Synthesis and micellar properties of surface-active ionic liquids: 1-Alkyl-3-methylimidazolium chlorides

Journal of Colloid and Interface Science 313 (2007) 296–304 www.elsevier.com/locate/jcis

Synthesis and micellar properties of surface-active ionic liquids: 1-Alkyl-3-methylimidazolium chlorides Omar A. El Seoud a,∗ , Paulo Augusto R. Pires a , Thanaa Abdel-Moghny b , Erick L. Bastos c a Instituto de Química, Universidade de São Paulo, C.P. 26077, 05513-970 São Paulo, SP, Brazil b Egyptian Petroleum Research Institute, Cairo, Egypt c Centro de Ciências Naturais e Humanas, Fundação Universidade Federal do ABC, Santo André, SP, Brazil

Received 1 February 2007; accepted 15 April 2007 Available online 20 April 2007

Abstract A series of surface-active ionic liquids, RMeImCl, has been synthesized by the reaction of purified 1-methylimidazole and 1-chloroalkanes, RCl, R = C10 , C12 , C14 , and C16 , respectively. Adsorption and aggregation of these surfactants in water have been studied by surface tension measurement. Additionally, solution conductivity, electromotive force, fluorescence quenching of micelle-solubilized pyrene, and static light scattering have been employed to investigate micelle formation. The following changes resulted from an increase in the length of R: an increase of micelle aggregation number; a decrease of: minimum area/surfactant molecule at solution/air interface; critical micelle concentration, and degree of counter-ion dissociation. Theoretically-calculated aggregation numbers and those based on quenching of pyrene are in good agreement. Gibbs free energies of adsorption at solution/air interface, G0ads , and micelle formation in water, G0mic , were calculated, and compared to those of three surfactant series, alkylpyridinium chlorides, RPyCl, alkylbenzyldimethylammonium chlorides, RBzMe2 Cl, and benzyl(3-acylaminoethyl)dimethylammonium chlorides, R AEtBzMe2 Cl, respectively. Contributions to the above-mentioned Gibbs free energies from surfactant methylene groups (in the hydrophobic tail) and the head-group were calculated. For RMeImCl, the former energy is similar to that of other cationic surfactants. The corresponding free energy contribution of the head-group to G0mic showed the following order: RPyCl ≈ RBzMe2 Cl > RMeImCl > R AEtBzMe2 Cl. The head-groups of the first two surfactant series are more hydrophobic than the imidazolium ring of RMeImCl, this should favor their aggregation. Micellization of RMeImCl, however, is driven by a relatively strong hydrogen-bonding between the chloride ion and the hydrogens in the imidazolium ring, in particular the relatively acidic H2. This interaction more than compensates for the relative hydrophilic character of the diazolium ring. As indicated by the corresponding G0mic , micellization of R AEtBzMe2 Cl is more favorable than that of RMeImCl because the –CONH– group of the former surfactant series forms hydrogen bonds to both the counter-ion and the neighboring molecules in the micelle. © 2007 Elsevier Inc. All rights reserved. Keywords: Ionic liquids, surface active; Ionic liquids, micellar properties of; Imidazole-based cationic surfactants, synthesis of; Surface tension; Conductance; ISE; Fluorescence

1. Introduction The amphiphilic nature of surfactants is responsible for their useful properties, hence applications: adsorption at interfaces (wetting, dispersion of solids, etc.), and aggregation in aqueous and non-aqueous solutions (detergency, emulsification, solubilization, preparation of nanoparticles, catalysis of chemical * Corresponding author.

E-mail address: [email protected] (O.A. El Seoud). 0021-9797/$ – see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2007.04.028

reactions, etc.) [1,2]. Therefore, the study of the relationship between structure of surfactants and their properties in solution will always be an important subject, both scientifically, and from the application point of view. The structural variables include the length of the hydrophobic tail, the nature of the counter-ion, and the nature/charge of the head-group. The latter has been much less studied than the former ones; this is interesting, because: Many of the above-mentioned applications reflect substrate-head-group interactions; Electrostatic interactions make a large positive contribution to the free energy of micellization [1,3–9]. Rates of

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

reactions that are sensitive to changes of solvent polarity (e.g., the spontaneous decarboxylation of the 6-nitrobenzisoxazole3-carboxylate anion) or to desolvation of the attacking nucleophile (e.g., elimination by the E2 pathway; SN 2; and acyl transfers) are enhanced by an increase of the hydrophobic character of the head-group, e.g., upon going from trimethyl- to tri-n-butylammonium. The reason advanced is that large headgroups partially exclude water from the interface; this results in a less polar reaction environment [7,10]. Cationic surfactants are amenable to this line of study because the structure of the head-group can be changed, e.g., by increasing the length of the alkyl groups attached, while maintaining constant the nature of the head-ion, e.g., quaternary ammonium. This is not the case for anionic surfactants where a change in the structure of the head-group entails a corresponding change in its chemical nature, carboxylate, sulfate, phosphate, etc. An additional reason for interest in studying cationic surfactants is that they show antibacterial properties and are used as hair conditioners, fabric softeners, antistatic agents, and corrosion inhibitors [11–14]. Ionic liquids, ILs, are low-melting electrolytes, thus forming liquids that consist only of cations and anions. The current convention is that a salt melting below the boiling point of water is known as an IL. Of these, the compounds based on imidazole, e.g., RMeImX have received much attention because they constitute “green” substituents to classic (volatile) organic solvents. Here R, Me, Im, and X stand for alkyl group, methyl, imidazolium, and counter-ion, respectively; for simplicity we do not show the charges. Note that (R) and (Me) are usually attached to positions 1 and 3 of the imidazolium ring; typical (X) include − − simple and complex ions, e.g., Br− , Cl− , BF− 4 , PF6 , (CN)2 N , etc. In addition to extremely low vapor pressure, ILs are highly polar, chemically inert, noninflammable and thermally stable. A myriad of chemical reactions can be carried out in ILs; residual reagents, reaction products and catalysts are separated by suitable procedures (e.g., removal under reduced pressure or extraction) and the IL, in principle, can be recycled into the process indefinitely [15–19]. RMeImX with a long-chain (R) are expected to be surface active ionic liquids, SAILs. Indeed, conductimetric, potentiometric, surface tension and volumetric (from solution density) studies have indicated that C10 MeImBr aggregates in water and, at low concentrations, behaves as a classic cationic surfactant [20,21]; surface tension measurements have indicated a similar behavior of C16 MeImCl and C16 MeImBF4 both in aqueous solution, and in the IL ethylammonium nitrate [22]. A few articles dealt with other physico-chemical properties of longchain RMeImX, including their partition coefficients between n-octanol and water (RMeImCl, R = C8 , C10 , and C12 , respectively) [23]; solubility of C12 MeImCl in several alcohols [24] and dipolar aprotic solvents [25]. Additionally, the hydrated crystals of RMeImCl, R = C12 –C18 , were postulated to have a double-layer structure with stabilization derived from direct, or water mediated hydrogen-bonding between Cl− and the hydrogens of the imidazolium moiety [26,27]. It is surprising that there has been no systematic study on the effects of the structure SAILs on the properties of their aqueous

297

solutions. This lack of information, coupled with their potential applications has prompted us to synthesize and investigate the properties of a series of SAILs, RMeImCl, R = C10 , C12 , C4 and C16 , respectively. Solution conductivity, surface tension, electromotive force, fluorescence of micelle-solubilized pyrene, and static light scattering have been employed to study their adsorption at solution/air interface and aggregation in aqueous solution. The data thus obtained was employed to calculate the following properties: Minimum area/surfactant at solution/air interface (Ainterface ); critical micelle concentrations (cmc); degrees of counter-ion dissociation (αmic ); aggregation numbers (Nagg ), as well as contribution of the surfactant discrete segments to Gibbs free energies of adsorption (G0ads ) and micellization (G0mic ). In order to assess the effect of surfactant head-group on its adsorption and aggregation in water, we compare data of the present series with those of cationic surfactants with different head-groups, pyridinium; benzyldimethylammonium, and benzyldimethylammonium that carries an amide group, vide infra. This comparison shows that micellization of cationic surfactants is affected by head-group hydrophobicity and, where operative, hydrogen bonding in the interfacial region. 2. Experimental 2.1. Materials The reagents were purchased from Acros and Apagão Química S.A. The solvents were purified as described elsewhere [28]; 1-methylimidazole was distillated from CaH2 ; 1-chloroalkanes (C10 , C12 , C14 , and C16 , respectively) were repeatedly fractionally distilled in a 50 cm long Vigreux column, under reduced pressure. Gas chromatographic analysis showed that 1-methyimidazole, 1-chlorodecane, and 1-chlorododecane are chromatographically pure; satisfactory purity, >99.5%, was obtained for the remaining 1-chloroalkanes. Pyrene and 2,3,6,7tetrahydro-9-(trifluoromethyl)-1H, 5H, 11H -[1]benzopyranpo [6,7,8-ij ]quinolizin-11-one (coumarine 153, here after designated as “C153”) were used as received. 2.2. Apparatus Melting points were determined with IA 6304 apparatus (Electrothermal). Gas chromatographic analyses were carried out on Shimadzu 17A-2 chromatograph, equipped with a FID detector and Supelcowax 10 capillary column. NMR spectra were recorded with Varian Innova 300 spectrometer; elemental analyses were performed on Perkin-Elmer 2400 CHN apparatus at the Elemental Analyses laboratory of this Institute. 2.3. Surfactants RMeImCl were prepared according to the general procedure, described elsewhere [29]. The reaction mixture was kept under dry, oxygen-free nitrogen atmosphere and was protected from light. To 1 mol of 1-methylimidazole, dissolved in 150 mL of

298

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

acetonitrile were added, under constant stirring 1.2 mol of the appropriate 1-chloroalkane. The resulting emulsion was continuously stirred for 48 h, at ca. 90 ◦ C; the medium turned homogeneous. The solution was cooled and transferred, via a filter cannula, to a vigorously-stirred, cold anhydrous ethyl acetate (300 mL, −20 ◦ C); the two-phase mixture was kept at this temperature for 2 h, followed by removal of the supernatant (by a syringe). This procedure was repeated until no more n-alkyl chloride was detected in the supernatant (by gas chromatography). The products were dried under reduced pressure at 40 ◦ C, over P4 O10 , for 24 h. Note: Elemental analyses, results of 1 H and 13 C NMR spectra are listed in the following tables of supplementary material (SM): SM-1, SM-2 and SM-3, respectively: 1-Decyl-3-methylimidazolium chloride: C10 MeImCl: Slightly yellowish liquid; yield, 91%; mp 17–19 ◦ C. 1-Dodecyl-3-methylimidazolium chloride: C12 MeImCl: White waxy solid; yield, 82%; mp 32–34 ◦ C. 1-Tetradecyl-3-methylimidazolium chloride: C14 MeImCl: White waxy solid; yield, 77%; mp 51–52 ◦ C, 1-Hexadecyl-3-methylimidazolium chloride: C16 MeImCl: White waxy solid; yield, 68%; mp 64–65 ◦ C, 2.4. Measurement of properties of surfactant solutions Glass double-distilled, de-ionized water was used throughout. Each surfactant was weighted, dried under reduced pressure, over P4 O10 until constant weight before making up the solution. 2.4.1. Surface tension Solution surface tension was measured at 25 ◦ C with a Lauda TE1C digital ring tensiometer, equipped with a thermostated solution compartment. The standard deviation of four successive readings was <0.10 mN m−1 . 2.4.2. Solution conductivity Conductivity measurements were recorded at 25 ◦ C with a PC-interfaced Fisher Accumet 50 ion-meter, provided with a DM-C1 micro-conductivity cell (Digimed, São Paulo) and Schott Titronic T200 programmed burette. A home-developed software package was used both for programmed dilution of the concentrated surfactant solution, and acquisition of conductance data. 2.4.3. EMF measurements The above-mentioned ion-meter was employed for measuring EMF, at 25 ◦ C. A Corning model 476416 double-junction Ag/AgCl reference electrode and Orion model 9417BN chloride ion selective electrode (ISE) were employed. An example of the dependence of the EMF on surfactant concentration is shown in Fig. SM-3 (Fig. SM-3 in supplementary material). 2.4.4. Fluorescence measurements The emission spectra of micelle-solubilized pyrene (hereafter referred to as “probe”) were employed for the determi-

nation of both cmc and Nagg . Determination of the former property is carried out by examining the effect of increasing [surfactant] on the ratio between the intensities of the first and third vibronic bands (I1 , at 374 nm and I3 , at 384 nm, respectively) of the emission spectrum of pyrene, whose concentration was kept constant. The appropriate volume of a 1 × 10−4 mol/L pyrene solution in ethanol was introduced into a volumetric flask, the solvent was dried with a stream of nitrogen. Water was added and the flask was agitated overnight (tube rotator), the resulting solution filtered through a 2 µm membrane; final [pyrene] = 2 × 10−6 mol L−1 . This solution was employed as “solvent” for making up the surfactant stock solution (hence [probe] is constant). A series of solutions of constant [probe] and increasing [surfactant] were obtained by mixing the appropriate volumes of the two solutions, i.e., pyrene/water and pyrene/surfactant/water. The probe was excited at 337 nm and 25 ◦ C by employing a Varian-Cary Eclipse fluorimeter. The cmc was calculated from the first derivative of plot of the (I1 /I3 ) ratio versus [surfactant], as shown in Fig. SM-1. Micellar Nagg were determined from C153-mediated steadystate fluorescence quenching of micelle-solubilized probe [30–32]. The resulting Stern–Volmer plot (effect of [C153] on the intensity of fluorescence, Fig. SM-2) and details of the calculation of Nagg are given in the appropriate section of SM. Ethanolic solution of pyrene was pipetted into a volumetric flask, followed by evaporation of the solvent (by N2 ). A surfactant solution of known concentration, 10 times the cmc (determined from conductivity measurement) was added, and the solution made up to the mark; complete probe dissolution was assured by sonication under inert atmosphere (10 min, 40 kHz, 25 ◦ C). The fluorescence quencher, C153 was dissolved in the above-mentioned probe/surfactant solution; final [C153] = 1.0 × 10−2 mol L−1 . A series of solutions of constant [pyrene]; constant [surfactant], and increasing [C153], 1.0 × 10−7 –5.0 × 10−5 mol L−1 , were obtained by mixing the appropriate volumes of the above-mentioned solutions, i.e., probe/surfactant/water and probe/surfactant/C153/water. Before fluorescence measurements, all solutions were sonicated under N2 and allowed to stand for 10 min. 2.4.5. Static light scattering Static light scattering (SLS) was recorded at 25 ± 0.1 ◦ C (Malvern 4700 system; 25 mW He/Ne laser source). The “solvent” employed to dissolve the surfactant was 0.86 × 10−4 mol L−1 NaCl (electrolyte concentration = surfactant cmc); all solutions were filtered through 0.22 µm cellulose acetate membrane. SLS measurements were recorded at 90◦ scattering angle. Solution refractive index increment (∂n/∂[Surf]) was measured with Optilab 903 interferometric differential refractometer (Wyatt) operating at 633 nm. A home-developed software package was used for acquisition of the scattering data, calculation of the cmc, and of the (weight averaged) micellar molecular weight from Debye plot.

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

299

3. Results and discussion

3.3. Adsorption at solution/air interfaces

3.1. Synthesis of RMeImCl

The effect of traces of alkaline earth ions on the adsorption of anionic surfactants at solution/air interface has been discussed [43]. Although these impurities, if present, are not expected to perturb cationic micelles, we have soaked all glassware before use in EDTA solution (1 × 10−4 mol L−1 ), followed by washing with water. The effectiveness and efficiency of surfactant adsorption are given by solution surface tension at the cmc, γcmc , and surfactant concentration required to reduce the surface tension of water by 20 mN m−1 , C20 , respectively [3]. These properties, which are relevant to wetting and emulsification, are listed in Table SM-4. The data of the latter table show that, as a function of increasing length of R, the effectiveness shows little variation, whereas the efficiency increases. The same tendencies were observed for other cationic surfactants [3]. The values of C20 show that the adsorption of RMeImCl (4.5 to 0.4 × 10−3 mol L−1 , for C12 MeImCl and C16 MeImCl, respectively) is somewhat more efficient than RPyCl (C20 from 8.0 to 0.5 × 10−3 mol L−1 , for C12 PyCl [44] and C16 PyCl [45], respectively). The minimum area/surfactant molecule at solution/air interface (Ainterface ) and Gibbs free energy of adsorption, G0ads for the four surfactant series are listed in Table 1. As expected, Ainterface of RMeImCl decreases as a function of increasing the length of R, due to concomitant closer packing of monomers at the interface [3,4]. It is not possible to compare the two series with heterocyclic head-groups because (Ainterface ) reported for RPyCl do not show a trend. For RMeImCl and R AEtBzMe2 Cl, plots (not shown) of these areas versus the number of carbon atoms, Nc , in R or R , are linear, leading to the following equations, where (r) refers to the correlation coefficient:

The synthesis of ILs is generally not beset by any major problem. Due to their extremely low vapor pressure, vide supra, which renders distillation unfeasible, laborious and often time-consuming purification is necessary before measuring their physico-chemical properties. These difficulties may have led to some conflicting physico-chemical properties of “purified” ILs. Examples are m.p. of C2 MeImBF4 , (5.8 to 15 ◦ C) [33–36], and viscosities of C2 MeImBF4 (37 and 66.5 mPa s) and of C8 MeImPF6 (691 and 866 mPa s) [18]. Therefore, employing 1-alkyl chlorides and 1-methylimidazole as received [22,26,37], i.e., without determining their purities may lead to surface-active impure products that are practically impossible to purify by conventional methods, e.g., crystallization or flash-column chromatography. Note that satisfactory 1 H NMR data does not necessarily indicate a surface-active pure substance (as indicated from the surface tension, γ , versus log[surfactant] plots) [38–40]. The absence of water, light and oxygen during the synthesis, and continuous (not interrupted) reflux/stirring are important for obtaining of pure SAILs. Since “hydrophobic” ILs absorb water at room temperature [41,42] frequent drying under reduced-pressure, over P4 O10 is necessary. In summary, the laborious reagent purification, and the above-mentioned precautions during the synthesis are necessary in order to obtain surface-active pure products. 3.2. Properties of aqueous solutions: Note: Details of calculations of all quantities discussed are given in SM The Discussion below is arranged in the following order: adsorption of surfactants at the solution/air interface is addressed; the micellization process is then considered; finally, in order to probe the effect of head-group structure on adsorption and micellization, we compare data of the present series with those of other cationic surfactants, of the same hydrophobic tails (C10 – C16 ). The latter include three classes, namely: surfactants with a heterocyclic head-group, alkylpyridinium chlorides, RPyCl; surfactants with an aromatic head-group, alkylbenzyldimethylammonium chlorides, RBzMe2 Cl; surfactants with a headgroup that carries a benzyl moiety and a group (–CONH–) that forms hydrogen-bonds, benzyl (3-acylaminoethyl) dimethylammonium chlorides, RCONH(CH2 )2 N+ (CH3 )2 CH2 C6 H5 Cl− , or R AEtBzMe2 Cl. In these acronyms, R, Py, R , A, Et, Bz, and Me, stand for the alkyl group, the pyridinium ring, the acyl group, –NH–, –C2 H4 –, C6 H5 CH2 –, and CH3 –, respectively. The series alkyltrimethylammonium chlorides, RMe3 Cl will be used when the data of head-group that carries no aromatic or heterocyclic ring is required. A note is in order with regard to R AEtBzMe2 Cl: for comparison, we consider surfactants with R or R of the same number of carbon atoms, e.g., we compare C12 MeImCl with C11 COAEtBzMe2 Cl.

Ainterface = 1.461 − 0.062Nc (RMeImCl),

r = 0.9896,

(1)



Ainterface = 1.095 − 0.0225Nc (R AEtBzMe2 Cl), r = 0.9927.

(2)

That is, as a function of increasing the length of (R), the imidazole-based surfactants pack more closely at the interface than R AEtBzMe2 Cl. The reason is that the hydrophilic group of the latter is voluminous because it includes the amide group [38]. The transfer of the discrete surfactant segments from bulk water to the interface contributes to G0ads . These contributions are due to the terminal CH3 group of the hydrophobic chain, G0CH3 ; the methylene groups of the alkyl chain, (NCH2 G0CH2 ), and the head-group, G0head-group , as given by the following equation [3,4]: G0ads = G0CH3 + NCH2 G0CH2 + G0head-group .

(3)

Equation (3) predicts a linear correlation between G0ads and NCH2 , where the intercept includes contribution from the terminal methyl plus the head-group. Since G0CH3 is independent of the chain length of the surfactant, its contribution is constant in a homologous series. That is, differences of the intercepts

300

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

Table 1 Relationship between surfactant structure and the minimum area/surfactant molecule at the solution/air interface, Ainterface , and Gibbs free energy of adsorption, G0ads R or R

RMeImCla Ainterface (nm2 )

C10

0.85

−30.16

C12

0.72

−32.03

0.62d 0.525e 0.73f

C14

0.56

−33.81

1.11 f

−35.23

0.49f

C16

0.49

G0

G0CH

a b c d e f h g

Ainterface (nm2 )

G0ads

(kJ mol−1 )

Ainterface (nm2 )

−31.6d

0.62g

G0ads

(kJ mol−1 )

Ainterface (nm2 )

G0ads (kJ mol−1 )

−19.7

0.87

−33.3

−23.0

0.82

−35.9

−26.4

0.79

−38.9

−29.4

0.73

−41.3

0.420h

(kJ mol−1 ) Series

(kJ mol−1 )

R AEtBzMe2 Clc

RBzMe2 Clb

RPyCl G0ads

2

G0CH

3

G0CH

2

+ G0head-group −0.85

−22.6

G0CH

G0CH

3

+ G0head-group –

–

2

G0CH

3

G0CH

+ G0head-group −1.6

−3.8

2

G0CH + 3

G0head-group -1.4

−22.5

Present work, at 25 ◦ C. At 25 ◦ C. From: Blois and Swarbrick [64]. At 25 ◦ C. From: Shimizu and El Seoud [38]. At 25 ◦ C. From: Rosen et al. [44]. At 30 ◦ C. From: Mataa et al. [65]. At 25 ◦ C. From: Semmler et al. [66]. At 35 ◦ C. From Okano et al. [68]. At 35 ◦ C. From: González-Pérez et al. [67].

essentially reflect the transfer of the head-group from bulk solution to the solution/air interface. Equation (3) applies satisfactorily to the data of SAILs (r = 0.9982); data for three surfactant series are listed in Table 1. As expected, the values of G0CH2 are similar because the transfer of a CH2 group from aqueous pseudo-phase to the solution/air interface should be negligibly dependent of the structure of the surfactant. On the other hand, (G0head-group + G0CH3 ) are much more negative for RMeImCl and R AEtBzMe2 Cl than the RBzMe2 Cl series. That is, the adsorption of the former two series is more favorable than that of the latter one, due to hydrogen-bonding, as will be discussed below. 3.4. Aggregation: Critical micelle concentrations, degrees of micelle dissociation, aggregation numbers, and Gibbs free energies of micellization Representative plots of solution conductance, surface tension, intensity ratio of I1 /I3 vibronic bands of micelle-solubilized pyrene, and free counter-ion, [Cl− ]f , as a function of [surfactant] are shown in parts A to D of Fig. 1; the complete set of experimental data for all surfactants is shown in Fig. SM-1. Parts A, B, and D consist of two straight lines intersecting at the cmc. The dependence of I1 /I3 on [RMeImCl] is sigmoidal; the inflection point (i.e., where the first derivative, (∂(I1 /I3 )/∂[Surf]) is zero) is the cmc. These values are listed in Table 2, along with αmic and Nagg ; both properties determined by use of two techniques. Regarding Table 2, the following is relevant:

(i) Results of the EMF experiment were used to calculate cmc and αmic . Calculation of the last property requires measurements up to high surfactant concentrations, ca. 60 × cmc; high solution viscosity precluded carrying out the experiment with C10 MeImCl. In addition to Fig. 1B, surface tension plots for the other surfactants (Fig. SM-1) indicated that the compounds studied are surface-active pure. (ii) The agreement between the values of cmc, and of αmic calculated from data of different techniques is satisfying, considering that a micelle-solubilized probe was used in the fluorescence experiment. Values of αmic decrease linearly as a function of increasing the chain-length of the alkyl group (slope = −0.0135, r = 0.9959), because of the accompanied increase of Nagg (vide infra) and the concomitant decrease in the surface charge density [3,46]. (iii) As shown by Eq. (12) of supplementary material, calculation of G0mic requires knowledge of cmc and αmic . Conductance data have been employed in these calculations, because both quantities are calculated from the same set of experimental data. It is possible to calculate αmic by Frahm’s method, i.e., by dividing the slopes of the straight lines above and below the cmc. Although this method is simple, it is only a useful approximation when Nagg is not available [47]. The reason is that the conductivity of the micelle (a “macro-ion”) is not taken into account; this leads to relatively high αmic . E.g., αmic for C10 MeImCl (Fig. 1A), calculated by Frahm’s method is 0.31. As shown in Table 2, values of αmic calculated by Evans equation (where micellar conductivity is taken into account) are in excellent agreement with those based on the use of ISE [48].

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

301

Table 2 Relationship between surfactant structure and critical micelle concentration, cmc, aggregation number, Nagg , degree of counter-ion dissociation, αmic , and Gibbs free energy of micelle formation, G0mic a Surfactant

103 × cmc (mol L−1 ) surface tension

103 × cmc (mol L−1 ) conductivity

103 × cmc (mol L−1 ) EMF

103 × cmc (mol L−1 ) fluorescence

Nagg b

αmic c

G0mic (mol L−1 )d

C10 MeImCl C12 MeImCl C14 MeImCl C16 MeImCl

39.90 13.17 2.98 0.87, 0.89e

40.47 13.47 3.68 0.86

– 12.8 3.52 0.92

40.52 14.0 3.65 1.30

40, 48 58, 67 79, 89 99, 114, 118f

0.26, – 0.23, 0.24 0.20, 0.18 0.19, 0.18

−31.14 −36.51 −42.91 −49.68

a All measurements were carried out at 25 ◦ C. The uncertainty limits of cmc, determined by each technique were found to be 0.5%. The reasons for the observed small dependence of cmc on the technique employed have been discussed elsewhere [69,70]. For example, Mukerjee and Mysels [70] have compiled 54 cmc’s for sodium dodecyl sulphate and cetyltrimethylammonium bromide (measurements at 25 ◦ C), differing, for the same technique, by 100 and 22%, respectively. b N agg are listed in the following order: from fluorescence, by using molecular geometry and, for C16 MeImCl, from light scattering. c α mic is listed in the following order: from conductivity and from EMF measurements. d Calculated using conductivity-based cmc and α mic . e From Thomaier and Kunz [22]. f In the presence of 0.86 × 10−4 mol L−1 of NaCl. Under these conditions, the SLS-based cmc of C MeImCl was found to be 0.31 × 10−3 mol L−1 . The 16 (∂n/∂[Surf]) calculated was 0.183 mL g−1 .

Fig. 1. Dependence on surfactant concentration of the following properties: solution conductance of C10 MeImCl (A); solution surface tension of C12 MeIm (B); Intensity ratio of I1 /I3 vibronic bands of C14 MeImCl-solubilized pyrene (C); concentration of the free counter-ion, [Cl− ]f , calculated from EMF measurements of C16 MeImCl (D). For (A) the points are experimental and the line has been calculated by Eq. (4) of supplementary material.

(iv) Calculation of αmic according to Evans equation requires knowledge of Nagg . These have been calculated from fluorescence quenching of the micelle-solubilized probe by C153 and, for C16 MeImCl, from results of SLS. Nagg was also theoretically calculated from the length of the surfactant monomer, and the diameter of the (spherical-shaped) aggregate, by assuming that the monomer has a stretched,

all-trans conformation within the micelle (Spartan-Pro program package, version 5.1, Wave Function Inc.). A plot of theoretically calculated Nagg versus those calculated from fluorescence measurements is linear, with intercept, slope, and (r) of 2.80, 1.11 and of 0.9991, respectively. This average difference of 11% between the two sets of Nagg merits comments. First, the theoretical calculation overes-

302

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

timates Nagg , because the molecular geometry is optimized for an isolated molecule in vacuum. Unlike molecular dynamics simulations, repulsive electrostatic interactions between monomers in the micelle [49–51], and their possible conformations, e.g., the average number of gauche bonds per hydrophobic chain [52], were not considered in our calculations. Their inclusion should increase the effective volume of the micellized monomer, hence decrease Nagg . Second, αmic , calculated by Evans equation is not very sensitive to Nagg , as shown by the following values for αmic of RMeImCl, based on theoretical Nagg : 0.27, 0.24, 0.19, and 0.19, for R = C10 , C12 , C14 , and C16 , respectively. The SLS based Nagg is higher because NaCl screens the electrostatic repulsion between the head-group, this should lead to a decrease in cmc (from ca. 0.9 to 0.3 × 10−3 mol L−1 , see footnote f of Table 2) and an increase of the aggregation number [4,9]. (v) Equation (4) for G0mic is written analogously to Eq. (3) [3,4]: G0mic = G0CH3 + NCH2 G0CH2 + G0head-group ,

(4)

where the terms G0head-group , G0CH3 , and G0CH2 refer to the contribution to G0mic of the surfactant moieties, as discussed above for adsorption. Application of Eq. (4) to RMeImCl gave a straight line (r = 0.9987); data for the four surfactant series are collected in Table 3. (vi) Before addressing the data of the last table, a comment on αmic and NCH2 employed in the calculations is in order, see Tables 2 and SM-5. Consider first the degree of dissociation of the counter-ion. These values are available from Evans equation for RMeImCl, RBzMe2 Cl and R AETBzMe2 Cl, but not for RPyCl. As shown in Table SM-5, Frahm’s equation-based αmic for RPyCl are 0.4. These values are high (vide supra), consequently, are not convenient for comparison with the results of the other series. McGinnis and Woolley satisfactorily fitted their osmometric and calorimetric data for the micellization of RPyCl by using the mass-action law and the following values for αmic and Nagg : 0.27, 0.23, 0.20, 0.16 (αmic ); 36, 52, 70, and 85 (Nagg ), for R = C10 , C12 , C14 , and C16 , respectively [53]. We employed these αmic values (underlined in Table SM-5) in calculations of G0mic . For RMeImCl, RBzMe2 Cl and RPyCl, the values of NCH2 employed in Eqs. (3) and (4) were 9, 11, 13, and 15 for R = C10 , C12 , C14 , and C16 , respectively. The corresponding figures for R AEtBzMe2 Cl are 8, 10, 12, and 14, respectively, i.e., the carbon atom of the –CONH– group was not included as a part of the hydrophobic tail. The reason is that IR and NMR studies showed that the micellar interfaces of surfactants that carry the head-group –CONH(CH2 )n –N+ (CH3 )2 – CH2 C6 H5 ; n = 2 and 3, lie at the amide group, i.e., behind the quaternary ammonium ion [38,40]. Therefore, the carbon atom of the –CONH– is not present in the micellar interior, and should not be included in NCH2 . (vii) Table 3 shows that G0CH2 is similar for the four surfactant series (−3.35 ± 0.25 kJ mol−1 ), since the transfer of a

Table 3 Contribution of the surfactant discrete segments, CH2 and CH3 + head-group to the G0mic , at 25 ◦ Ca,b G0mic (kJ mol−1 ) RMeImCl RPyCl RBzMe2 Clc R AEtBzMe2 Cld

CH2

CH3 + head-group

r

−3.10 −3.45 −3.61 −3.34

−2.85 1.60 1.77 −6.31

0.9987 0.9991 0.9991 0.9995

a At 25 ◦ C; data obtained by application of Eq. (4). In applying the latter equation, values of Nc for RMeImCl, RPyCl and RBzMe2 Cl are 9, 11, 13, and 15. For R AEtBzMe2 Cl, the equivalent numbers are 8, 10, 12, and 14, respectively. b For the first, third and fourth series, α mic is that calculated by Evans equation. For RPyCl, αmic is that calculated from osmotic pressure data [53]. c Data taken from Ref. [40]. d Data taken from Ref. [38].

CH2 group from water to the micelle should be independent of head-group structure. This average value is also similar to that of G0transfer of the methylene groups of alkanes from water to bulk hydrocarbons, −3.56 kJ mol−1 [54]. On the other hand, G0head-group shows the following order RPyCl ≈ RBzMe2 Cl > RMeImCl > R AEtBzMe2 Cl. How can this order be explained? To a first approximation, because G0CH2 is independent of surfactant structure, the dependence of G0mic on NCH2 is given by (G0head-group ), i.e., is controlled by the lipophilicity of the head-group. One empirical scale to quantify this property is log P which is based on the partition coefficient of a substance between n-octanol and water (where both solvents are mutually saturated): log P = log([substance]n-octanol /[substance]water ) [55]. Indeed, for the same range of R (C10 to C16 ), G0head-group for RBzMe2 Cl are smaller than those of RMe3 Cl [39,56]. This is due to the fact that the benzyl group is more hydrophobic than the methyl group; log P for benzyldimethylamine and trimethylamine (models for the head-groups) are 2.31 and 0.16, respectively [57]. However, the order of (G0head-group ) for RMeImCl, RBzMe2 Cl and RPyCl is the inverse of that expected, based only on the lipophilicity of the head-group; values of log P for (the precursors) 1-methylimidazole, benzyldimethylamine and pyridine are −0.06, 2.31 and 0.65, respectively [57]. Since the contribution of head-group lipophilicity to Gmic is certainly operative, we explore another structural factor. Recent X-ray results of crystalline C14 MeImCl, and theoretical calculations on the interactions in liquid C4 MeImCl have been explained based on the formation of hydrogen-bonding between the counter-ion and the protons of the heterocyclic ring, in particular the relatively acidic H2 [58,59]. Note that this hydrogen-bonding persists in aqueous solutions of ILs [60]. We have calculated the partial charges on the different hydrogens of 1-methylimidazole, 1,3-dimethylimidazolium cation, pyridine, and 1-methylpyridinium cation, respectively (models for the head-groups), solvated in water. The results of these calculations are

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

listed in Table SM-6. They clearly show that H-2 of the diazole is much more acidic than any other hydrogen in both heterocyclic rings; the same applies to the corresponding cations. In fact all hydrogens of the imidazolium ring are more acidic than the corresponding ones of the pyridinium ring, due to the presence of two heteroatoms in the former. That is, the favorable effect of (more lipophilic) head-groups of RPyCl and RBzMe2 Cl on (G0head-group ) is more than compensated for by electrostatic attraction between the counter-ion and (more acidic) hydrogens of the imidazolium ring. The importance of hydrogen-bonding to (G0head-group ) is also clear from the fact that aggregation of R AEtBzMe2 Cl is more favorable than that of RMeImCl. Hydrogen bonding is operative for both head-groups, but is stronger for the former series, because the –CONH– group may form direct, or water-mediated hydrogen-bonds to both the counterion and the neighboring molecule in the micelle [38]. Finally, there is an excellent linear correlation between fluorescence-based Nagg of RMeImCl and those calculated for RPyCl by McGinnis and Woolley; values underlined in Table SM-5: Nagg (RMeImCl) = −1.933 + 1.158Nagg (RPyCl); r = 0.9999. This 15% difference probably reflects the efficiency of packing of the smaller imidazolium ring, relative to its pyridinium counterpart. 4. Conclusion The intense interest in studying short-chain ILs has not been matched by effort to study their surface active counterparts, although the latter have very interesting properties, in addition to being surface active: low melting points (32–34 and 86–87 ◦ C, for C12 MeImCl and C12 PyCl, respectively) [61], high polarities (situated between those of dipolar aprotic solvents and protic ones) [62], and high chemical and thermal stabilities. Additionally, the presence of two alkyl groups in the imidazolium ring introduces a structural flexibility because its bulk can be tailored to ones need. These favorable characteristics and the lack of information on SAILs have prompted us to synthesize a series of pure 1-alkyl-3-methylimidazolium chloride surfactants and to investigate their adsorption at solution/air interface, and aggregation in aqueous solutions by employing several techniques, namely, conductivity, surface tension, EMF, fluorescence and light scattering. The emphasis has been on: (i) determination of the effects of increasing the alkyl chainlength on adsorption and micelle formation; (ii) comparison of the results of this novel series with those of other cationic surfactants. Gibbs free energies of adsorption and/or micellization have been calculated, and then separated into contributions from the discrete surfactant segments, (G0CH2 ) and (G0CH3 + G0head-group ). Whereas the former free energy is similar to that of other surfactants, the latter one is lower for RMeImCl than that of corresponding RPyCl, due to direct, or water-mediated hydrogen bonding between the counter-ion (Cl− ) and the hydrogens of the imidazolium ring, in particular H2. This interaction more than compensates for the favorable effect on G0mic of the more lipophilic pyridinium ring. Thus

303

the interfacial water of SAILs is probably more structured than that of other cationic surfactants, e.g., RMe3 Cl or RPyCl, this may lead to favorable enthalpy/entropy of activation of micellemediated reactions, e.g., acyl transfers [63]. Acknowledgments We thank the State of São Paulo Research Foundation (FAPESP) for financial support and a postdoctoral fellowship to E.L. Bastos; the National Council for Scientific and Technological Research (CNPq) for a research productivity fellowship to O.A. El Seoud; the Egyptian Ministry of Higher Education for a travel/research fellowship to T. Abdel-Moghny. Supplementary material The online version of this article contains additional supplementary material. Please visit DOI: 10.1016/j.jcis.2007.04.028. References [1] D. Attwood, A.T. Florence, Surfactant Systems: Their Chemistry, Pharmacy, and Biology, Chapman and Hall, London, 1984. [2] O.A. El Seoud, Organized Assemblies in Chemical Analysis, in: W.L. Hinze (Ed.), JAI Press, Greenwich, 1994. [3] M.J. Rosen, Surfactants and Interfacial Phenomena, Wiley, New York, 1989. [4] P.C. Hiemenz, R. Rajagopalan, Principles of Colloid and Surface Chemistry, Marcel Dekker, New York, 1997. [5] C.A. Bunton, F.H. Quina, L.S. Romsted, Acc. Chem. Res. 24 (1991) 357. [6] C.A. Bunton, G. Savelli, Adv. Phys. Org. Chem. 22 (1986) 213. [7] C.A. Bunton, J. Mol. Liq. 72 (1997) 231. [8] K.S. Birdi, Handbook of Surface and Colloid Chemistry, CRC Press, Boca Raton, 1997. [9] D.F. Evans, H. Wennerström, The Colloidal Domain, second ed., Wiley– VCH, Weinheim, 1999, 157. [10] T.J. Broxton, J.R. Christie, D. Theodoridis, J. Phys. Org. Chem. 6 (1993) 535. [11] E. Jungerman, Cationic Surfactants, Marcel Dekker, New York, 1969. [12] J. Cross, E.J. Singer, Cationic Surfactants: Analytical & Biological Evaluation, Marcel Dekker, New York, 1994. [13] P.M. Holland, D.N. Rubingh (Eds.), Cationic Surfactants: Physical Chemistry, Marcel Dekker, New York, 1991. [14] J.M. Richmond, Cationic Surfactants: Organic Chemistry, Marcel Dekker, New York, 1990. [15] J.D. Holbrey, K.R. Seddon, Clean Prod. Proc. 1 (1999) 233. [16] T. Welton, Chem. Rev. 99 (1999) 2071. [17] W. Keim, P. Wasserscheid, Angew. Chem. Int. Ed. 39 (2000) 3772. [18] C.F. Poole, J. Chromatogr. A 1037 (2004) 49. [19] C. Chiappe, D. Pieraccini, J. Phys. Org. Chem. 18 (2005) 275. [20] J. Sirieix-Plénet, L. Gaillon, P. Letellier, Talanta 63 (2004) 979. [21] L. Gaillon, J. Sirieix-Plénet, P. Letellier, J. Sol. Chem. 33 (2004) 1333. [22] S. Thomaier, W. Kunz, J. Mol. Liq. 130 (2007) 104. [23] U. Domanska, E. Bogel-Lukasik, R. Bogel-Lukasik, Chem. Eur. J. 9 (2003) 3033. [24] U. Domanska, E. Bogel-Lukasik, R. Bogel-Lukasik, J. Phys. Chem. B 107 (2003) 1858. [25] U. Domanska, L. Mazurowska, Fluid Phase Equilib. 221 (2004) 73. [26] A.E. Bradley, C. Hardacre, J.D. Holbrey, S. Johnston, S.E.J. McMath, M. Nieuwenhuyzen, Chem. Mater. 14 (2002) 629. [27] A. Downard, M.J. Earle, C. Hardacre, S.E.J. McMath, M. Nieuwenhuyzen, S.J. Teat, Chem. Mater. 16 (2004) 43.

304

O.A. El Seoud et al. / Journal of Colloid and Interface Science 313 (2007) 296–304

[28] W.L.F. Armarego, D.D. Perrin, Purification of Laboratory Chemicals, fourth ed., Butterworths–Heinemann, Oxford, 1998. [29] J. Dupont, C.S. Consorti, P.A. Suarez, R.F. Souza, Org. Synth. 79 (2002) 236. [30] N.J. Turro, A. Yekta, J. Am. Chem. Soc. 100 (1978) 5951. [31] R. Sabate, J.J. Estelrich, Phys. Chem. B 107 (2003) 4137. [32] P.J. Tummino, A. Gafni, Biophys. J. 64 (1993) 1580. [33] J.D. Holbrey, K.R. Seddon, J. Chem. Soc. Dalton Trans. (1999) 2133. [34] A.B. McEwen, H.L. Ngo, K. LeCompte, J.L. Goldman, J. Electrochem. Soc. 146 (1999) 1687. [35] J. Fuller, R.T. Carlin, R.A. Osteryoung, J. Electrochem. Soc. 144 (1997) 3881. [36] A. Noda, M. Watanabe, Electrochim. Acta 45 (2000) 1265. [37] D. Demberelnyamba, K.-S. Kim, S. Choi, S.-Y. Park, H. Lee, C.-J. Kim, Bioorg. Med. Chem. 12 (2004) 853. [38] S. Shimizu, O.A. El Seoud, Langmuir 19 (2003) 238. [39] S. Shimizu, P.A.R. Pires, W. Loh, O.A. El Seoud, Colloid Polym. Sci. 282 (2004) 1026. [40] P.A.R. Pires, O.A. El Seoud, Prog. Colloid Polym. Sci. 133 (2006) 131. [41] C.D. Tran, S.H.P. Lacerda, D. Oliveira, Appl. Spectrosc. 57 (2003) 152. [42] K.R. Seddon, A. Stark, M.-J. Torres, Pure Appl. Chem. 72 (2001) 2275. [43] J. Eastoe, S. Nave, A. Downer, A. Paul, A. Rankin, K. Tribe, Langmuir 16 (2000) 451. [44] M.J. Rosen, M. Dahanayake, A.W. Cohen, Colloids Surf. 5 (1982) 159. [45] D. Varade, T. Joshi, V.K. Aswal, P.S. Goyal, P.A. Hassan, P. Bahadur, Colloids Surf. A 259 (2005) 95. [46] R. Zana, J. Colloid Interface Sci. 78 (1980). [47] J. Frahm, S. Diekmann, A. Haase, Ber. Bunsenges. Phys. Chem. 84 (1980) 566. [48] H.C. Evans, J. Chem. Soc. (1956) 579.

[49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70]

S. Yuan, L. Ma, X. Zhang, L. Zheng, Colloids Surf. A 289 (2006) 1. S. Fujiwara, M. Hashimoto, T. Itoh, J. Plasma Phys. 72 (2006) 1011. J. Gao, W. Ge, G. Hu, J. Li, Langmuir 21 (2005) 5223. D.W.R. Gruen, Prog. Colloid Polym. Sci. 70 (1985) 6. T.P. McGinnis, E.M. Wooley, J. Chem. Thermodyn. 29 (1997) 401. H.D. Nelson, C.L. de Lingy, Recl. Trav. Chim. Pays-Bas 87 (1968) 528. A.J. Leo, C. Hansch, Perspect. Drug Discovery Des. 17 (1999) 1. P.A.R. Pires, O.A. El Seoud, J. Colloid Interface Sci. 304 (2006) 474. C. Hansch, A.J. Leo, D. Hockman, Exploring QSAR: Hydrolysis and Steric Constants, Am. Chem. Soc., Washington, DC, 1995. S.C. Handy, M. Okello, J. Org. Chem. 70 (2005) 1915. S. Kossmann, J. Thar, B. Kirchner, P.A. Hunt, T. Welton, J. Chem. Phys. 124 (2006) 174506-1. J. Dupont, J. Braz. Chem. Soc. 15 (2004) 341. R.S. Shelton, M.G.V. Campen, C.H. Tilford, H.C. Lang, L. Nisonger, F.J. Bandelin, H.L. Rubenkoenig, J. Am. Chem. Soc. 68 (1946) 757. C. Reichardt, Green Chem. 7 (2005) 339. S. Possidonio, F. Siviero, O.A. El Seoud, J. Phys. Org. Chem. 12 (1999) 325. D.W. Blois, J. Swarbrick, J. Colloid Interface Sci. 36 (1971) 226. J. Mataa, D. Varadea, P. Bahadurb, Thermochim. Acta 428 (2005) 147. A. Semmler, H.-H. Kohler, J. Colloid Interface Sci. 218 (1999) 137. A. González-Pérez, L.M. Varela, M. García, J.R. Rodríguez, J. Colloid Interface Sci. 293 (2006) 213. L.T. Okano, F.H. Quina, O.A. El Seoud, Langmuir 16 (2000) 3119. P. Mukerjee, Adv. Colloid Interface Sci. 1 (1967) 241. P. Mukerjee, K.J. Mysels, Critical Micelle Concentrations of Aqueous Surfactant Systems, NSRDS-NBS, 36th ed., U.S. Government Printing Office, Washington, DC, 1971.