM/IE ratio is well correlated to
the bulk viscosity of aqueous [bmim][BF4],17a the two quantitiesbeing related empirically by (Figure 3, insert):
(r2 ) 0.9943; standard error of the fit (sfit) ) 1.810).Excited-state Intensity Decays of Pyrene. Time-resolved
intensity decay profiles for pyrene dissolved in [bmim][BF4] +PB were collected as another means to investigate this binarymixture. The time-resolved intensity decay traces for pyrene inpure PB were rigorously single-exponential with a recoveredexcited-state fluorescence lifetime of 86 ns. Interestingly,however, we found that for all mixtures containing [bmim][BF4],a triple-exponential decay model was required to adequatelydescribe the excited-state intensity decay, I(t):
In this expression, Ri denotes the pre-exponential amplitudeassociated with the ith component possessing an excited-statelifetime τi. The fractional contribution (fi) of each excited-statelifetime to the observed emission is given by
Representative pyrene excited-state fluorescence intensitydecay data and the fits to a triple-exponential decay model forxPB ) 0, 0.3, 0.6, and 0.9 under ambient conditions are presentedin Figure 4. We note that including fewer than three componentsresulted in statistically poorer fits based on the residuals,autocorrelations, and 2 values. Recovered fitting parameters,including a goodness-of-fit parameter (2) and the intensity-weighted mean lifetime <τ>, are compiled in Table 2. The latterwas calculated from
Careful examination of the entries in Table 2 reveals severalkey findings. The most striking feature of these results is thefact that the fractional contribution from the longest-lived pyrenespecies (τ3) is remarkably constant at ca. 96% of the total pyrenesignal within all [bmim][BF4] + PB mixtures, although τ3
decreases monotonically from 203 ns in pure [bmim][BF4] to124 ns at an xPB value of 0.9. Naturally, the principal contribu-tion from τ3 is reflected in the <τ> values, which similarlyplummet from 196 to 117 ns over the range of xPB studied. Thetwo minor species are much shorter lived. The first (τ1) accountsfor ∼1.9% of the observed emission, with an excited-statelifetime that increases from 0.33 to 0.96 ns as xPB increasesfrom 0.0 to 0.9. Species τ2 has an average lifetime near 6 nsand constitutes roughly 2% of the total fluorescence, althoughthe measured values vary considerably about these averages.Because these samples were not deoxygenated, one plausibleexplanation for the trend in <τ> is that the increased fluidityaccompanying higher xPB results in a larger diffusion coefficient
Figure 3. Variation in BPP IE/IM (left axis) with xPB in the [bmim][BF4]+ PB mixture. The dashed line indicates the IE/IM response calculatedusing a version of eq 4, and the solid curve is a fit to a Redlich-Kistermodel. The reciprocal viscosity (η-1 where η is the bulk viscosity)plotted as a function of xPB is also included. Provided in the inset isthe correlation between η and IM/IE for this system, as described byeq 5.
for oxygen and more efficient collisional quenching of pyrene.In support of this proposition, in earlier work the lifetime forpyrene in [bmim][PF6] was reportedly 250 ns.3h The longer-lived pyrene in this case is fully consistent with a furtherlowering in oxygen diffusivity resulting from the higher η of[bmim][PF6] relative to [bmim][BF4]. Nonetheless, this sug-
gestion remains purely speculative at this point and deservescloser scrutiny.
Nitromethane Quenching of Pyrene Fluorescence. Fluo-rescence quenching is a process whereby the fluorescenceintensity of a sample decreases via interaction with a quencherspecies.14d,19 In dynamic (collisional) quenching, the quenchermust diffuse to the fluorophore during its excited-state lifetime,whereas a complex is formed between the fluorophore and thequencher in static quenching. In either case, molecular contactbetween the fluorophore and the quencher is required. Thisrequirement can be exploited to obtain key dynamical insightinto complex systems like [bmim][BF4] + PB. For this purpose,we have selected the pyrene/nitromethane (CH3NO2) fluoro-phore-quencher pair. It is well-established that the quenchingof pyrene fluorescence by nitromethane is due to an electron/charge transfer reaction wherein electron/charge is transferredfrom excited-state pyrene to nitromethane, which acts as anelectron/charge acceptor.14a-e,k-n The quenching efficiency forthis fluorophore-quencher dyad is understood to dependcritically on the nature of the medium.
Figure 5 presents F0/F versus [CH3NO2] quenching profilesfor [bmim][BF4] + PB mixtures under ambient conditions. Bythis designation, F0 and F are the fluorescence intensities of
Figure 4. Representative pyrene excited-state intensity decay curves in [bmim][BF4] at 298 K for select values of xPB (0, 0.3, 0.6, 0.9); (λem ) 376nm; solutions were not deoxygenated). The bottom curves denote the instrumental response function measured using a dilute colloidal silica (Ludox)suspension. The top panels provide triple-exponential fits to experimental data, and the lower panels show weighted residuals of the correspondingfits. The model parameters are collected in Table 2.
a τi and fi are determined from triple-exponential fits ofexcited-state intensity decay data to eqs 6 and 7; uncertainties areexpected to be on the order of ( 5% for both τi and fi. <τ > is themean excited-state lifetime calculated from eq 8.
Interesting and Unusual Solvatochromic Probe Behavior J. Phys. Chem. B, Vol. 113, No. 10, 2009 3093
pyrene in the absence and presence of nitromethane, respec-tively. Unexpectedly, with the exception of PB, F0/F versus[CH3NO2] plots for all samples showed distinct upwardcurvature at higher quencher concentrations, suggesting that thisluminophore-quencher pair does not obey simple Stern-Volmerbehavior in the presence of [bmim][BF4].14d,19 If dynamicquenching alone were operative, linear plots of F0/F versus [Q]would be anticipated, where [Q] is the molar quencherconcentration.14d Indeed, this is the case for [CH3NO2] quenchingof pyrene in neat PB. For samples other than xPB ) 1.0, however,the classic Stern-Volmer model of a single species beingquenched bimolecularly is entirely inadequate at describing thequenching behavior.
For quenching within rigid solutions, Perrin introduced theconcept of an “active sphere”,20 a volume of interaction arounda quencher molecule such that a fluorophore excited within thisvolume is instantaneously quenched. This model was latermodified by Frank and Vavilov21 by combining the Perrin modelwith a dynamic quenching one. The Frank-Vavilov modelprovides a good description of quenching within solutions ofhigh viscosity with short-range electron/charge-exchange inter-action, as in the present case, with an active sphere radius of∼10-15 Å.14e The form of the Stern-Volmer equation modifiedfor this case is given by eq 9,
where V is the volume of the quenching sphere-of-action, andKD is the dynamic (collisional) quenching constant. All curvesin Figure 5 (save for an xPB of unity) represent fits of our F0/Fdata to this quenching sphere-of-action model. The recoveredparameters resulting from fits to eq 9 are collected in Table 3.A quenching sphere-of-action radius (R/Å), calculated from V,has also been included. In each case, the quenching behavior iswell described by a sphere-of-action model (r2 > 0.9936), evenin the buffer-rich regime. It is noteworthy that the radius ofquenching (R ≈ 12.3 Å) is independent of solution composition
and is appropriate to the short-range electron/charge-exchangemechanism of interaction between CH3NO2 and pyrene givingrise to quenching in viscous solutions.14e On the contrary, therecovered KD values increase nearly 8-fold as xPB increases from0.0 to 0.9 in the [bmim][BF4] + PB system. We attribute this,in large part, to the decreased viscosity of the solution as thefraction of PB increases.14d,e,19 To support this proposition,bimolecular quenching rate constants (kq) were calculated usingthe following equation
and are also listed within Table 3. The kq value, in turn, can bedissected into a quenching efficiency (fQ) and a diffusion-controlled bimolecular rate constant (k0), which relates to probe/quencher sizes and diffusion coefficients (D), as given by theSmoluchowski equation14d
By this relation, the collision radius is assumed to be the sumof the molecular radii of the fluorophore and quencher (Rf andRq, respectively), Df and Dq are the corresponding diffusioncoefficients, and N is Avogadro’s number. A general relationshipbetween D and η is provided by the Stokes-Einstein equation:14d
where R is the molecular radius. From these relationships, it isreasonable to expect that kq should dramatically increase withxPB as a result of the reduction in η. The relative increase in kq,however, fails to keep pace with the declining η, suggestingpossible changes in the effective radii of the diffusing species,Rf and/or Rq.
Fluorescence Correlation Spectroscopy Diffusion Studies.Fluorescence correlation spectroscopy (FCS) is an elegant toolfor studying diffusional processes in complex media. Offeringfew- or even single-molecule detection sensitivity, FCS is basedupon analysis of temporal fluctuations in the fluorescenceintensity arising from a well-defined open volume of solution.In earlier work from one of these authors, FCS was used to
Figure 5. Stern-Volmer plots showing nitromethane quenching ofpyrene in nondeoxygenated [bmim][BF4] + PB mixtures at 298 K (xPB
values are indicated alongside quenching profiles). The dashed profilesidentify results in neat solvent components (i.e., xPB ) 0.0 or 1.0) andall curves represent fits to a quenching sphere-of-action model (eq 9),except for xPB ) 1.0, in which case quenching followed simple, linearStern-Volmer behavior (for V ) 0, eq 9 becomes: F0/F ) 1 +KD[CH3NO2]). The typical error in F0/F is <10%. Within the highlightedwindow, F0/F increases with xPB in the expected order for a given[CH3NO2] concentration.
F0/F ) (1 + KD[CH3NO2])exp(V[CH3NO2]) (9)
TABLE 3: Summary of Pyrene/Nitromethane“Sphere-of-Action” Quenching within [bmim][BF4] + PBa
probe the diffusion of differentially charged molecular probesfreely dissolved in dry and hydrated [bmim][PF6].4q In thepresent work, in order to explore molecular mobility within the[bmim][BF4] + PB mixture, we extend the utility of FCS toinclude determination of the diffusion rates of two tracer speciespossessing significantly different molecular masses: a typicallysized organic dye (BODIPY FL; ∼300 Da) and a fluorescentbiomacromolecular conjugate (Texas Red-dextran; nominally104 Da).
Experimentally, the time-dependent fluorescence intensityF(t) in FCS is analyzed in terms of its temporal autocorrelationfunction G(τ), which compares the fluorescence intensity at timet with the intensity after a time delay τ, as given by eq 13:
In this expression, the angle-bracketed quantities denote timeaverages, where deviations in the fluorescence light intensityfrom the average (δF(t) ) F(t) -⟨F(t)⟩) may result from changesin either the number or the quantum yield of molecules residentin the detection volume. FCS data were initially processed withVista FCS software (ISS, Champaign, IL) to construct auto-correlation curves, with further analysis performed with GlobalsUnlimited software (Laboratory for Fluorescence Dynamics,Urbana, IL) to calculate diffusion coefficients using a three-dimensional Gaussian model. The following functional form wasused to fit our experimental data
where N is the average number of molecules in the observationvolume, D is the diffusion coefficient, ω0 is the beam waist inthe xy-plane, and ωz is the beam waist in the z-plane. The valuesof ω0 and ωz were calibrated using an analytically prepared 35nM solution of Rhodamine 110 (R110; Invitrogen, Carlsbad,CA). The autocorrelation curve of R110 was fitted with thesingle-component model using its known diffusion coefficientof 270 µm2 s-1. Typically, the resulting ω0 value and ωz/ω0
ratio were 0.3 µm and 4.0, respectively. The autocorrelationcurves for our unknown samples were then fitted against eq 14using these fixed ω0 and ωz values and the diffusion coefficientsrecovered from the fit.
Figure 6 shows autocorrelation curves, G(τ), for BODIPYFL (panel A) and Texas Red-dextran (B) within various[bmim][BF4] + PB mixtures at 298 K. The solid curvesrepresent fitting results of G(τ) according to eq 14. The lowermobility of the much larger 10 kDa dextran labeled with TexasRed (TR) can be directly observed by the lengthening in thetranslational diffusion time, apparent from the temporal shiftin the G(τ) curves for TR-dextran relative to those of BODIPYFL. Recovered D values for both diffusing fluorescent probesare plotted as the insets within Figure 6. As expected, Ddecreases with increasing vol % [bmim][BF4] in both cases.This decrease in D, due in part to the rise in viscosity, isanalogous to the kq results observed for the CH3NO2-mediatedquenching of pyrene discussed in the previous section. It shouldbe noted that a truncated composition range was employed inthe FCS experiments, allowing us to focus on the buffer-richcompositions spanning xPB from about 0.8-1.0.
In Figure 7 we plot FCS-determined values of D versus η-1
for both BODIPY FL and TR-dextran. As is evident, a fairly
poor linear correlation between D and η-1 exists in either case(r2 values of 0.865 and 0.898, respectively). On the other hand,the experimental trend shows intriguing evidence for a bilinearbehavior with a break point near a similar value of η for eitherprobe (∼1.3 cP). Interestingly, despite an over 30-fold differencein the size of the diffusing unit, BODIPY FL and TR-dextranindeed generate fairly similar D-η-1 profiles (see Figure S1 inthe Supporting Information). For example, at 25 vol %[bmim][BF4] the experimental D values are roughly half theexpected values for both probes. The solid upper curvesenclosing the shaded regions within Figure 7 show the behaviorexpected from the Stokes-Einstein equation (eq 12), based on
G(τ) ) ⟨δF(t)δF(t + τ)⟩⟨F(t)⟩2
(13)
G(τ) ) γN(1 + 8Dτ
ω02 )-1(1 + 8Dτ
ωz2 )-1/2
(14)
Figure 6. Experimental FCS autocorrelation function curves and theassociated fits for (A) BODIPY FL and (B) Texas Red-dextran in PBcontaining various amounts of [bmim][BF4]. The bar chart inserts showthe corresponding translational diffusion coefficients (D) recovered fromeq 14.
Figure 7. Plot of D versus η-1 for BODIPY FL (upper) and TexasRed-dextran (lower) diffusional probes in the [bmim][BF4] + PBsystem. In each case, the solid line occupying the longest leg of theshaded scalene triangle represents the anticipated response calculatedbased on changes in bulk viscosity as [bmim][BF4] is added to PB,and assuming Stokes-Einstein behavior for an invariant R (eq 12).The dashed lower segments denote best fits of measured results tobilinear functions.
Interesting and Unusual Solvatochromic Probe Behavior J. Phys. Chem. B, Vol. 113, No. 10, 2009 3095
the increase in bulk η as [bmim][BF4] is added to PB. Theexperimental results clearly indicate diffusivities much lowerthan these estimates. For instance, in the case of BODIPY FL,D suffers a 15-fold decrease, whereas the corresponding bulkviscosity is only lowered by a factor of about five. This impliessignificant changes in effective probe volume, presumably as aresult of changes in local solvent composition and/or structuring.
Discussion
The behavior of solvatochromic absorbance and fluorescenceprobes reveals important information regarding solute-solventand solvent-solvent interactions within complex fluid systems.Limited solvent-solvent interaction results within aqueous[bmim][BF4] have been reported in a few previous studiesrelying upon noninvasive (probe-free) methods.4 In what fol-lows, we briefly discuss several of these reports within thecontext of our current results.
On the basis of ROESY spectra, Mele and co-workersprovided strong evidence for the presence of water-[bmim+]interactions within aqueous [bmim][BF4]4a These authors showedthe interaction of water to be specific and localized at theC(2)-H proton for low levels of hydration. At high watercontents, the interactions became less selective, and otherprotons on [bmim+] also became involved. Far-IR spectroscopicinvestigation by Lendl and co-workers strongly implied thepresence of stretched H-bonds between [BF4
-] and watermolecules.4f A very recent investigation based on moleculardynamics simulation with supporting NMR results from Raoset al. showed that water clusters form almost exclusively vialinear chains of H-bonded molecules within aqueous [bmim]-[BF4].4e Nanoscale structuring of the mixture was proposed withno possibility of macroscopic phase separation among thecomponents. Two solvation regimes, one at low water contentswhere [bmim+] and [BF4
-] are selectively coordinated byindividual water molecules, and another where the [bmim][BF4]ionic network appears to be somewhat disrupted or swollen ina nonspecific way by the water clusters, were proposed. Aninteresting outcome was proposed by Bowers and co-workersfrom surface tension measurements of [bmim][BF4] as a functionof water concentration.6a The behavior of the IL was found toa considerable extent to be analogous to that of a surfactantwith a critical aggregation concentration of ∼800 mM at whichthere was an aggregate formation of the IL. On the basis ofSANS data analysis, the aggregation of [bmim][BF4] was bestmodeled by treating it as a dispersion of polydisperse sphericalaggregates.6a The same type of aggregate formation was detectedby Koga et al. from thermodynamic analysis.6b,c These authorsobserved that at concentrations above 0.5 mole fraction the ILformed clusters as in the pure state, and water molecules, withoutinteracting among themselves, interacted with these clusters.
These aforementioned observations on solvent-solvent in-teractions furnish key insight into the probe behavior withinaqueous [bmim][BF4]. It is clear that the absorbance probebehavior and empirical Kamlet-Taft indices (i.e., ET
N, π*, R,and ) in the mixture are dominated by their values in neat[bmim][BF4]. Whereas R and show relatively little changeon buffer addition from their values in neat [bmim][BF4], theincreases in ET
N and π* are rather gradual. It appears that theabsorbance probes are preferentially solvated within an IL-richcybotactic region,4e minimizing the influence of water on probebehavior. Conversely, dramatic changes in solvatochromic probebehavior are observed for addition of just 10 mol %[bmim][BF4] to PB. As shown earlier, the probe behaviorgenerally changes little upon further increase in the amount of
[bmim][BF4] in the mixture. This suggests that the localenvironment becomes essentially saturated in IL by selectivesolvation. As demonstrated by Bower6a and Koga,6b,c themajority of the IL [bmim][BF4] is present in aggregated format xPB g 0.1. It is easy to envisage that the probes may beinteracting with and/or partitioning into these IL aggregatesand, in turn, experiencing solvation characteristic of neat[bmim][BF4]. Similarly dramatic changes in probe behavior wererevealed for the polarity probes pyrene and PyCHO, furthersubstantiating this viewpoint. The unusually high experimentalI1/I3 values (hyperpolarity) measured for pyrene in the IL-richregime may be attributed to the unusual dipolarity of the pyrenecybotactic region due to the interaction between pyrene’sπ-cloud and [bmim+] resulting from preferential solvation ofexcited-state pyrene by [bmim][BF4]. PyCHO, while showinginteraction with aggregated [bmim][BF4] in the buffer-richregion, has an overall behavior more in line with predictions.The presence of the hydrogen-bond amenable aldehyde groupfavors interaction with water and may be the reason behind thesimplified behavior. Although preferential solvation of BPP by[bmim][BF4] would certainly result in decreased fluidity, thepresence of strengthened hydrogen-bonded networks betweenwater and [bmim][BF4] components might possibly yield asimilar effect. The trend in the measured bulk viscosity of themixture also supports the presence of strong solvent-solventinteractions within [bmim][BF4] + PB.
The rate of nonradiative deactivation for excited-state pyrenedepends strongly upon the surroundings and is observed to befacilitated within aqueous media.14e As a consequence, the meanexcited-state lifetime of excited-state pyrene is expected to belowered in the presence of water. Although, as alreadymentioned, there exists the possibility for pyrene preferentialsolvation by [bmim][BF4] as well as interaction between theπ-cloud of pyrene and [bmim+], the average lifetime of pyrenedecreases gradually as water is incrementally added to [bmim]-[BF4]. Excited-state fluorescence intensity decays of pyreneappear to be highly sensitive to the presence of water. A factthat supports our earlier proposition is that the change in pyreneexcited-state decay behavior is relatively minor upon going fromneat [bmim][BF4] to [bmim][BF4] containing 10 mol % PB,whereas a more dramatic change is observed on proceeding from[bmim][BF4] containing 90 mol % PB to neat PB (Table 2).
Information about bimolecular quenching rate constants fromfluorescence quenching behavior is nicely complemented bytranslational diffusion coefficients obtained from FCS analysis.The fact that upward curvature is observed in F0/F quenchingprofiles even for 5 mol % [bmim][BF4] in PB is in accord withpyrene localization, partitioning, or preferential solvation withinaggregated [bmim][BF4] clusters. The unusually enhancedmicroviscosity manifests in upward curvature, a behavior wellaccounted for by using a sphere-of-action quenching model.Understandably, the volume of the quenching sphere-of-actionshows little variation, whereas the bimolecular quenching rate(kq) representing diffusion between quencher and fluorophoreincreases as xPB in [bmim][BF4] rises. It is satisfying to notethat the value of kq in neat PB nearly approaches the value forthe diffusion-controlled bimolecular rate constant (k0) given bythe Smoluchowski equation (eq 11).14d Translational diffusioncoefficients of BODIPY FL and TR-dextran obtained via FCSmeasurements further substantiate our fluorescence quenchingresults. Although these two fluorescence probes are disparatein size, the trend in measured diffusion coefficients reflectssimilar fluidity changes in the [bmim][BF4] + PB system. Asexpected, with increasing [bmim][BF4], the translational diffu-
3096 J. Phys. Chem. B, Vol. 113, No. 10, 2009 Sarkar et al.
sion coefficients of the two probes decrease. The absence ofgood linear correlation between the translational diffusioncoefficients of these probes and η-1 reflects marked microhet-erogeneous changes in the aqueous [bmim][BF4] system at abuffer-rich composition where [bmim][BF4] likely exists inaggregated form.
Summary
Unusual solvatochromic probe behavior within the [bmim]-[BF4] + PB system strongly suggests the presence of pro-nounced solute-solvent interactions within the system. That is,the unusual probe behavior is a clear manifestation of thepresence of specific interactions among the solution components.The most prominent deviation from ideal behavior for all thesolvatochromic probes studied is observed within the buffer-rich region of the mixture. This is due, in major part, to theinteraction of the solute with the aggregated [bmim][BF4]clusters within the mixture. The presence of strong solvent-solvent interactions (likely driven by H-bonding) is furtherreflected in changes in the bulk viscosity as well as themicroviscosity of the mixture, which in turn control the diffusivebehavior of solutes dissolved therein. It may be concluded thatthe failure of simple quenching and diffusion models to explainthe observed results is a manifestation of the inherent micro-heterogeneity associated with the [bmim][BF4] + PB mixture.This investigation not only affirms the presence of strongsolvent-solvent interactions and microheterogeneity within thisimportant aqueous IL system, but their direct influence on themobility and diffusion of molecular species within the mixturecan be properly rationalized and appreciated for the first time.These more novel aspects of this work should be of generalinterest as they offer a broadly applicable strategy for tacklingissues concerning solute dynamics, interaction, and diffusionin other IL mixtures.
Acknowledgment. This work was generously supported bythe Department of Science and Technology (DST), India throughthe grant SR/S1/PC-38/2004 to S.P. The authors thank Dr. N. K.Chaudhury, Division of Biocybernetics, Institute of NuclearMedicine and Allied Sciences, Delhi, India for his assistancewith pyrene lifetime measurements. M.A. would like to thankUGC, India for a fellowship. G.A.B. also thanks Basic EnergySciences, U.S. Department of Energy, under Contract DE-AC05-0096OR22725 with Oak Ridge National Laboratory, for finan-cial support.
Supporting Information Available: D-η-1 profiles forBODIPY FL and TR-dextran. This material is available free ofcharge via the Internet at http://pubs.acs.org.
References and Notes
(1) (a) Ionic Liquids in Synthesis; Wasserscheid, P., Welton, T., Eds.;Wiley-VCH Verlag: Weinheim, 2003. (b) Weingartner, H. Angew. Chem.,Int. Ed. 2008, 47, 654. (c) Baker, G. A.; Baker, S.; Pandey, S.; Bright,F. V. Analyst 2005, 130, 800. (d) Pandey, S. Anal. Chim. Acta 2006, 556,38. (e) Baldelli, S. Acc. Chem. Res. 2008, 41, 421. (f) Burrell, A. K.; DelSesto, R. E.; Baker, S. N.; McCleskey, T. M.; Baker, G. A. Green Chem.2007, 9, 449. (g) Ionic Liquids in Chemical Analysis; Koel, M., Ed.; CRCPress: Boca Raton, 2009. (h) Electrochemical Aspects of Ionic Liquids;Ohno, H., Ed.; Wiley-Interscience: Hoboken, NJ, 2005. (i) Poole, C. F.J. Chromatogr. A 2004, 1037, 49. (j) Liu, J. F.; Jonsson, J. A.; Jiang, G. B.Trends Anal. Chem. 2005, 24, 20. (k) Berthod, A.; Ruiz-Angel, M. J.; Carda-Broch, S. J. Chromatogr. A 2008, 1184, 6. (l) Zhao, H.; Baker, G. A.;Song, Z.; Olubajo, O.; Crittle, T.; Peters, D. Green Chem. 2008, 10, 696.(m) Jin, H.; O’Hare, B.; Dong, J.; Arzhantsev, S.; Baker, G. A.; Wishart,J. F.; Benesi, A. J.; Maroncelli, M. J. Phys. Chem. B 2008, 112, 81. (n)Hough, W. L.; Rogers, R. D. Bull. Chem. Soc. Jpn. 2007, 80, 2262. (o)Plechkova, N. V.; Seddon, K. R. Chem. Soc. ReV. 2008, 37, 123. (p)
Greaves, T. L.; Drummond, C. J. Chem. ReV. 2008, 108, 206. (q) Castner,E. W.; Wishart, J. F.; Shirota, H. Acc. Chem. Res. 2007, 40, 1217. (r) Angell,C. A.; Byrne, N.; Belieres, J. P. Acc. Chem. Res. 2007, 40, 1228. (s) Freire,M. G.; Santos, L. M. N. B. F.; Fernandes, A. M.; Coutinho, J. A. P.;Marrucho, I. M. Fluid Phase Equilib. 2007, 261, 449. (t) Hardacre, C.;Holbrey, J. D.; Nieuwenhuyzen, M.; Youngs, T. G. A. Acc. Chem. Res.2007, 40, 1146. (u) Luo, H.; Baker, G. A.; Dai, S. J. Phys. Chem. B 2008,112, 10077. (v) Dominguez de Maria, P. Angew. Chem., Int. Ed. 2008, 47,6960.
(2) Fletcher, K. A.; Pandey, S. Langmuir 2004, 20, 33.(3) (a) Fletcher, K. A.; Pandey, S. Appl. Spectrosc. 2002, 56, 266.
(b) Pandey, S.; Fletcher, K. A.; Baker, S. N.; Baker, G. A. Analyst 2004,129, 569. (c) Sarkar, A.; Pandey, S. J. Chem. Eng. Data 2006, 51, 2051.(d) Fletcher, K. A.; Pandey, S. Appl. Spectrosc. 2002, 56, 1498. (e)Fletcher, K. A.; Pandey, S. J. Phys. Chem. B 2003, 107, 13532. (f)Fletcher, K. A.; Baker, S. N.; Baker, G. A.; Pandey, S. New J. Chem.2003, 27, 1706. (g) Baker, S. N.; Baker, G. A.; Bright, F. V. GreenChem. 2002, 4, 165. (h) Baker, S. N.; Baker, G. A.; Kane, M. A.; Bright,F. V. J. Phys. Chem. B 2001, 105, 9663.
(4) (a) Mele, A.; Tran, C. D.; Lacerda, S. H. D. P. Angew. Chem., Int.Ed. 2003, 42, 4364. (b) Saha, S.; Hamaguchi, H. J. Phys. Chem. B 2006,110, 2777. (c) Cammarata, L.; Kazarian, S. G.; Salter, P. A.; Welton, T.Phys. Chem. Chem. Phys. 2001, 3, 5192. (d) Koddermann, T.; Wertz, C.;Heintz, A.; Ludwig, R. Angew. Chem., Int. Ed. 2006, 45, 3697. (e) Moreno,M.; Castiglione, F.; Mele, A.; Pasqui, C.; Raos, G. J. Phys. Chem. B 2008,112, 7826. (f) Dominguez-Vidal, A.; Kaun, N.; Ayora-Canada, M. J.; Lendl,B. J. Phys. Chem. B 2007, 111, 4446. (g) Jeon, Y.; Sung, J.; Kim, D.; Seo,C.; Cheong, H.; Ouchi, Y.; Ozawa, R.; Hamaguchi, H. J. Phys. Chem. B2008, 112, 923. (h) Almasy, L.; Turmine, M.; Perera, A. J. Phys. Chem. B2008, 112, 2382. (i) Li, W.; Zhang, Z.; Han, B.; Hu, S.; Xie, Y.; Yang, G.J. Phys. Chem. B 2007, 111, 6452. (j) Martins, C. T.; Sato, B. M.; El Seoud,O. A. J. Phys. Chem. B 2008, 112, 8330. (k) Baker, S. N.; Baker, G. A.;Munson, C. A.; Chen, F.; Bukowski, E. J.; Cartwright, A. N.; Bright, F. V.Ind. Eng. Chem. Res. 2003, 42, 6457. (l) Zhang, L.; Xu, Z.; Wang, Y.; Li,H. J. Phys. Chem. B. 2008, 112, 6411. (m) Kato, H.; Nishikawa, K.; Murai,H.; Morita, T.; Koga, Y. J. Phys. Chem. B. 2008, 112, 13344. (n) Gardas,R. L.; Dagade, D. H.; Coutinho, J. A. P.; Patil, K. J. J. Phys. Chem. B.2008, 112, 3380. (o) Chang, H. C.; Jiang, J. C.; Chang, C. Y.; Su, J. C.;Hung, C. H.; Liou, Y. C.; Lin, S. H. J. Phys. Chem. B. 2008, 112, 4351.(p) Ali, M.; Baker, G. A.; Pandey, S. Chem. Lett. 2008, 37, 260. (q) Werner,J. H.; Baker, S. N.; Baker, G. A. Analyst 2003, 128, 786. (r) Baker, S. N.;Brauns, E. B.; McCleskey, T. M.; Burrell, A. K.; Baker, G. A. Chem.Commun. 2006, 2851. (s) Baker, G. A.; Heller, W. T. Chem. Eng. J. 2009,147, 6. (t) Chakrabarty, D.; Chakraborty, A.; Seth, D.; Sarkar, N. J. Phys.Chem. A 2005, 109, 1764. (u) Seth, D.; Chakraborty, A.; Setua, P.; Sarkar,N. J. Chem. Phys. 2007, 126, 224512.
(5) (a) Tokuda, H.; Baek, S. J.; Watanabe, M. Electrochemistry 2005,73, 620. (b) Ozawa, R.; Hayahi, S.; Hamaguchi, H. Chem. Lett. 2003, 32,498.
(6) (a) Bowers, J.; Butts, C. P.; Martin, P. J.; Vergara-Gutierrez, M. C.;Heenan, R. K. Langmuir 2004, 20, 2191. (b) Miki, K.; Westh, P.; Nishikawa,K.; Koga, Y. J. Phys. Chem. B 2005, 109, 9014. (c) Katayanagi, H.;Nishikawa, K.; Shimozaki, H.; Miki, K.; Westh, P.; Koga, Y. J. Phys. Chem.B 2004, 108, 19451.
(7) (a) Fletcher, K. A.; Pandey, S. Appl. Spectrosc. 2002, 56, 1498.(b) Ali, M.; Sarkar, A.; Tariq, M.; Ali, A.; Pandey, S. Green Chem 2007,9, 1252.
(8) Baker, G. A.; Baker, S. Aust. J. Chem. 2005, 58, 174.(9) O’Connor, D. V.; Phillips, D. Time-Correlated Single Photon
Counting; Academic Press: Orlando, London, 1984.(10) Gupta, R.; Mozumdar, S.; Chaudhury, N. K. Biosens. Bioelectron.
2005, 20, 1358.(11) Tetin, S. Y.; Ruan, Q. Q.; Saldana, S. C.; Pope, M. R.; Chen, Y.;
Wu, H. Q.; Pinkus, M. S.; Jiang, J. J.; Richardson, P. L. Biochemistry 2006,45, 14155.
(12) Sarkar, A.; Trivedi, S.; Pandey, S. J. Phys. Chem. B 2008, 112,9042.
(13) Trivedi, S.; Sarkar, A.; Pandey, S. Chem. Eng. J. 2009, 147, 36.(14) (a) Acree, Jr., W. E. Absorption and Luminescence Probes. In
Encyclopedia of Analytical Chemistry: Theory and Instrumentation; Meyer,R. A., Ed.; John Wiley & Sons, Ltd.: Chichester, 2000 andreferences citedtherein. (b) Reichardt, C. Green Chem. 2005, 7, 339. (c) Reichardt, C. PureAppl. Chem. 2004, 76, 1903. (d) Lakowicz, J. R. Principles of FluorescenceSpectroscopy, 3rd ed.; Kluwer Academics/Plenum Publishers: New York,2006. (e) Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Interscience: New York, 1970. (f) Turro, N. J. Modern MolecularPhotochemistry; University Science Books: Sausalito, CA, 1991. (g) Gore,M. G. Spectrophotometry and Spectrofluorimetry; Oxford University Press:London, 2000. (h) Rettig, W.; Strehmel, B.; Schrader, S.; Seifert, H. AppliedFluorescence in Chemistry, Biology, and Medicine; Springer: New York,1999. (i) Guilbault, G. G. Practical Fluorescence; Marcel Dekker: NewYork, 1990. (j) Suppan, P.; Ghoneim, N. SolVatochromism; Royal Societyof Chemistry: Cambridge, 1997. (k) Handbook of Photochemistry, 3rd ed.;
Interesting and Unusual Solvatochromic Probe Behavior J. Phys. Chem. B, Vol. 113, No. 10, 2009 3097
Montalti, M., Credi, A., Prodi, L. Gandolfi, M. T., Eds.; CRC Press/Taylorand Francis Group: Boca Raton, 2006. (l) AdVances in MultidimensionalLuminescence; Warner, I. M., McGown, L. B., Eds.; Jai Press Inc.:Greenwich, 1993; Vol. 2. (m) Bredereck, K.; Forster, Th.; Oenstein, H. G.Luminescence of Inorganic and Organic Materials; Kallman, H. P., Spruch,G. M., Eds.; Wiley: New York, 1960. (n) Pandey, S. ISRAPS Bull. 2006,18, 9. (o) Pandey, S.; Redden, R. A.; Fletcher, K. A.; Sasaki, D. Y.; Kaifer,A. E.; Baker, G. A. Chem. Commun. 2004, 1318.
(15) (a) Fletcher, K. A.; Storey, I. A.; Hendricks, A. E.; Pandey, S.;Pandey, S. Green Chem. 2001, 3, 210. (b) Fletcher, K. A.; Pandey, S.;Storey, I. A.; Hendricks, A. E.; Pandey, S. Anal. Chim. Acta 2002, 453,89. (c) Pandey, S. J. Dispersion Sci. Tech. 2005, 26, 381. (d) Pandey, S.;Reddon, R. A.; Hendricks, A. E.; Fletcher, K. A.; Palmer, C. P. J. ColloidInterface Sci. 2003, 262, 579. (e) Baker, G. A.; Bright, F. V.; Pandey, S.Chem. Educ. 2001, 6, 223. (f) Waris, R.; Acree, W. E., Jr.; Street, K. W.,Jr. Analyst 1988, 113, 1465. (g) Dong, D. C.; Winnik, M. A. Can. J. Chem.1984, 62, 2560. (h) Karpovich, D. S.; Blanchard, G. J. J. Phys. Chem. 1995,99, 3951. (i) Mazur, M.; Blanchard, G. J. J. Phys. Chem. B 2005, 109,4076. (j) Kalyanasundaram, K.; Thomas, J. K. J. Phys. Chem. 1977, 81,2176. (k) Pandey, S.; Redden, R. A.; Fletcher, K. A.; Palmer, C. P.Macromol. Chem. Phys. 2003, 204, 425. (l) Behera, K.; Dahiya, P.; Pandey,
S. J. Colloid Interface Sci. 2007, 307, 235. (m) Behera, K.; Pandey, S. J.Colloid Interface Sci. 2007, 316, 803. (n) Behera, K.; Pandey, M. D.; Porel,M.; Pandey, S. J. Chem. Phys. 2007, 127, 184501. (o) Behera, K.; Pandey,S. J. Phys. Chem. B 2007, 1, 13307. (p) Sarkar, A.; Trivedi, S.; Baker,G. A.; Pandey, S. J. Phys. Chem. B 2008, 112, 14927.
(16) Acree, W. E., Jr.; Wilkins, D. C.; Tucker, S. A.; Griffin, J. M.;Powell, J. R. J. Phys. Chem. 1994, 98, 2537.
(17) (a) Liu, W.; Zhao, T.; Zhang, Y.; Wang, H.; Yu, M. J. Soln. Chem2006, 35, 1337. (b) Murrell, J. N.; Jenkins, A. D. Properties of Liquidsand Solutions, 2nd ed.; John Wiley & Sons: Chichester, England, 1994.
(18) (a) Snare, M. J.; Thistlethwaite, P. J.; Ghiggino, K. P. J. Am. Chem.Soc. 1983, 105, 3328. (b) Zachariasse, K. A. Chem. Phys. Lett. 1978, 57,429. (c) Baker, G. A.; Baker, S. N.; McCleskey, T. M. Chem. Commun.2003, 2932.
(19) Valeur, B. Molecular Fluorescence: Principles and Applications;Wiley-VCH Verlag GmbH: Weinheim, 2002.
(20) Perrin, F. C. R. Acad. Sci. Paris 1924, 178, 1978.(21) Frank, J. M.; Vavilov, S. I. Z. Physik. 1931, 69, 100.
JP8098297
3098 J. Phys. Chem. B, Vol. 113, No. 10, 2009 Sarkar et al.
