Interplay of Dynamical Properties between Ionic Liquids and Ionic

Subscriber access provided by UNIV OF MISSISSIPPI

Article

The Interplay of Dynamical Properties Between Ionic Liquids and Ionic Surfactants: Mechanism and Aggregation Michael McCutchen, Lang G Chen, Harry Bermudez, and Silvina Matysiak J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.5b05151 • Publication Date (Web): 30 Jun 2015 Downloaded from http://pubs.acs.org on July 3, 2015

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

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

Page 1 of 26

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

The Journal of Physical Chemistry

The Interplay of Dynamical Properties between Ionic Liquids and Ionic Surfactants: Mechanism and Aggregation Michael McCutchen,† Lang Chen,‡ Harry Bermudez,∗,‡ and Silvina Matysiak∗,† Fischell Department of Bioengineering, University of Maryland, College Park, Maryland 20742, USA, and Department of Polymer Science and Engineering, University of Masachusetts, Amherst, Massachusetts 01003, USA E-mail: [email protected]; [email protected]

Abstract The dynamical and aggregation behavior of sodium dodecylsulfate (SDS) in 1-ethyl3-methyl imidazolium ethylsulfate [EMIM+ ][EtSO4 − ] are characterized experimentally and computationally. A retardation of the ionic liquid (IL) and SDS diffusion coefficients with a concentration increase of SDS is observed. In agreement with experiments, aggregation is detected for concentrations higher than the experimental critical micelle concentration (CMC), which is mostly driven by alkyl tail aggregation. Solvent-exposed hydrophobic patches are observed on the micelle’s surfaces. The hydrophobic tails of the IL molecules are found to fill those micelle’s hydrophobic patches. Also, penetration of the IL is found in the SDS micelles indicating that the IL acts as a cosurfactant, allowing the formation of “mixed” micelles. A higher level of Na+ counterion dissociation ∗

To whom correspondence should be addressed University of Maryland ‡ University of Massachusetts †

1

ACS Paragon Plus Environment


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

compared to previous studies of SDS micelles in aqueous solutions is also observed. A multi-layering effect of alternating IL anions and cations is detected at the surface of the formed aggregates. The observed increase in system ordering with SDS concentration is what hinders the mobility of each chemical species. Keywords: Ionic liquids, surfactant aggregation, all-atom simulations, cosurfactant, counterion dissociation, multielectric layer

Introduction Surfactants, such as sodium dodecylsulfate (SDS), are extensively used as detergents, solubilizers and emulsifiers due to their ability to self-organize into many supramolecular forms. 1–3 The organization and form of surfactant aggregates highly depends on the chemical composition of the solvent, temperature, and concentration. Towards understanding the formation and stability of surfactant (e.g., SDS) micelles in aqueous environments, extensive experimental 4,5 and molecular dynamic (MD) simulations 6,7 have been conducted. Experimental studies have recently shown that surfactant aggregation can also occur in a variety of roomtemperature ionic liquids (ILs). 8–13 However, an understanding at the microscopic level regarding the aggregation mechanism and the stability of such aggregates in ILs is presently lacking. Room-temperature ionic liquids (ILs) are a class of organic salts whose unique properties make them a promising alternative to conventional solvents. 14–18 The combination of surfactants and ILs is particularly relevant to applications such as heterogeneous catalysis (e.g., SILP 19,20 ) and emulsion technologies that rely on control over interfacial properties. 21–23 Furthermore the use of surfactants to modulate the interfacial properties of ILs relies on mixing rather than the time-consuming synthesis of new ILs. Despite the endless chemical diversity of ILs, the empirical approach to achieving desired properties has motivated the need for computational investigations. As a recent example, all-atom simulations of surfactants adsorbed at a vacuum/IL interface revealed that the mobility of the IL species depends on the 2


Page 2 of 26

Page 3 of 26

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


surfactant coverage at the interface. 24 Thus, characterization of the molecular interactions responsible for the behavior of surfactants in ILs will be useful for controlling interfacial properties and surfactant phase behavior. Here we explore the aggregation and dynamics of the system composed of 1-ethyl-3methyl imidazolium ethylsulfate [EMIM+ ][EtSO4 − ] and SDS. To obtain a microscopic understanding of the observed experimental trends, corresponding molecular dynamic simulations were performed. A detailed characterization of the structural changes of the IL due to SDS aggregation and the structure of the SDS aggregates is obtained. To our knowledge, we provide the first simulation of ionic surfactant aggregation in a neat IL starting from a random initial conformation and its comparison with experimental results.

Methods Experimental All solutions are prepared by directly dissolving a certain amount of SDS surfactants in [EMIM+ ][EtSO4 − ] at elevated temperature. PGSE-NMR diffusion measurements are carried out on a 400 MHz Bruker NMR spectrometer equipped with a temperature controller. The self-diffusion measurements are performed with a Gaussian-shape pulsed field gradient stimulated echo, whose magnitude is 5.35 Gauss/mm. The maximum gradient is chosen to be 95% of the full power to maintain the linearity of the amplifier power. The gradient strength is systematically modified from 95% to 2% in order to get 16 NMR spectra. The diffusion time, ∆, between the two pulses is set between 200-500 ms, and the gradient pulse duration, δ, is set between 2 and 6 ms, depending on the diffusion coefficient of the mobile species. The diffusion coefficient value is determined from the intensity change equation: 25

I = I0 e−Dγ

2 g 2 δ 2 (∆− δ ) 3

3

.


(1)


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

Here, I and I0 are the areas of the signal obtained with or without gradient pulses respectively, D is the diffusion coefficient. γ is the gyromagnetic ratio of proton, whose value is given by 2.675 ∗ 108 T −1 s−1 , g is the magnitude of the two gradient pulses.

Simulation Four ionic liquid/surfactant concentrations are simulated at atomistic level with the GROMACS software package. 26 For all systems, [EMIM+ ][EtSO4 − ] is used as the solvent and SDS as the surfactant, in the following concentrations (C) 575mM, 290mM and 52mM, with the last system being a bulk ionic liquid system. More details of the systems are presented in Table 1. The molecular structure of these molecules are shown in Figure 1. A nonpolarizable force field developed by Canongia et al., 27 which is compatible with the OPLS-AA force-field of Jorgensen et al. 28 is used to model the IL. The SDS molecule and the Na+ counter-ion is modeled with the OPLS-AA force field. Standard geometric combining rules are used for intramolecular nonbonded interactions between all pair of atoms (i, j) separated by 3 or more bonds or between molecules. 28 Each system consist of a random starting configuration of SDS and Na+ counter-ions in an xyz periodic cube solvated with an equilibrated box of [EMIM+ ][EtSO4 − ]. The SDS is solvated with equal numbers of EtSO4 − and EMIM+ molecules to maintain a net zero charge on the system. To randomize the systems, an NVT simulation at 4000K are run after energy minimization till any pre-formed micelles are destroyed. This is followed by a 1 ns simulation in NVT ensemble, to equilibrate the temperature at 400K, and by at least 1ns in NPT ensemble, to equilibrate the pressure at 1 bar. A temperature of 400K is chosen to achieve equilibrium in a reasonable amount of computational time. 29 Molecular dynamics are performed for 600ns in the NPT ensemble at 400K and 1atm (using a 2fs step size), for the 575 mM and 290mM systems, 150ns for the 52mM system, and 7 ns for the bulk IL system. Different lengths of production runs are used since each system achieve equilibrium at different times. Temperature coupling is performed each step, and 4


Page 4 of 26

Page 5 of 26

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


isotropic pressure coupling every 5 steps. Temperature and pressure coupling are achieved with the velocity rescale method and the Parrinello Rahman barostat respectively. 30 The system is assigned an isothermal compressibility of 3.41 ∗ 10−5 bar. 31 A Lennard-Jones cutoff of 1.2 nm is used. The long-range electrostatic interactions with periodic boundary conditions (xyz) are calculated by the particle-mesh Ewald method. 32 With this setup the density of bulk [EMIM+ ][EtSO4 − ] is found to be 1172 kg/m3 in agreement with previous experimental observations. 33

Analysis Methods Matching From the trajectory data, micelles are identified using a geometric criteria in similar styles, as in previous atomistic studies of surfactant self-assembly. 7 Two surfactants are members of the same micelle if any pair of carbons, belonging to different surfactants, are closer than a threshold value. By carefully experimenting and visualizing the micellar configurations we have setup a threshold value of 4.2 Å. This value corresponds to the peak of the carboncarbon radial distribution function (RDF), as depicted in Figure S2. These micelles are then re-evaluated based on a second criterion: each surfactant that is not paired to at least three other members of its micelle is removed. This process is repeated recursively in order to eliminate the case where a single surfactant molecule connects two independent micelles (resulting in the two micelles being erroneously reported as a single micelle). All the analysis described below are done using only trajectory data after equilibrium is reached. For the concentrations that surfactant aggregation is observed, equilibrium is defined when the average micelle size and the fraction of free monomers remains constant. On the other hand, for the systems that do not exhibit any micelle formation or for bulk IL, equilibrium is reached when the RDFs between different pair of molecules do not change with different periods of time.

5



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

Page 6 of 26

Critical micelle concentration The critical micelle concentration (CMC) is calculated from the equilibrium part of the trajectories that exhibit micelle aggregation. This is done by measuring the free monomer concentration in these micellar solutions, which analytic theory predicts to be equal to the CMC. 34 Therefore, CMC is calculated as nf ree . Vtotal − Vmicelles

(2)

Here, nf ree is the number of free monomers (not belonging to any micelle), Vtotal is the volume of the simulation box and Vmicelles is the volume occupied by micelles. The volume of a single micelle is estimated as the number of surfactants in a micelle times the volume of a surfactant. The volume of the surfactant is calculated from the van der Waals radius of its atoms.

Alignment and radial distribution functions The largest micelle at each frame is used for the analysis, after equilibrium is reached. To characterize the alignment of the EMIM+ ring, the angle between the normal vector to the EMIM+ ring’s plane and a position vector of the EMIM+ ring is calculated. The position vector is defined between each EMIM+ molecule ring center of mass and the closest atom on the micelle surface.

Shape The shape of the micelles is calculated using the radius of gyration tensor of each aggregate. By diagonalizing the radius of gyration tensor, the three principal radii (R1 , R2 and R3 ) are found. The radii are ordered so that R1 >R2 >R3 . When R1 >R2 ≈ R3 the aggregate has the shape of a cylinder, for a sphere R1 ≈ R2 ≈ R3 , while for a disk R1 ≈ R2 > R3 . 35

6


Page 7 of 26

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


Diffusion Coefficient The diffusion coefficient (D) is calculated by computing the mean square displacement (MSD) of molecules and the Einstein’s relation: 36 < |r(t) − r(0)|2 > . t→+∞ 6t

D = lim

(3)

Here, r(t) is the location of the center-of-mass of a molecule at time t. At long times, the motion of a molecule becomes diffusive. The diffusion coefficient is computed from the gradient of MSD in this region. To improve the statistics, multiple time origins after equilibrium are taken to compute the diffusion coefficient average and error.

Results Experiments As depicted in Table 2, there is a reduction of the diffusion coefficient for all species with an increase of SDS concentration (C). The experimental CMC is 208mM. 37 Increasing the concentration beyond the CMC causes a major diffusion coefficient drop for the ionic liquid and SDS molecules. The drop of the diffusion coefficient from C < CMC to C > CMC is 43% for EMIM+ , 53.6% for EtSO4 − and 61.6% for SDS. Increasing the concentration even further reduces the diffusion coefficient of all species but by a lesser amount. For example, a reduction of 21% is observed for SDS surfactants, 13% for EMIM+ and 16.7 % for EtSO4 − from CMC < C < 2*CMC to C > 2*CMC. Clearly, micellization causes the retardation of molecular diffusion. Interestingly, EMIM+ exhibits the fastest diffusion of all species at all concentrations. The fact that EtSO4 − diffuses faster than SDS can be explained by the difference in alkyl tail length.

7



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

Simulations To obtain a molecular understanding of why micellization causes such a drop in the diffusion coefficient of the ionic liquid species, we have analyzed from the molecular dynamics simulations the self-assembly pathway, dynamics, energetics and solvent structural order. To assess possible finite size effects, we perform simulations of 128 and 256 SDS molecules at a fixed 575mM concentration and compare the time evolution of the micelle size and fraction of free surfactants (not belonging to a micelle), as shown in Figure S3. The micelles are structurally identical in the two systems within the studied time window of 600ns. Both systems produce an average micelle size of 12-15 and fraction of free monomers of 0.5-0.6. Because the average micelle size and fraction of free monomers are similar in the two systems, we use a size of 128 SDS molecules for all the concentrations and analysis described in this manuscript. The evolution of the fraction of free surfactants (not belonging to a micelle) and the average micelle size are shown in Figure 2a and 2b for different concentrations. The system with a concentration of 52mM does not exhibit surfactant aggregation. For 575mM (blue curve) and 290mM (red curve) after 200ns the fraction of free monomers remains essentially constant at 0.5 and 0.7, while the average micelle size converges at 16 and 12 surfactants for 575mM and 290mM, respectively. The evolution of micelles sizes exhibit the same aggregation behavior for both systems, as shown in Figure 3. That is, the aggregates grow in size at the expense of the free monomers from solution up to 200ns after which the fraction of free monomers remains constant. However, the distribution of micelle sizes still keeps shifting after 200ns towards higher micelles sizes up to 420ns. This means that smaller size micelles must be fusing together resulting into bigger micelle sizes, since the fraction of free monomers does not change after 200ns. Aggregation is considered as complete and equilibrium reached when the fraction of free surfactants, average micelle size and the probability distribution of micelle sizes have stabilized. Figure 4c shows a representative micelle structure for 575mM and time=598ns. This structure exhibits the typical hydrophobic core inside the micelle with the polar headgroups facing the solvent. Interestingly, solvent-exposed hydrophobic patches 8


Page 8 of 26

Page 9 of 26

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


on the micelle’s surface are present. By using the last 100ns of the equilibrated part of the trajectory for 575mM (for 128 and 256 SDS molecules) and 290mM (128 SDS molecules), the CMC is calculated as the concentration of free surfactants (see Method section). The average CMC is found to be 244.7mM ± 71.5mM and is in quantitative agreement with the experimental value of 275mM. 37 Therefore, the system with 575mM of SDS in IL has a concentration higher than 2*CMC, for 290mM the concentration is between CMC and 2*CMC, whereas 52mM is below CMC. To characterize the overall micelle’s size and shape, we have computed the radius of gyration and shape of the largest micelle per simulation frame. As shown in Figure S4, the radius of gyration converges after 300 ns at 3.5 ± 0.2 nm for C>2*CMC and 4.5 ± 0.2 nm for CMC 2*CMC, the diffusion coefficient reduces by 34.8% for EMIM+ , 37.14% for EtSO4 − , 48.5% for SDS and 46.47 % for Na+ . Experiments show 9



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

that, with an increase in concentration, the diffusion coefficient of SDS, EMIM+ and EtSO4 − reduce a 30% less from CMC < C < 2*CMC to C> 2*CMC compared to the large drop that is observed from C < CMC to CMC < C < 2*CMC. On the other hand, in simulations, a similar diffusion coefficient drop is observed from C < CMC to CMC < C < 2*CMC and from CMC < C < 2*CMC to C > 2*CMC. In agreement with the experimental observations, EMIM+ exhibits a faster diffusion coefficient than the anion. The highest diffusion of EMIM+ can be attributed to its planar shape 40 that provides less friction compared to EtSO4 − . Interestingly, at concentrations above CMC, Na+ ions diffuses similarly to SDS molecules, suggesting counter-ion condensation. Whereas, for concentrations below CMC, the diffusion coefficient of Na+ is less than any other species, including SDS diffusion. The slower mobility of the sodium molecules could be attributed to the fact that it can bind to SDS and EtSO4 − molecules. To provide an understanding of why the diffusion coefficient for the different chemical species decrease with concentrations above CMC, the solvent structure is analyzed by means of radial distribution functions (RDFs). Only the largest micelle per frame is considered for C > 2*CMC. Figure 6 depicts the RDFs of selected micelle’s atoms and IL and Na+ molecules as a function of their distance from (a) the SDS’s sulfur atom, (b) 10% of the furthest SDS carbons from a micelle’s center-of-mass and (c) the micelle’s center of mass. A solvent layering effect is observed in Figure 6a where EMIM+ and EtSO4 − alternate layers near the SDS headgroup. The EMIM+ tail carbon atom (CE) is slightly closer than the EMIM+ ring nitrogens to the SDS headgroup, while the terminal alkyl tail carbon of EtSO4 − is substantially closer than the EtSO4 − headgroup. This observation means that most of the IL’s alkyl tails in contact with a micelle are pointing towards the micelles, which is also evident from Figures 4a and 4b. Due to the small size of the Na+ molecules, the RDF between Na+ and the SDS sulfur atoms exhibits a prominent peak at small distances. In the second coordination shell (second peak in the sodium RDF) the sodium molecules pairs with the first layer of EtSO4 − headgroups. 10


Page 10 of 26

Page 11 of 26

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


By integrating the RDF from zero to the first minima of each RDF curve, the number of molecules inside the first coordination shell of the SDS’s sulfur atoms is estimated. On average 9.8 Na+ , 164 EMIM+ ’s nitrogens (that would correspond to 82 imidazolium molecules) and 28 EtSO4 − alkyl tail (CT) atoms are observed in the first coordination shell. Since the average size of the largest micelles is 25 ± 3 SDS molecules, not all SDS molecules have a bound counterion and the degree of counterion dissociation is more than 50%. This percentage is larger than what is observed for surfactant aggregates in water from published experimental and simulation data. 6,41 One explanation of the large counterion dissociation is that Imidazolium cations can also neutralize the headgroup charge of the SDS molecules. Also, there is more charge screening in the IL system due to the large number of mobile charge carriers compared to liquid water. Therefore, from a charge screening point of view a large counterion dissociation is expected when compared to liquid water. Another interesting observation from the ions’ RDF to the SDS’s sulfur atoms (Figure 6a) is that a multi electric layer of almost 1.25nm from a micelle’s surface is formed with alternating layers of anions and cations. This alternating layer is as follows: Na+ | EMIM+ | EtSO4 − | Na+ . The presence of this layer can hinder the mobility of the IL ions. For micelles that have hydrophobic patches exposed to the solvent (as shown in Figures 4a, 4b, S5a and S5b), we see the alkyl tails of EtSO4 − and EMIM+ cover those solvent exposed hydrophobic patches since both the EtSO4 − and EMIM+ terminal carbon RDF exhibit their first peak at short distances, with respect to external alkyl carbons of SDS, shown in Figure 6b. Therefore, the alkyl tails of the IL point towards solvent exposed hydrophobic patches in micelles resulting in a lower surface energy. These alkyl tails of the solvent stabilize structures with solvent exposed hydrophobic patches as it is the case of the disklike structures observed at CMC < C < 2*CMC. As shown in Figure 6c, penetration of the IL is also observed, since the RDFs are nonzero for distances to the center-of-mass less than the location of the sulfur atoms of the SDS molecules (solid red line). The solvent penetration inside the micelle is due to the

11



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

amphiphatic nature of the IL. Figures S5a and S5b show how an IL molecule intercalates between the SDS surfactants of a micelle. It is worth mentioning, that the solvent structure around micelles for CMC

Interplay of Dynamical Properties between Ionic Liquids and Ionic

Recommend Documents