Effect of alkyl chain length and head group on surface active and aggregation behavior of ionic liquids in water

Fluid Phase Equilibria 327 (2012) 22–29

Contents lists available at SciVerse ScienceDirect

Fluid Phase Equilibria journal homepage: www.elsevier.com/locate/fluid

Effect of alkyl chain length and head group on surface active and aggregation behavior of ionic liquids in water Nandhibatla V. Sastry a,∗ , Nilesh M. Vaghela a , Vinod K. Aswal b a b

Department of Chemistry, Sardar Patel University, Vallabh Vidyanagar 388120, Gujarat, India Solid State Physics Division, Bhabha Atomic Research Center (BARC), Trombay, Mumbai 400085, India

a r t i c l e

i n f o

Article history: Received 2 August 2011 Received in revised form 12 April 2012 Accepted 19 April 2012 Available online 8 May 2012 Keywords: Surfactant-like ionic liquids Critical aggregation concentration Surface activity Thermodynamic parameters SANS

a b s t r a c t A series of surfactant-like ionic liquids (ILs), typically consisting of a long hydrocarbon tail and an ionic head group have been synthesized by a direct reaction of 1-methylimidazole and 1-choloroalkane, RCl, R = C10 , C12 , C14 , C16 and C18 , 1-methylpipridine or 1-methylpyrrolidine and 1-chlorooctadecane respectively. Surface activity and aggregation of these surfactants have been explored by surface tension and solution conductivity measurements. New results (critical aggregation concentrations (cac), and surface active parameters (at 298.15 K), thermodynamic parameters of aggregation (at 298.15, 303.15 and 313.15 K)) are reported. The increase in length of R decreased cac, minimum area/surfactant molecule at air/water interface and while the adsorption efficiency, surface excess concentrations, standard entropy of aggregation were increased indicating that the aggregation with in the temperature limits of present study is an entropy driven process. The analysis of the small angle neutron scattering (SANS) curves revealed that the aggregates are of oblate ellipsoidal shape and the aggregation numbers increased with the increase in the chain length (C10 –C18 ) of alkyl branch and therefore it is suggested that the longer the alkyl chains, parallel would be their alignment in the core part of the aggregates. Comparison of surface active parameters for the three ILs with a common octadecyl chain but different cationic head groups revealed that the methylimidazolium moiety is more effective than methylpiperidine and methylpyrrolidine at air/water interface. Similarly, the number of molecules in an aggregate was found to be more, when the cationic head group is made up of ␲ electron ring systems as compared to the one with point charge. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Ionic liquids (IL) consist entirely of ions and are emerging as an important class of compounds with unique properties such as negligible vapor pressure, wide liquidous range, excellent solvation, high thermal stability, ionic conductivity, selective catalytic behavior and wide electrochemical windows etc. [1–8]. The chemical structure of ionic liquids mimic that of amphiphilic cationic surfactants consisting of a non-polar hydrophobic tail and a polar cationic head group. Therefore based on structure activity relationship, a typical long alkyl chain IL could possess surface active properties similar to the conventional cationic surfactants and have tendency to associate in to nano micellar structures [9]. A long alkyl chain based IL, when dissolved in water thus exhibits amphiphilic character. Very little is known about the surface active and micellization features of such ILs and it is unclear as how the surfactant-like tendency of ILs depends upon their molecular characteristics, concentration,

∗ Corresponding author. Fax: +91 2692 234675. E-mail address: nvsastry [email protected] (N.V. Sastry). 0378-3812/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.fluid.2012.04.013

temperature and additives etc. In view of this, the study on the relationship between the molecular structural features and microstructural characteristics of micellar associates of ILs in aqueous solutions is always an interesting and an important subject area of investigation in the discipline of colloid and interface science. The understanding of the solution behavior of surfactant-like ionic liquids in aqueous media is highly necessary to invent and propose very useful and still yet fully unexplored applications such as wetting, dispersion, emulsification, detergency, solubilization, catalysis of chemical reactions, etc. ILs have already been emerging as one of the alternatives to volatile organic solvents because of their negligible vapor pressure. Similarly, ILs can be potential green surface active agents especially to replace the cationic surfactants of alkyltrimethylammonium bromide type that are widely used in hair conditioning products and reported to be toxic to micro-organisms inhabiting activated sludge systems [10,11]. A proper and comprehensive understanding, especially of the adsorption behavior at air/water interface, and aggregation or association phenomenon in aqueous and non-aqueous solutions is a prerequisite before ILs are explored for surfactant based applications [9]. The surface, interfacial and aggregation characteristics of ILs based

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29

on 1-alkyl (butyl to hexadecyl)-3-methylimidazolium bromide, or chloride in water have been extensively studied by using various methods such as surface tension [12,13,16,21–23,25–27], electrical conductivity [14–25,31,32], fluorescence [16,20,21,25,26], 1 H NMR [19,26,28], steady state fluorescence spectroscopy [25,29], isothermal titration calorimetry [30]. The main focus of these investigations were to determine the critical aggregation concentrations (CAC) mostly at 298.15 K. Some discrepancies between the CAC data as determined from different methods for the same IL were noted [31] and these discrepancies can be attributed to the presence of impurities of the sample as well as to the sensitivity of the method used. The CAC values of ILs with bromide counter anion were found to be lower than that of with chloride counter anion. Bromide ions, being larger have weaker hydration and hence are easily adsorbed on the surface of the aggregates and effectively decrease the electrostatic repulsions between the head groups which eventually facilitate the onset of aggregation at much lower concentrations. Luczak et al. [9] and Smirnova et al. [31] have collected the literature CAC data of homologous ILs based on alkylmethylimidazolium of type [Cn mim][Cl] or [Cn mim][Br] and examined the dependence of their CAC on the alkyl chain length. It was observed that the variation of log CAC values with the alkyl chain can be fitted to an empirical Stauff–Klevens rule, log CAC = A − Bx, where A and B are constants at a given temperature and x is the number of carbon atoms in the alkyl chain. The CAC values of ILs otherwise were found to lie in between the values of molecularly comparable cationic and anionic surfactants. As compared to extensive data on the CAC, there exists very little information on the surface active and morphological features such as size and the shape of the aggregates formed by homologous ILs in general and their variation with the molecular characteristics, i.e. alkyl chain length, type of ring cation, nature of anion and solution conditions such as concentration of IL, temperature and presence of additives etc. in particular. Our laboratory has recently reported the surface active features and aggregation characteristics of surfactant-like ionic liquids based on 1-alkyl-3-methylimidazolium or 1-alkyl-1methylpyridinium halides of type. [Cn mim][X] or [Cn py][X], where n = 4, 6 or 8 and X = Cl− , Br− and I− in water as well as in aqueous solutions containing sodium chloride, sodium sulfate and 2-butoxyethanol as additives, by employing surface tension, electrical conductance, 1 H NMR and SANS methods [32–34]. Even though the studied ILs displayed characteristic CAC, the analysis of the results especially from the latter two methods revealed that true micelle-like aggregates are not evident in ILs when n = 4. 1-Octyl-3-methylimidazolium or -pyridinium chloride aqueous solutions do contain aggregates but the structural organization is far different from conventional micelles formed by amphoteric surfactant substances. The ionic liquids with hexyl chains are found to be of transitory type with no detectable aggregate structures. This prompted us to take up systematic investigations with a focus mainly on examining the surface activity and aggregate formation of ionic liquids based on 1-alkyl-3-methylimidazolium halides of type, [Cn mim][Cl], where n = 10, 12, 14, 16 and 18, 1-octadecyl-1-methylpiperidinium chloride, [C18 mpip][Cl] and 1octadecyl-1-methylpyrrolidinium chloride, [C18 mpyrro][Cl] and access the effect of chain length and head group nature on (i) the critical aggregation concentration, (ii) the surface active properties such as the effectiveness of surface tension reduction,  CAC , adsorption efficiency, pC20 , surface excess,  max and minimum area per ionic liquid molecule, as1 , (iii) thermodynamic parameters of aggregation and (iv) size and shape of the aggregate structures, by surface tension, electrical conductance and SANS measurements. SANS profiles at different concentrations for a given ionic liquid were measured and analyzed through model fits to ascertain the effect of concentration on the geometrical features of the aggregates. To the best of our knowledge, this is the first report that

23

describes the geometrical features for homologous series of ILs of type, [Cn mim][Cl] where n ≥ 10 in water. 2. Experimental 2.1. Reagents Acetonitrile (GR reagent grade, Merck 99.5%) was used as such without any further purification. Ethyl acetate (Merck, ≥98%) was dried over fused calcium chloride and distilled and the fraction boiling in the temperature range of 349.15–350.15 K was collected. 1-methylimidazole (Alfa Aesar, 99%), 1-methylpyrrolidine (Fluka, ≥99%), and 1-methylpiperidine (Merck, ≥98%), were dried over potassium hydroxide and freshly distilled. The liquids were fractionally distilled. 1-Chlorodecane (≥98%), 1-chlorododecane (95%), 1-chlorohexadecane (96%) and 1-chlorooctadecane (98%) were of Merck make (with the purities shown in the parenthesis) and used as received. The water was triple distilled over potassium permanganate and the samples with a conductivity of 1.2 × 10−6 S/cm were used in the solution preparation. 2.2. Synthesis and characterization of ionic liquids The ionic liquids were prepared by carrying out a direct reaction between 1-methylimidazole, or 1-methylpyrrolidine, or 1-methylpiperidine with an excess amount of respective 1chloroalkane in acetonitrile under reflux at 353.15 K for 48 h. The reactions were conducted under nitrogen blanket in a round bottom flask equipped with reflux condenser and magnetic stirrer. The resulting product in its solution form was cooled. The product was crystallized from ethylacetate under vigorous stirring in the temperature range of 273.2–278.2 K depending upon the reactants. The crystallized products were further dissolved in acetonitrile and recrystallize them once again. This process was repeated for four times to remove the un-reacted reagents. The solvent residues were finally removed by placing the product solution under vacuum at 343.2 K over a period of 12 h. The structures of the products were confirmed by 1 H and 13 C NMR spectra in CDCl3 . The melting points of the ILs were obtained from differential scanning calorimetry (DSC) traces. The results of 1 H and 13 C NMR spectra are summarized in Table S1 1-Decyl-3-methylimidazolium chloride: [C10 mim][Cl]: colorless liquid: yield, 90%. 1-Dodecyl-3-methylimidazolium chloride: [C12 mim][Cl]: white waxy solid: yield, 91%: m.pt. 317.2 K. 1-Hexadecyl-3-methylimidazolium chloride: [C16 mim][Cl], white waxy solid: yield, 91%: m.pt. 332.8 K. 1-Octadecyl-3-methylimidazolium chloride: white waxy solid: yield, 95%: [C18 mim][Cl], m.pt. 343.1 K. 1-Octadecyl-1-methylpyrrolidnium chloride: white waxy solid: [C18 mpyrro][Cl], m.pt. 324.1 K. 1-Octadecyl-1-methylpiperdinium chloride: white waxy solid: [C18 mpip][Cl], m.pt. 365.2 K. The water content of ILs was determined by using Karl Fischer titrator and the results showed that mass fractions of water ranged from 5.3 × 10−6 to 8.2 × 10−6 . The purity of the ionic liquids was adjudged by potentiometric titration using argentometric method in which 10 cm3 of IL aliquot solution was titrated with a sodium

24

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29

nitrate solution (standardized against a 0.015 mol dm−3 sodium chloride) using silver chromate as indicator. The estimated purities of the samples were ≥97–98% on mole basis.

The flow times were in the range of 130–360 s and hence no kinetic corrections were invoked. The flow volume was greater than 5 cm3 making drainage corrections unimportant. 3. Results

2.3. Methods 3.1. Critical aggregation concentration (CAC) The stock solutions of ILs in water were made by mass (±0.01 mg) by dissolving known weights by agitation under magnetic stirring over a period of 2 h at room temperature. The stock solutions were then diluted by four times fresh distilled water to desired concentrations. The surface tension measurements were performed using a Model DCAT 11 Surface Tensiometer (Dataphysics, Germany) equipped with a Wilhemy plate. The plate was dipped in triple distilled water and a stock IL solution was dispensed into the measuring bowl by a liquid dispensing unit. The desired dilutions were pre-programmed to get automatically the final concentrations. Therefore the plate could not be washed in between two measurements for the given system. However once the measurements were completed for a given system, the plate was thoroughly clean washed with triple distilled water repeatedly and heated red hot on a Bunsen flame to remove the adsorbed solutes, if any from its surface. The adsorption effects on the plate surface due to its contact with high IL concentration were minimized by this procedure. The temperature (accurate to ±0.05 K) during the measurements was maintained by circulating water through inlet and outlet jacketed chutes provided on the base using a INSREF make (India) circulator. The uncertainty in the measured surface tensions was within ±0.03 mN m−1 . The electrical conductivities were measured by a digital conductivity meter (Equiptronics, India) using a cell with platinum electrodes. The cell was calibrated with 0.01 and 1.0 mol dm−3 potassium chloride aqueous solutions. At least three measurements were made every time and the specific conductivity was calculated from the mean value. The uncertainty in conductivity measurements was less than ±0.3%. The cell assembly was dipped in glass vials kept in water bath thermostatted at the measuring temperature maintained to ±0.05 K. The SANS experiments were carried out using SANS diffractometer at the DHRUVA reactor facility, BARC, Trombay, India. The ˚ The diffractometer uses neutrons of mean wave length,  of 5.2 A. momentum transfer or accessible wave transfer Q (=4 sin 0.5/, where  is the wave length of the incident neutrons and  is the scattering angle, i.e. angle between the transmitted and scattered beam) range of this instrument is between 0.018 and 0.35 A˚ −1 . The angular distribution of the scattered neutron is recorded using a one-dimensional position sensitive detector located at a distance of 1.8 m from the sample. The diffractometer makes use of a BeO filtered beam, has a resolution (Q/Q) of about 30% at Q = 0.05 A˚ −1 . IL solutions were prepared by dissolving their known amount in D2 O by mass. The use of D2 O instead of H2 O (water) for preparing solutions provides a very good contrast between the aggregates and the solvent. The solutions were held in UV-grade quartz sample holders of 0.5 cm path length with tight-fitting teflon stoppers, sealed with parafilm. The measured raw scattering intensities of neutrons were corrected for the background, empty cell scattering and sample transmission. The intensities then were converted to differential cross-section units, d˙/d˝. The d˙/d˝ were normalized to an absolute scale with reference to a scattering from a standard sample. The uncertainty in the measured scattering intensities is estimated to be 10%. The flow times of ionic liquid aqueous solutions were obtained by using two Ubbelohde suspended level viscometers placed in a thermostatic water bath maintained at constant temperature accurate to ±0.05 K. Three consecutive flow times agreeing within ±0.02 s were registered and the mean flow time was considered.

The CAC values of ionic liquids were determined by surface tension as well as electrical conductivity methods at T = 298.15, 303.15 and 313.15 K. The surface tension isotherms for a series of homologous ionic liquids at 298.15 K are given in Fig. S1 (supplementary material). The isotherms are found to be typical with an inflection point but surface tension values increased after this point at higher concentrations. Because of this, the isotherms appear to have a minima around the vicinity of aggregation onset. The specific conductivity isotherms of the same ionic liquids at 298.15, 303.15 and 313.15 K were constructed and representative isotherms at 298.15 K are shown in Fig. S2 (supplementary material). The profiles are typical with a clear inflection point. Phillips method [35] was used to locate the exact inflection points corresponding to CAC values. 3.2. Surface active parameters The surface active parameters namely the effectiveness of surface tension reduction,  CAC , i.e. surface tension at CAC, adsorption efficiency, pC20 , (pC20 = −log C20 , where C20 is the molar concentration required to reduce the surface tension of water by 20 mN m−1 ), surface excess,  max , ( max = − 1/2RT(∂/∂ln C)) and minimum area per ionic liquid molecule, as1 , (as1 = 1016 /NA max where NA = Avogadro number) at 298.15 K were calculated from Gibbs equation [23] and the same are summarized in Table 1. 3.3. Temperature dependence of CAC and thermodynamic parameters The CAC data at different temperatures are listed in Table 1 and Table S2 (supplementary material). CAC values slightly increased with the increase in the temperature. Applying a phase separation model for the monomer–aggregate equilibrium, the free energy of aggregation is related to CAC by, ◦

Ga = (1 + ˇ)RT ln xCAC

(1)

where xCAC is the critical aggregation concentration in mole fraction, ˇ is the degree of ionization, which represents the fraction of amphiphile ions in the aggregate neutralized by counter anions. The values of ˇ were calculated from the ratio of slopes of the two linear lines drawn through regions corresponding to below and above CAC. The summary of ˇ values is given in Table 2 and Table S2. The ◦ ◦ enthalpy and entropy of aggregation, Ha and Sa are calculated by the relations, ◦

Ha = (1 + ˇ)RT 2

d ln xCAC dT

(2)

and ◦

Sa =

Hao − Gao T

(3)

The summary of the thermodynamic parameters of aggregation is also given in Table 2 and Table S2. 3.4. SANS measurements SANS method is a powerful technique for getting structural information on the particles with characteristic size range of

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29

25

Table 1 Critical aggregation concentration, CAC (from surface tension isotherms); effectiveness of surface tension reduction,  CAC ; surface pressure, ˘ CAC ; adsorption efficiency, pC20 ; surface excess,  max and minimum area for adsorbed molecule, as1 for ionic liquids in water at 298.15 K. IL

CAC (mmol dm−3 )

 CAC (mN m−1 )

˘ CAC (mN m−1 )

pC20

 max (×1010 mol cm−2 )

as1 (A´˚ 2 )

[C10 mim][Cl]

40.00 ± 1.55 39.90 [16] 53.80 [23] 55.00 [26] 14.80 ± 1.00 15.00 [26] 13.20 [16] 1.14 ± 0.05 1.14 [23] 0.89 [44] 0.87 [10] 0.40 ± 0.02 0.45 [23] 0.42 ± 0.02 0.45 ± 0.02

27.3

44.5

2.5

1.9 ± 0.1

85 ± 1 85 [16]

38.7

33.6

2.4

2.3 ± 0.1

72 ± 1 72 [16]

37.0

34.8

3. 2

3.4 ± 0.1

49 ± 1 49 [16]

42.0

29.8

3.6

3.7 ± 0.1

45 ± 1

36.5 37.7

35.3 34.1

3.7 3.8

3.5 ± 0.1 3.3 ± 0.1

48 ± 1 50 ± 1

[C12 mim][Cl]

[C16 mim][Cl]

[C18 mim][Cl] [C18 mpyrro][Cl] [C18 mpip][Cl]

Standard uncertainty in  CAC , ˘ CAC ,  max and as1 are ±0.03 mN m−1 , ±0.03 mN m−1 , ±0.1 × 10−10 mol cm−2 and ±1 A´˚ 2 respectively. Table 2 ◦ ◦ ◦ The CACs (from conductivity isotherms), degree of ionization, ˇ standard molar Gibbs free energy, Ga enthalpy, Ha and entropy of aggregation Sa for long chain ionic liquids in water at 298.15 K. ◦

◦

◦

IL

CAC (mmol dm−3 )

ˇ

Ga (kJ mol−1 )

Ha (kJ mol−1 )

[C10 mim][Cl]

53.81 ± 2.58 59.90 [23] 40.47 [16] 19.96 ± 1.97 14.00 [16] 1.25 ± 0.08 1.26 [23] 0.86 [16] 0.44 ± 0.02 0.40 [23] 0.42 ± 0.04 0.45 ± 0.03

0.490 0.490 [23] 0.74 [16] 0.560 0.77 [16] 0.450 0.450 [23] 0.81 [16] 0.399 0.400 [23] 0.472 0.435

−25.61 ± 0.16 −25.22 [23] −31.14 [16] −30.65 ± 0.17 −36.51 [16] −38.45 ± 0.14 −38.42 [23] −49.68 [16] −40.72 ± 0.15 −41.07 [23] −43.01 ± 0.17 −41.69 ± 0.15

−1.70 ± 0.05

80.2 ± 0.8

−2.62 ± 0.03

94.0 ± 0.5

−7.15 ± 0.03

105.0 ± 0.5

−7.50 ± 0.05

111.4 ± 0.5

−8.16 ± 0.06 −7.67 ± 0.06

116.9 ± 0.6 114.1 ± 0.6

[C12 mim][Cl] [C16 mim][Cl]

[C18 mim][Cl] [C18 mpyrro][Cl] [C18 mpip][Cl]

◦

◦

Sa (J mol−1 K−1 )

◦

Standard uncertainty in CAC, Ga , Ha , and Sa are ±0.01 mg, ±0.17 kJ mol−1 , ±0.06 kJ mol−1 , ±0.8 J mol−1 K−1 respectively.

3.4.1. Analysis of SANS data The SANS intensities for monodisperse micelles of volume (Vm ) and of uniform scattering length density ( m ) is given by the modified Hayter–Penfold approach as, d˙ 2 ( m − s )2 [F 2 (Q ) + F(Q )2 (S(Q ) − 1)] = nm Vm d˝

(4)

12

10

8 -1

dΣ/dΩ (cm )

10–150 A˚ in general and micellar aggregates in particular [36,37]. SANS measurements were made in solutions of ionic liquids in D2 O over a wide concentration range at 303.15 K to obtain information on the effect of concentration on the shape and size of the micelle like aggregates in D2 O solutions. Sufficient care was taken to avoid high IL concentrations to ensure that the solutions are not highly viscous and are free of air bubbles during the measurements. The representative plots of coherent differential scattering crosssection, d˙/d˝ (also known as intensity) versus scattering wave vector, Q for 100 mM solutions of each of the six ILs are shown in Fig. 1, while the SANS curves for the other concentrations of ILs are depicted in Fig. S3 (supplementary material). The SANS curves displayed well defined correlation peaks and the nature of the curves is typical of self associating systems. The presence of sharp correlation peaks is attributable to the strong repulsive interaction between the positive charges on the head groups of the ring. The SANS curves were subjected to the modified Hayter–Penfold type analysis as applied to the colloidal solutions consisting of particles with an ellipsoid shape [38–43].

6

4

2

0 0.0

0.1

0.2

0.3

0.4

Q (Å ) -1

Fig. 1. SANS distributions and model fits for solutions of the long alkyl chain ionic liquids in D2 O for 100 mmol dm−3 concentration at 303.15 K; (+) [C10 mim][Cl], () [C12 mim][Cl], () [C16 mim][Cl], () [C18 mim][Cl], () [C18 mpyrro][Cl], and (×) [C18 mpip][Cl]; symbols are experimental data points and lines are fitted curves.

26

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29

where d˙/d˝ is the coherent differential scattering cross section for a amphiphile solution in a solvent (D2 O) of scattering length density s . F(Q) is the single particle form factor and S(Q) is the inter particle structure factor. The aggregation number (N) of the micelle is related to the micellar volume through the monomer volume. For an ellipsoidal micelle, P(Q) is given as



1

[F(Q, )]2 d

P(Q ) = F 2 (Q ) =

(5)

0

where is the cosine of the angle between the axis of revolution A and the wave-vector transfer (Q). The F(Q, ) for an ellipsoidal micelle with semiminor and semimajor axes, a and b, respectively, is given as F(Q, ) =

3(sin ω − ω cos ω) ω3

(6)

where ω = Q[a2 2 + b2 (1 − 2 )]1/2 .The S(Q), which specifies the correlation between the centers of different micelles, is the Fourier transform of the radial distribution function g(r) for the mass centers of the micelles. For an isotropic system, S(Q) is expressed as [42],



S(Q ) = 1 + C(Q )nm where C(Q ) = P(Q )−1

exp(iQ · r)[g(r) − 1]dr



1 0

(7)

2 [F(Q, )]d

.

The factor C(Q) tends to be unity as Q approaches zero. The analysis of the scattering cross section thus reduces to the calculation of g(r). The S(Q) for an ellipsoidal model using the mean spherical approximation [40] assuming it to be a sphere of effective diameter = 2(a2 b)1/3 and interacting through a screened coulomb potential is, u(r) =

u0 r exp[−k(r − )]

r>

(8)

where is the equivalent particle diameter, k is the Debye–Huckel inverse screening length (which depends on the critical aggregation concentration and effective charge z on the aggregate), and u0 is the contact potential. The fractional charge on the micelle ˛ (=z/N) is an additional parameter in the calculation of S(Q). For the quantitative analysis of the SANS data, one calculates the cross sections using Eq. (4) and compare the same with the experimental distributions. In the analytical program, the scattering length and volume of hydrophobic part are fed along with the trial input fitting parameters of semi major, minor axes and fractional charge. This procedure limits the number of fitting parameters and hence provides better statistics. A number of models were tried to fit the SANS data. Oblate core–shell model F(Q) in combination with the Hayter–Penfold charged sphere S(Q) were found to give best fits. The key parameters extracted from the analysis are summarized in Table 3. It is worth mentioning that the SANS intensities for the solutions of short alkyl chain based ILs namely 1-hexyl-, or, 1-octyl3-methylimidazolium or pyridinium chloride could not be fitted to any well meaning model fits that assume a definite shape and size for the aggregates [32]. Not only this, as compared to the sharp peaks observed in the SANS spectra of Fig. 1, we could observe only broad and diffused peaks for the short chain ionic liquid solutions. 3.5. Dilute solution viscosities The viscosity parameters provide a simple guide about the aggregate growth. We calculated the relative and reduced viscosities of the ionic liquid aqueous solutions within the limits of concentrations corresponding to SANS measurements. The dependence of relative viscosity and reduced viscosity on the

concentration of ILs is shown in Figs. S4 and S5 (supplementary material). Relative viscosities increased gradually not only with the concentration for a given ionic liquid but also with the increase in the carbon chain length in the alkyl part of the IL for a given concentration. Among the three 1-octadecyl chain based ILs, the decrease in the relative viscosity followed the order: mpip < mpyrro < mim and this trend is similar to the one observed in the aggregation numbers of respective ILs. The shape of the reduced viscosity versus concentration curves for ionic liquid aqueous solutions is found to be typical of polyelectrolyte particles in solution. The observed sharp increase of reduced viscosities in the low concentration region can be attributed to the maximum probable repulsive interactions among IL molecules. Longer the length of the alkyl chain, more was the reduced viscosity values indicating a close relation between the reduced viscosity and the aggregation number.

4. Discussion 4.1. Surface tension isotherms The appearance of a minima in surface tension isotherms of classical ionic or nonionic surfactant aqueous solutions is often attributed to the presence of impurities which get adsorbed at the air/water interface around the onset of aggregate formation. The minima displayed in some of the isotherms shown in Fig. S1 persisted even after using the IL samples obtained by repeated crystallization or from independent batches of synthesis. Therefore it can be reasonably claimed that the observed minima is not due to the presence of any impurity in the samples. The appearance of such minima in surface tension isotherms has also been reported for 1-octyl-3-methylimidazolium chloride, [C8 mim][Cl] [12], nalkyl-3-methylimidazolium bromides, [Cn mim][Br], where n = 4, 6, 8, 10 [13], 1-decyl-3-methylimidazlium bromide, [C10 mim][Br] [15] and n-alkyl-3-methylimidazolium halides, [Cn mim][X], where n = 4, 6, 8, X = Cl− , Br− and I− [32]. Goodchild et al. [13] had examined in detail the surface structure of 1-octyl-3-methylimidazolium bromide molecules in the minima region of the surface tension isotherms using neutron reflectivity measurements. It has been found that the appearance of minima in surface tension values coincided with a minimum of average molecular area and maximum of thickness of adsorbed surface film and therefore the authors have suggested that the structure at the surface is more complex and may be it deviates from a simple monolayer around CAC. The increase of surface tension values at higher concentrations were attributable to the depletion of solute molecules at the air/water interface as a result of their tendency to get aggregated in the bulk rather than increased surface excess of water [15]. The minimum in the surface tension isotherms got narrowed and disappeared for 1-octadecyl-3-methylimidazolium and 1-octadecyl-1-methylpyrrolidinium chlorides aqueous solutions. It is also possible that a reverse pull away from the surface may be exerted by the long alkyl chain such that the cationic head group orients properly at the air/water interface. However a deep minima is clearly visible for the aqueous solutions of 1-octadecyl1-methylpiperidinium chloride. With out any further speculation, we state here that the surface structure of long alkyl chain ILs seems to be complex and depends on the nature of both the cationic head group and the anion. The precise location of the CAC from the isotherms with a minima, by simple intersection method is difficult and therefore we analyzed them using Phillips [35] definition of CAC, as a point corresponding to the maximum change in the gradient of any physical property of the solution against concentration. Such maximum value of concentration corresponding to CAC is indicated by arrow marks. The CAC values of ILs as determined

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29

27

Table 3 Semi major axis, b, and semi minor axis, a, fractional charge, ˛ of aggregates corresponding to 100 mmol dm−3 of ionic liquids solution in D2 O at 303.15 K. IL

Semi major axis, b (Å)

[C10 mim][Cl] [C12 mim][Cl] [C16 mim][Cl] [C18 mim][Cl] [C18 mpyrro][Cl] [C18 mpip][Cl]

33.4 29.6 61.4 58.1 51.8 48.9

± ± ± ± ± ±

2.2 1.5 2.7 3.3 3.5 3.3

Semi minor axis, a (Å) 12.8 13.3 20.4 21.4 22.1 20.5

± ± ± ± ± ±

0.5 0.2 0.4 0.5 0.5 0.4

Fractional charge 0.17 0.34 0.13 0.09 0.09 0.17

± ± ± ± ± ±

0.01 0.06 0.06 0.06 0.02 0.06

Standard uncertainty in b and a is ±10%.

by both surface tension and electrical conductivity isotherms at different temperatures are also compared with the available literature values of the respective compounds. A perusal of column 2 of Tables 1 and 2 shows that the CAC values for lower alkyl chain (C10 and C12 ) homologue ILs as determined by surface tension isotherms and electrical conductivity measurements differ from each other. Not only this, the CAC values for such ILs determined by the same method also deviate from one research group to another. Our CAC value of 40 mM for [C10 mim][Cl] even though matched well with the literature value of 39.90 [16] but it deviates by ∼13.8–15.0 mM with the other reported literature data of 53.8 [22] and 55.0 [26]. We believe that the variation in CAC values for the same IL as determined by the same method but by different authors may either result from the presence of impurities in the samples or can be attributed to the inaccurate location of the inflection point especially in surface tension isotherms displaying a minima. It is of course difficult to emphasize which of the methods gives accurate CAC values. However it is reasonable to state that electrical conductivity measurements may give more accurate values when the surface tension isotherms for the same IL aqueous solutions show a minima around the onset of aggregation. When no such minima are not recorded, the CAC values for The ILs from both the methods may be close to each other and are therefore complimentary. 4.2. CAC and its dependence on the alkyl chain length CAC decreased with the increase in the chain length of alkyl branch and such systematic decrease could be due to enhanced hydrophobic character of higher alkyl chains. Applying the Stauff–Klevens rule to the CAC data of homologous series of imidazolium based ionic liquids, collected from the present as well as previous work [32], we calculated the values of A and B and the same were found to be 4.348 and B = 0.267 and these values are close to 4.51 and 0.28 reported by Luczak et al. [9] for similar ionic liquids. The value of the constants A and B depends upon the nature and number of hydrophilic groups and the effect each additional methylene group on the CAC respectively. In fact, the value of the constant B represents the free energy of transfer of a methylene group from aqueous back ground to the micellar interior. For the alkyl chain based single head group ionic surfactants, B value is close to 0.28–0.30. Therefore it is suggested that the aggregation process of long alkyl chain ILs in water is needed to be close to that of the micellization process of classical ionic surfactants in aqueous media. 4.3. Surface active parameters and nature of cationic ring A perusal of  CAC and pC20 and  max values of Table 1 clearly shows that effectiveness of ILs increases with the chain length of the alkyl branch. By comparing the data for the three ionic liquids with same alkyl chain (octadecyl) and common chloride counter ion but different cationic rings, it was found that methylimidazolium ring is more effective than the methylpiperidine and methylpyrrolidone rings even though the three of them have more or less same efficiency. The minimum area per adsorbed ionic liquid

at air/water interface, as1 decreases with the length of the alkyl chain and it implies that longer the alkyl chains, more compact would be the monolayers at air/water interface. Structurally, [C18 mpip][Cl] is more hydrophobic than [C18 mpyrro][Cl] and [C18 mim][Cl] and hence has slightly higher as1 values. 4.4. Temperature dependence of CAC and thermodynamic parameters of aggregation The degree of ionization, ˇ and the thermodynamic parameters of aggregation for ILs in water were calculated at three different temperatures of 298.15, 303.15 and 313.15 K. The data at T = 298.15 K are listed in Table 2 and the same at the other two temperatures is given Table S2 (supplementary material). The perusal of the data from Table 2 shows that the ˇ value increased with the increase in the alkyl chain length (from decyl to dodecyl) and however it showed a decrease with the increase in the alkyl chain length (from dodoecyl to octadecyl). The small value of ˇ in general indicates that the aggregate structures are compact for 1-octadecyl-3-methylimidazolium chloride in water. The free energy of aggregation in general for the six ionic liquids in water is negative and expectedly became more negative with the increase in the number of carbon atoms in the alkyl chain. Such trend is in consistent with the fact that the higher alkyl chains facil◦ itate the spontaneous aggregate formation. Ha is also negative and increased with the carbon chain length among the four mim based ILs studied in the present work. Within the temperature lim◦ its of 298.15–313.15 K, for a given IL, Ha become more negative with the increase in the temperature. This implies that the aggregate formation remains mainly exothermic in the above mentioned ◦ temperature limit. −T Sa term within the temperature limits of the present study systematically become more negative with the increase in carbon chain length at a given temperature or for a given IL with the increase in the temperature. These characteristic changes indicate that the aggregation process for the ILs studied in the present case within the temperature limit of 2915–313.15 K is dominated by positive entropy values. 4.5. Micro-structural features of IL aggregates SANS analysis based on model fits described in the previous section revealed that the aggregates formed by the ILs studied in the present investigations are of prolate ellipsoidal shaped structures with a core part made up of hydrocarbon tails while the outer shell consist mainly of the hydrophilic head groups along with the bound water molecules. A perusal of the various aggregate parameters as listed in Table 3 and Table S3 (supplementary material) reveals that the axial ratios did not show much variation either with the change in the concentration for a given IL or with the increase in the length of the alkyl chain at a given concentration. We believe that these results are first of its kind and clearly establish that the prolate shaped micellar-like aggregates of ILs are stable in wide concentration range. This is in strong contrast to the aggregates of short alkyl chain based IL namely 1-octyl-3-methylimidazolium chloride [12,33], 1-octyl-3-methylimidazolium iodide [12] and

28

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29

60

the ring system. As discussed previously, the rings in methylimidazolium and methylpyrrolidinium are five membered but the positive charge is delocalized in the former and exists as a point charge in the latter. Therefore head group. . .head group repulsions are more prevalent among methylpyrrolidinium molecules and such repulsive interactions would cause steric hindrance in forming the aggregate structures. Therefore the aggregation number for such structures would be low. In the methylpiperidinium case, the ring is six membered with a point cationic charge on it. So it is bulkier than other two rings and hence the outer shell of the aggregates can accommodate less number of such voluminous molecules and hence the calculated lowest aggregation number for this IL are justified. Thus in conclusion, it can be stated that both the length of the alkyl branch, the nature and size of the cationic ring equally play a decisive role in controlling the number of molecules in an aggregate.

40

5. Conclusions

200 180 160

N

140 120 100 80

0

100

200

300

400

-3

C ( mmol dm ) Fig. 2. The concentration dependence of aggregation numbers for the long alkyl chain ionic liquids in D2 O at 303.15 K; () [C10 mim][Cl], () [C12 mim][Cl], () [C16 mim][Cl], () [C18 mim][Cl], () [C18 mpyrro][Cl], and () [C18 mpip][Cl].

1-octyl- or 1-decyl-3-methylimidazolium bromides [13] in which the increase in the concentration resulted in to the formation of structures consisting of the individual molecules organized in to a cubic type packing. Note that the shape of the SANS curves within the concentration regime of this study is almost same and only intensities increased at high concentrations. This indicates that more and more IL molecules get into the favorable hydrophobic domains to avoid the unfavorable interactions as the concentration of ILs is increased progressively. Also, longer the alkyl chains, parallel would be their alignment in the aggregates. The effect of concentration and alkyl chain length on the aggregation numbers is depicted in Fig. 2. It is expected that the conformation of the alkyl chains within the confines of the aggregate structures plays very important role in their packing and hence aggregation number. A recent molecular dynamics study [45] on the aqueous solutions of 1-alkyl (decyl, dodecyl, tetradecyl and hexadecyl)-3methylimidazolium bromides carefully examined the structural nature of the aggregates for the initial stages of formation by analyzing the conformational states of alkyl chains. The noted trends in our experimental data are close to the results of this independent computational study, which assigns a quasi-spherical shape to the aggregates and most importantly, it was found that the arrangement of cations with decyl- and dodecyl chains was random, while with the tetradecyl tails displayed a small degree of order, and the organic cations with the hexadecyl chains showed a large of crystalline domains formed by inter-digitized or parallel alkyl chains within the interior core of the aggregates. In either of the cases, the aggregation numbers ought to be more which is the case in the present study. 4.6. Effect of head group For the same alkyl chain length (C18 ), the size of the prolate shaped aggregates was almost same for both of the ILs with methylpyrrolidinium and methylpiperidinium cationic head groups, but were smaller as compared to the methylimidazolium based system. The aggregation number for comparable concentrations followed the order: mim > mpyrro > mpip. This trend is rather not totally unexpected and can be explained by considering both the ring structure and the nature of charge localization with in

The surface activity and aggregation characteristics of ILs of type [Cn mim][Cl] where n = 10–18, are highly dependent on the chain length of the alkyl branch and type of the head group. CAC values systematically decrease with the increase in the alkyl chain length and such dependence is analogous with the conventional ionic surfactants. The effectiveness of surface tension reduction and efficiency of adsorption increase and become better with the increase in the alkyl chain length. The minimum area per adsorbed molecule on the other hand decreases sharply with the alkyl chain length indicating that compact mono layers are formed at air/water interface on one hand and facilitate the formation of well defined aggregates in the bulk on the other, when the alkyl chains are longer. The aggregation process of high alkyl chain ILs (within the temperature limits of 298.15–313.15 K) is entropy driven. The SANS results reported in the present work established a clear distinction between the nature of the aggregates of short alkyl chain (≤C8 ) and long alkyl chain (≥C10 ) based 1-alkyl-3-methylimidazolium chlorides in water. The long chain ILs form oblate ellipsoidal shape aggregate structures similar to the micelles of conventional surfactants and increase in the concentration did not result in any appreciable change in the size and does not perturb the shape of the aggregates. This is in contrast to the concentration dependent structural features of the aggregates formed by short alkyl chain ILs wherein increasing concentration resulted in the increase of size and formation of elongated structures [12,13,32]. The increase in the alkyl chain length leads to the parallel packing of chains in the interior of aggregates and hence accommodate more IL molecules in a given volume, i.e. increases the aggregation number. The cationic rings with delocalized ␲-electron clouds favor high aggregation number and while rings with point like cationic charge cause more head group. . .head group repulsions and decrease the aggregation number. The bulkier the head group, the lower would be the aggregation number. The aggregates have considerable fractional charge and exhibit polyelectrolytic behavior at low IL concentrations. Therefore, similar to classical cationic surfactants, ionic liquids with longer alkyl chains are equally surface active at air/water interface and form well defined aggregates in bulk solution with distinct hydrophobic and hydrophilic regions and have a potential for the development of novel green surface active agents. List of symbols effectiveness of surface tension reduction  CAC pC20 adsorption efficiency molar concentration required to reduce the surface tenC20 sion of water by 20 mN m−1  max surface excess minimum area per ionic liquid molecule as1

N.V. Sastry et al. / Fluid Phase Equilibria 327 (2012) 22–29 ◦

Ga ◦ Ha ◦ Sa ˇ NA R T xCAC Q   d˙/d˝ Vm s m F(Q) S(Q) N

a b g(r) ˛

free energy of aggregation enthalpy of aggregation entropy of aggregation degree of ionization Avogadro number gas constant absolute temperature critical aggregation concentration in mole fraction unit accessible wave transfer wave length of the incident neutrons scattering angle cross-section volume of monodispersed micellellar aggregates scattering length density of solvent number density of the micelles or aggregates single particle form factor inter particle structure factor aggregation number cosine of the angle between the axis of revolution semiminor axes semimajor axes radial distribution function effective diameter fractional charge on the micelle

Acknowledgment The authors thank UGC-DAE Consortium for Scientific Research, Mumbai Center for funding a Collaborative Research Scheme (CRS) under grant no. CSR/AO/MUM/CRS-M-127/08/448. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.fluid.2012.04.013. References [1] [2] [3] [4] [5]

M.J. Earle, K.R. Seddon, Pure Appl. Chem. 72 (2000) 1391–1398. P. Wasserscheid, R. Hal, A. Boesmann, Green Chem. (2002) 400–404. C. Lagrost, D. Carrie, M. Vaultier, R. Hapoit, J. Phys. Chem. A 107 (2003) 745–752. T. Welton, Chem. Rev. 99 (1999) 2071–2083. P. Wassersheid, P.T. Welton (Eds.), Ionic Liquids in Synthesis, Wiley-VCH, Weinheim, Germany, 2003. [6] J.G. Huddelston, H.D. Willauer, R.P. Swatloskai, A.E. Visser, R.D. Rogers, Chem. Commun. (1998) 1765–1766.

29

[7] J.D. Holbrey, K.R. Seddon, Clean Prod. Process. 1 (1999) 223–228. [8] M. Freemantle, Chem. Eng. News 76 (1998) 32–36. [9] J. Luczak, J. Hupka, J. Thoming, C. Jungnickel, Colloids Surf. A: Physicochem. Eng. Aspects 329 (2008) 125–133. [10] R. Matrgesin, F. Schinner, Int. Biodeterior. Biodegrad. 41 (1998) 139–143. [11] J. Hrenovic, T. Ivankovic, Central Eur. J. Biol. 2 (2007) 405–414. [12] J. Bowers, C.P. Butts, P.J. Martin, M.C. Vargara-Gutierrez, R.K. Heenan, Langmuir 20 (2004) 2191–2198. [13] I. Goodchild, L. Colier, S.L. Millar, I. Porkes, J.C.D. Lord, C.P. Butts, J. Bowers, J.R.P. Webster, R.K. Heenan, J. Colloid Interface Sci. 307 (2007) 455–468. [14] T. Inoue, H. Ebina, B. Dong, L. Zheng, J. Colloid Interface Sci. 314 (2007) 236–241. [15] A. Modaressi, H. Sifaoui, M. Mielcarz, U. Domanska, M. Rogalski, Colloids Surf. A: Physicochem. Eng. Aspects 302 (2007) 181–185. [16] O.A. El Seoud, P.A.R. Pires, T. Abdel-Moghny, E.L. Bastos, J. Colloid Interface Sci. 313 (2007) 296–304. [17] M. Blesic, A. Lopes, E. Melo, Z. Pterovski, N.V. Plechkova, J.N.C. Lopes, K.R. Seddon, L.P.N. Rebelo, J. Phys. Chem. B 112 (2008) 8645–8650. [18] G. Bai, M. Lopes, M. Bastos, J. Chem. Thermodyn. 40 (2008) 1509–1516. [19] T. Singh, A. Kumar, Colloids Surf. A: Physicochem. Eng. Aspects 318 (2008) 263–268. [20] J.J. Wang, H.Y. Wang, S.L. Zhang, H.H. Zhang, Y. Zhao, J. Phys. Chem. B 111 (2007) 6181–6188. [21] S.L.I. Toh, J. MacFarlane, C. Tsousis, D.W. De Paoli, H. Luo, S. Dai, Solvent Extr. Ion Exch. 24 (2006) 33–38. [22] R. Vanyur, L. Biczok, Z. Miskolczy, Colloids Surf. A: Physicochem. Eng. Aspects 299 (2007) 256–261. [23] C. Jungnickel, J. Luczak, J. Ranke, J.F. Fernandez, A. Muller, J. Thoming, Colloids Surf. A: Physicochem. Eng. Aspects 316 (2008) 278–288. [24] J. Siriex-Plenet, I. Gallion, P. Letellier, Talanta 63 (2004) 979–986. [25] B. Dong, X. Zhao, L. Zheng, J. Zhang, Na. Li, T. Inoue, Colloids Surf. A: Physicochem. Eng. Aspects 317 (2008) 666–672. [26] M. Blesic, M.H. Marques, N.V. Plechkova, K.R. Seddon, L.P.N. Rebelo, A. Lopes, Green Chem. 9 (2007) 481–490. [27] Q.Q. Baltazar, J. Chandawalla, K. Sawyer, J.L. Anderson, Colloids Surf. A: Physicochem. Eng. Aspects 302 (2008) 150–156. [28] Zhao, S. Gao, J. Wang, J. Tang, J. Phys. Chem. B 112 (2008) 2031–2039. [29] T. Singh, A. Kumar, J. Phys. Chem. B 111 (2007) 7843–7851. [30] F. Geng, J. Liu, L. Zheng, L. Yu, Z. Li, G. Li, C. Tung, J. Chem. Eng. Data 55 (2010) 147–151. [31] N.A. Smirnova, A.A. Vanin, E.A. Safanova, I.B. Pukinsky, Y.A. Anufrikov, A.L. Makarov, J. Colloid Interface Sci. 336 (2009) 793–802. [32] N.V. Sastry, N.M. Vahgela, V.K. Aswal, Colloid Polym. Sci. 289 (2011) 309–322. [33] N.V. Sastry, N.M. Vahgela, V.K. Aswal, Colloids Surf. A: Physicochem. Eng. Aspects 373 (2011) 101–109. [34] N.V. Sastry, N.M. Vaghela, P.M. Macwan, S.S. Soni, A. Gibaud, J. Colloid Interface Sci. 394 (2012) 52–61. [35] J.N. Phillips, Trans. Faraday Soc. 51 (1955) 561–569. [36] P.S. Goyal, Phase Trans. 50 (1994) 143–148. [37] V.K. Aswal, P.S. Goyal, Curr. Sci. 79 (2000) 947–953. [38] S.H. Chen, Annu. Rev. Phys. Chem. 37 (1986) 351–358. [39] S.H. Chen, T.L. Lin, Methods of Experimental Physics, vol. 23B, Academic Press, New York, 1987, pp. 489–495. [40] J.B. Hayter, J. Penfold, J. Colloid Polym. Sci. 261 (1983) 1022–1030. [41] J.B. Hayter, J. Penfold, J. Mol. Phys. 42 (1981) 109–118. [42] J.P. Hansen, J.B. Hayter, Mol. Phys. 46 (1982) 651–656. [43] J.B. Hayter, J.J. Penfold, J. Chem. Soc. Faraday Trans. 1 (77) (1981) 1851–1861. [44] S. Thomaier, W. Kunz, J. Mol. Liq. 130 (2007) 104–107. [45] B.L. Bharghava, M.L. Klein, J. Phys. Chem. B 113 (2009) 9499–9505.