Molecular Dynamics Simulation of Pyrene Solubilized in a Sodium

Feb 23, 2012 - Department of Environmental and Chemical Engineering, Yellow River Conservancy Technical Institute, Kaifeng 475003, China. •S Support...
0 downloads 0 Views 5MB Size
Download PDF
Article pubs.acs.org/Langmuir

Molecular Dynamics Simulation of Pyrene Solubilized in a Sodium Dodecyl Sulfate Micelle Hui Yan,† Peng Cui,‡ Cheng-Bu Liu,† and Shi-Ling Yuan*,† †

Key Lab of Colloid and Interface Chemistry, Shandong University, Jinan 250100, China Department of Environmental and Chemical Engineering, Yellow River Conservancy Technical Institute, Kaifeng 475003, China

‡

S Supporting Information *

ABSTRACT: In the present work, the structural and dynamical aspects of the solubilization process of pyrene within a sodium dodecyl sulfate micelle were studied using molecular dynamics simulations. Our results showed that free pyrene as the fluorescence probe can be spontaneously solubilized into the micelle and prefers to be located in the hydrophobic core region. As the local concentration of pyrene increased, two molecular probes could enter into the core hydrophobic region and the excited dimer of pyrene molecules was formed, showing a stacking mode of π−π conjugation. Since the π−π stacking interaction between the two pyrene molecules was very weak, formation of the excimer was a dynamic process with the two pyrene molecules alternately separating and associating with each other. In this case, the two pyrene molecules were found to be mainly distributed in the palisade layer of the micelle due to the balance between the weak π−π stacking interaction and the hydrophobic interaction of probe molecules with the surfactant tails.

1. INTRODUCTION The study of amphiphilic molecules in bulk solutions has attracted wide interest over the past decades. Amphiphilic molecules are known to self-assemble into a variety of aggregate structures in aqueous solution, such as micelles, bilayers, vesicle, and lamellae.1 These aggregates are always formed in such a way that the hydrophobic parts of the molecules are assembled in the apolar interior, from where water is expelled, and the hydrophilic structural parts are in contact with water. Since there is a hydrophobic microenvironment in the solution brought about by the aggregates, the aqueous solubility of other slightly soluble nonpolar substances can be enhanced, which is known as solubilization. This solubilization plays a very important role in industrial and biological processes. In addition, the nonpolar substances can be solubilized in different regions of the aggregates, showing many unique properties. Thus, it is important and meaningful to investigate the hydrophobic microenvironments systematically in order to utilize the aggregates better.2 In light of the development of modern experimental techniques such as UV,3 Raman,4 NMR,5 and ESR spectroscopies6 in recent years, a considerable amount of experimental research has been carried out on the physicochemical properties of these aggregates. Among these techniques, the fluorescence method is very attractive because of its high sensitivity, high selectivity, and multiple choices in signals or parameters such as emission intensity, maximum wavelength, profile of an emission spectrum, anisotropy, lifetime, and even excimer or exciplex formation.7−9 For example, the fluorescence © 2012 American Chemical Society

probe method is an effective technique for estimating the properties of the aggregates or solubilizates when using UV spectroscopy.7 Fluorescence probe molecules are a large and diverse family of organic compounds, which usually have a planar conjugated structure,10 such as pyrene, diphenylhexatriene, and prodan. These molecules can emit a spectrum after excitation by UV or visible light. Fluorescence properties of these molecules are very sensitive to their surroundings, so they can be used to probe the microenvironmental changes of aggregates based on changes of fluorescence parameters.11,12 Pyrene is one of the most widely used fluorescence probe molecules. Its fluorescence spectrum can show a characteristic structure of five vibration peaks. In particular, the ratio of the third (368 nm) and first (375 nm) monomer emission intensities (i.e., Im3/Im1) of the pyrene spectrum is very sensitive to the microenvironmental polarity of the solubilized pyrene molecules.11,13 The ratio of Im3/Im1 is also influenced by the aggregation number, core cavity, core thickness of the micelles, hydrogen-bond-acceptor basicities, and coordinative covalency.14 Another important property of pyrene is its ability to form excimers and exciplexes, which emits fluorescence in the vicinity of 470 nm (Ie) and depends on the microviscosity conditions of the pyrene molecules and the occupancy numbers of the solubilized pyrene in the micelle.8,15−17 Thus, it is Received: June 1, 2011 Revised: February 22, 2012 Published: February 23, 2012 4931

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

important to understand the location of the solubilized guest molecule in the aggregates. To date, much experimental research has been concerned with the location of pyrene in various surfactant aggregates. It is generally recognized that the nonpolar derivatives solubilized in micelles are distributed in three solubilization loci: near the micelle surface, in the hydrocarbon core, and in the palisade layer. For example, Matzinger et al.18 showed that the probe molecules are distributed in the Stern layer and the hydrocarbon core in both long and short lifetimes of probe molecules in ionic trimethylammonium chloride and bromide micelles. Turro and co-workers concluded that pyrene exists in the palisade layer of ionic micelles.14 Honda et al.7 measured the fluorescence spectra of pyrene in aqueous solutions of heptaethyleneoxide monoalkyl ether as a function of surfactant concentration. They concluded that the pyrene molecules in the micelles were mainly distributed in the core region. Khan et al.19 performed a UV−vis study of pyrene in the sodium dodecyl sulfate (SDS) micelle and showed that one pyrene molecule preferentially resided inside the micellar core. Recently, Bales et al. developed an approach “probe−probe interaction” strategy to define relative positions of pyrene and spin probes in SDS micelles.20,21 Since the scale of the surfactant aggregates is usually several nanometers and coexists with the solvent, it is difficult for experimental research to show a direct microscopic view of the distribution of the probes. Recently, the molecular dynamics (MD) method has proven to be a valuable tool to study the structure of the assembly surfactants and can provide detailed, atomistic level insight into the structure of the studied systems. These kinds of studies allow us to extract information about dynamic and structural properties at a microscopic level, which is an important complement to experimental techniques. There is much literature that has reported the properties of pyrene in the membrane using computational calculations.22,23 These studies proposed that the pyrene molecule prefers to be positioned in the hydrophobic part of a membrane near the headgroup region. However, the process of pyrene solubilization within surfactant aggregates such as micelles in the solution is still poorly understood at the atomic level. In addition, there is no simulation study on this aspect to our knowledge. In this work, we employed MD simulations to study the solubilization of pyrene probes within the SDS micelle. The objective of this work was 2-fold. First, we investigated the process of one pyrene molecule solubilized into a micelle. Second, the influence of the local pyrene concentration on the solubilization was addressed in order to investigate pyrene dimer formation in a micelle. These studies were carried out through several computational models, and the details will be discussed in the following sections.

Scheme 1. Structures of (a) SDS and (b) Pyrene Molecules Used in This Study

Table 1. Summary of the Simulated Systems system label

no. of pyrene

no. of SDS

no. of water

description

time/ ns

I II III and III2 IIIa and IIIa2 IV

0 1 2

60 60 60

7356 7334 7313

free SDS micelle pyrene + free micelle pyrene + pyrene-micelle

200 200 150

2

60

7312

2 pyrene + free micelle

50

2

60

7356

50

IVa

2

60

7335

V

4

120

14785

mixed SDS monomers with 1 pyrene mixed SDS monomers with 2 pyrene 4 pyrene + 2 micelles

50 50

rectangular box containing 7356 water molecules (∼0.4 M of SDS solution). Then 60 sodium counterions were inserted into solution to keep the system electrically neutral. The simulation was carried out for 200 ns to equilibrate the system, and finally, the aggregated structure of SDS was obtained. The final density of the entire system was 1.003 g/cm3 (close to an experimental density of 1.0093 g/cm3 27). This model of the micelle solution was later used in the following pyrene− micelle systems (Systems II, III, and V). System II was performed to study the process of one pyrene molecule solubilized into a SDS micelle. To construct the simulation model, one free pyrene probe was manually inserted into the SDS micelle solution (obtained from System I) and randomly placed outside the micelle. The overlapping water molecules located within 0.3 nm around the probe were then removed. The number of the pyrene was 1, which was based on the experiments. In the experimental applications employing the fluorescence probe techniques, the concentration of the pyrene was very low to ensure the solubilized probes did not remarkably affect the structure of the aggregate. For example, the concentration of pyrene used in refs 19 and 29 was 4 × 10−6 M and 1.53 × 10−5 M, respectively. This simulation was performed for 200 ns for the single pyrene molecule to reach equilibrium with the hydrophobic environment. The single pyrene molecule finally entered into the interior of the micelle as expected (more specific details will be discussed in section 3). Since the concentration of pyrene used in the experiment was very low, the solubilizate distribution among micelles in solution is frequently described by the Poisson distribution model. It is rare to have two pyrene molecules in a SDS micelle. However, if the concentration of the probes is high enough, more probes will enter into a micelle and the excimer would be formed. To investigate the effect of local probe concentrations on the solubilization, we employed two simulation systems, named Systems III and IIIa. In these systems, two pyrene molecules were initially placed in the simulation box. The initial configuration of System III was constructed by adding one more pyrene to the final configuration obtained from System II, while System IIIa was built by adding two pyrene molecules to the final configuration of System I. The added pyrene molecules were initially placed outside the micelle in the bulk water for both Systems III and IIIa. To check whether the observations were generalized and investigate the effect of the initial positions of pyrene on the solubilization process, we performed two more simulations for Systems III and IIIa to discuss the behavior of the two pyrene

2. SIMULATION DETAILS 2.1. Model Systems. The initial coordinates for the surfactant monomer and pyrene were generated from ProDRG24 (shown in Scheme 1). Then, a series of simulations was performed to study the solubilization of pyrene in SDS micellar solution. All systems used in our study and their corresponding labels are summarized in Table 1. To give a clear description of the relationship among these systems, a flowchart is presented in Supporting Information, Figure S1. We first studied the behavior of 60 SDS monomers in pure solution at 298 K to get a SDS micelle (system I). The aggregation number of 60 is close to the experimental value of 65 at the cmc at 298 K25,26 and the same as the number used in previous simulation studies.27,28 To construct System I, SDS monomers were randomly placed in a 4932

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

molecules in the micelle, named III2 and IIIa2, respectively. In order to observe the difference between probes in micellar and premicelle solution, we studied the behavior of the probe in the premicelle solution. The model was constructed by randomly placing one or two pyrenes along with 60 SDS in a 7340 water box, named Systems IV and IVa, respectively. As is known, in a real system there are many micelles in solution; in order to elaborate on these effects on the solubilization process, we constructed a larger system containing two micelles, named System V. 2.2. Computational Details. Molecular dynamics simulations were performed using the GROMACS 4.0.5 package.30 The interatomic interactions were calculated according to the parameters and potential functions of the GROMOS 45a3 force field.31 In this force field a united-atom description was used for the SDS and pyrene molecules. Partial charges on SDS were adopted from Bruce et al.,27 while for the pyrene the atomic charges were derived from the work of Vattulainen et al.,23 who used ab initio quantum-mechanical calculation with the Hartree−Fock method and 6-31G basis set. Sodium ions were described by the Gromacs parameters. The simple point charge (SPC) model was adopted for the water molecules.32 The steepest descent method was used to minimize the energies of the initial configurations. After minimization, a brief equilibration under NPT ensemble at 1 atm and 298 K using the Berendsen barostat and thermostat33 with coupling time constants of 1.0 and 0.1 ps, respectively, was performed for each model system to make the system volume stable. Then a 50−100 ns NVT simulation was carried out at 298 K for each system to gather MD trajectories. In the simulation, bond lengths were constrained using the LINCS algorithm,34 periodic boundary conditions were applied in all directions, and a time step of 2 fs was used throughout the simulations. The nonbonded potential truncation was performed with a cutoff distance at 1.2 nm for Lennard−Jones interactions. The particle mesh Ewald method was employed for the long-range electrostatic interactions.35 During MD simulations the trajectories were stored every 100 ps. Trajectories were visualized using VMD 1.8.5.36

assembled to a micelle, as shown in Figure 2a. We note that the averaged micelle radius was about 1.95 ± 0.5 nm, and it

Figure 2. (a) Time profiles of the average distance between sulfur atoms and the center of mass (COM) of the micelle in pure surfactants solution. (b) Probability distribution in probe-free micelle of selected atoms measured with respect to COM of the micelle. Probability was defined as the number of instances the selected atom was found within a spherical shell of width 0.02 nm at a distance r from the micelle COM divided by r. (c) Optimized structure of a SDS monomer using the DFT method and a model of the pyrene-free SDS micelle. Asterisk (*) data from ref 38.

3. RESULTS AND DISCUSSION 3.1. Structure of a Free SDS Micelle. The macroscopic behavior of the pure surfactants solution was investigated by developing the molecular picture of System I. Figure 1 shows views of the configurations corresponding to the beginning and end of the simulation. We can see that the surfactants assembled into a micelle as expected. The equilibrium of the micelle was determined based on the micelle radius, Rs, which is represented by the distance between the headgroup S atom from the micelle center of mass (COM) after all surfactants

remained steady during the long MD run. Moreover, evolution of the total energy of the micelle was checked to make sure the system reached equilibrium (see the Supporting Information, Figure S2). To check the sphericity of the aggregate structure, the principle moments of inertia I1, I2, and I3 were averaged over the last 10 ns, where I1, I2, and I3 were along the x, y, and z axis, respectively. The calculation was performed using the program g_gyrate in the Gromacs package. A value for Imax/Imin of 1.08 was obtained, indicating the shape of the micelle was slightly less spherical. These results are in good agreement with previous simulations27,28 and experimental data.37 The micelle structure was also observed by the probability distribution of the selected atoms of the surfactant with respect to the COM of the micelle. The distribution profiles are shown in Figure 2b. The density profile is in good agreement with the total radius of 2.23 nm from X-ray scattering.37 From the plot one can see that the sulfur density rapidly decreases between 2.0 and 2.5 nm. The distribution of C12 shows a broad plateau and has a tail in the region of the sulfur distribution, which indicates that the terminal methyl groups reorganized toward the micelle surface and the tails of the surfactants were bent inside the micelle. The high probability of C12 began to occur at about 0.7 nm, which revealed that there was a cavity in the interior of the micelle. In our earlier work,38 the structure of a single SDS monomer was optimized at the B3LYP/6-31+G* level, and the distance between sulfur and the terminal methyl was found to be 1.66 nm. This also supports the standpoint that the tails of the monomer are bent in the micelle. On the basis of

Figure 1. Views of the configurations in pure SDS solution system at the (a) beginning and (b) end of the simulation. Sulfur, oxygen, and carbon atoms that belong to the SDS molecule are shown as yellow, red, and gray spheres, respectively. Water molecules are represented by red and white lines. Na+ ions are shown as little pink dots (which may be not very clear in this figure). 4933

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

the above analysis, a schematic representation of the free micelle is presented in Figure 2c. This aggregation structure of SDS was used to solubilize one or two pyrene molecules as a preassembled micelle for the following simulation systems (Systems II and III series). 3.2. One Pyrene Solubilized in the Micelle. Equilibration of System II was monitored through the LJ potential between the SDS micelle and the pyrene molecule (see Supporting Information, Figure S3). The figure shows that about 20 ns were required for the system to reach a stable state, which we denote as the adsorption interval. Visual inspection of the initial (at 0 ns) and final views (at 200 ns) indicated that the probe molecule was solubilized into the micelle hydrophobic region, as shown in Figure 3. The solubilizing process of the pyrene into the micelle can be

Figure 4. (a) Time evolution of the distance between COMs of the pyrene molecule and SDS micelle. (b) Probability distribution of selected atoms and pyrene with respect to micelle COM averaged over the last 50 ns. Probabilities for pyrene are calculated by its COM. (c) Model of the location of one pyrene in the SDS micelle.

Figure 3. Views of the configurations of System II at the (a) beginning (at 0 ns) and (b) end (at 200 ns) of the simulation. Carbon atoms of the pyrene molecule are shown as green spheres. Color scheme of SDS atoms and Na+ is the same as in Figure 1. Water is omitted for clarity.

hydrophobic interior the aggregation size of the surfactants was slightly influenced. It is mainly due to the loose packing of the surfactants tails, leaving much freedom for the probe to reside. The highest probability of the pyrene distribution extends from 0.3 to 0.7 nm, which reveals that one pyrene distributes within the core region and is encapsulated by the surfactants around it. The longest diameter of the planar molecule pyrene (the distance between C8 and C15 shown in Scheme 1) is about 0.7 nm. Thus, it is possible for the pyrene molecule to be involved in this core region of the micelle. The hydrophobic interaction between pyrene and alkyl chains of surfactant is the important reason for the location of pyrene in the micelle. In Figure 4c, a model is presented to illustrate the distribution of the pyrene in the SDS micelle. The overlapping of the probe and C12 distribution also indicates the translational freedom of the terminal groups. To estimate the orientation of the pyrene molecule when it was adsorbed onto the interface micelle surface, we plotted cos θ as a function of the distance of the pyrene molecule from the COM of the micelle, where θ is the mean angle between the vectors Mpyrene→ C8/C15 (Mpyrene represents the COM of pyrene) and the tangent of the micelle sphere, as shown in Figure 5a. It should be noted that the tangent point is along the radial direction of the micelle COM to Mpyrene. When the value of cos θ is 1, it means that the plane face of pyrene is parallel to the surface of the micelle, while when the value is 0, the probe is perpendicular to the aggregation surface. Figure 5c shows that the preferential orientation of the probe as a function of distance from the COM of the SDS micelle. When the pyrene molecule is far away from the micelle COM, the orientation with respect to the micelle drastically changes. This indicates that as the probe molecule approaches the micelle surface it rotates in the bulk water. Since the radius of the micelle is about 1.95 nm, we mainly focused on the probe orientation when it

observed by checking the MD trajectories. To illustrate this process, the distance between the COMs of pyrene and micelle with time evolution was calculated, as shown in Figure 4a. The pyrene molecule gradually approached the micelle surface, due to its hydrophobic property. It is known that polycyclic hydrocarbons are weak bases showing an affinity with protons and subsequently for positively charged groups. Once the pyrene molecule contacts the surface, pyrene will enter into the micelle because of the long-range electrostatic interaction. Because the outer surface of the palisade layer of the SDS micelle is a negatively charged sphere, the π electron cloud of pyrene is repelled by the negative sulfate. Hence, the pyrene will easily pass through the surface and penetrate into the inner palisade layer of the micelle, which is defined as the region between the surface of the micelle and the micellar core. Our simulation shows that after long equilibrium MD runs (about 100 ns) the pyrene molecule penetrates into the core region of the micelle. The distance between the COMs of pyrene and the micelle remained steady during the last 50 ns MD run as shown in Figure 4a. The fluctuations of the COM distances indicated that the probe molecule was still quite mobile in the hydrophobic environment. The average distribution probabilities of pyrene COM and the selected atoms of the surfactant with respect to the COM of the micelle were calculated in order to get a more detailed picture of the probe molecule location. The probability distribution functions of pyrene, C12, C6, C1, and S atoms are plotted in Figure 4b. Compared with Figure 2b, one can note the distribution of C12 and S atoms is almost unchanged. This indicates that after the probe molecule was involved in the 4934

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

core. Figure 5b shows the view of the probe about to enter into the micelle (at about 3 ns). 3.3. Two Pyrene Molecules Solubilized in the Micelle. The System III series was performed to investigate the effect of the local pyrene concentration on solubilization. We first discuss the results obtained from System III. The convergence was first established. The LJ potential between the SDS micelle and the pyrene molecules was also checked. The quantity is presented in Supporting Information Figure S4. These results show that the equilibrium structural properties of the system can be reliably determined from the last 40 ns simulation. Visual inspection of the views at three different simulation times shown in Figure 6 indicates that the second pyrene can be adsorbed onto the mixed micelle surface, and then it can be solubilized into the hydrophobic interior. From the final view at 150 ns, it can be seen that the two pyrene molecules are located in the hydrophobic region of the micelle. The average distribution probabilities of the two pyrene COMs and the selected atoms of the surfactant with respect to the COM of the micelle were also calculated as shown in Figure 7a. The probe molecule which is far away from the micelle at the beginning of the simulation is named pyrene 1, while the other one that is in the core region of the micelle is named pyrene 2. We note the distribution of probes mainly in the region of 1.0 −1.5 nm with respect to the micelle COM. It means the probes are located at the palisade layer of the micelle. Compared with the distribution of probe in the one pyrene−micelle system (shown in Figure 4b), it was found that the probe molecules move toward the interface from the deeper region of the micelle. To investigate the behavior of the two probe molecules in the micelle we plotted their distance from the micelle COM as a function of time in Figure 7b. We found that pyrene 1 approaches the micelle interface, while pyrene 2 stays in the deeper region of the micelle during the first several nanoseconds of the simulation. When pyrene 1 proceeds near the interface of the micelle at about 4 ns, pyrene 2 begins to move toward the micelle surface. It is probable that the two pyrene molecules interact with each other. Then the two pyrene molecules are both at a position about 1.2 nm from the micelle COM. During the long MD run, the distances almost remained steady, which also means the system was well equilibrated. Another interesting feature is that during the equilibrium process when one pyrene moves closer to the micelle COM then the other one moves toward the interface. In Figure 7b, the labels A, B, and C, which represent different simulation times, show this phenomenon. It is possibly due to the

Figure 5. (a) Scheme of the mean angle between the vectors Mpyrene→ C8/C15 (Mpyrene represents the COM of pyrene) and the tangent of the micelle sphere. (b) View of probe about to enter into the micelle (at about 3 ns). (c) Probe orientations as a function of distance from the micelle COM.

was located at 1.8−2.0 nm, which is the region for pyrene being about to insert into the hydrophobic part of the micelle. The plots show that the value of cos θ is close to zero when the probe is adsorbed onto the micelle surface. This indicates that the pyrene molecule approaches close to the micelle with the planar face of the molecule perpendicular to the surface of the micelle. As discussed above, with one pyrene molecule solubilizated into the micelle, the aggregation size of the surfactants was slightly influenced and the calculated moments of inertia of the micelle show that the micelle is an ellipsoid but not a normal sphere. Therefore, our statistics may have a systematic deviation. However, they can still reveal the orientation of the probe when it is entering into the micelle

Figure 6. Views of the configurations at different times of the simulation: (a) initial (at 0 ns), (b) intermediate (at 90 ns), and (c) final (at 150 ns). All other details are identical to those in Figure 3. 4935

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

Figure 7. (a) Probability distribution of selected atoms and two pyrene molecules with respect to micelle COM averaged over the last 30 ns. Probabilities for pyrene are calculated by its COM. (b) Time evolution of distance between each pyrene molecule and micelle COM. (c) Model of the location of pyrene dimer in the SDS micelle.

Figure 8. Potential of mean force between pyrene molecule and micelle: (a) System II and (b) System III.

interaction or intermolecular structure between the two pyrene molecules. The free energy profile of the penetration of the pyrene molecule into a micelle was estimated using the potential of mean force (PMF) between a pyrene molecule and the micelle, which was calculated by their pair distribution function. The profiles are shown in Figure 8. We note that the global minimum of the energy is located in the core region of the micelle for the one pyrene system (System II), while for the two pyrene system (System III), it is located at about 1.2 nm, which corresponds to the palisade layer of the micelle. This explains why a significant population of probe molecules has been solubilized in different regions. In the two probes system, the two pyrene molecules become a conjugate structure. Then equilibrium between the weak π−π stacking interaction and the hydrophobic interaction of probe molecules with the surfactant tails contributes to the whole PMF and make them locate at the outer region of the micelle. On the other hand, the presence of large, rigid aromatic π stacking conjugation in the middle of the micelle would force the lipid acyl chains to arrange themselves around the probes. It would reduce the mobility of the chain ends, resulting in the reduction of micelle entropy, which is highly unfavorable. To characterize the intermolecular structure of the two probes, we calculated the distance and orientation between the two pyrene molecules as a function of time as shown in Figure 9. The orientation between the two probes is defined by the angle α between the two polycyclic planes in each molecule. From Figure 9a, one can note that after the equilibrium MD run (about 10 ns) the distance between the two probes occasionally changes and the closest separation of the two probes is 0.42 nm. This reflects that the two pyrene molecules irregularly and intermittently separate from each other. The oscillation of the distance also shows the mobility of the probes

Figure 9. (a) Distance and (b) orientation between the two pyrene molecules as a function of time. (c) Views of two pyrene molecules selected from the MD trajectories at 45 ns.

in the internal micelle. When the two probes are in the closest separation, it can be found that the angle between the polycyclic planes is 0° or 180°, as shown in Figure 9b. The polycyclic planes are parallel with each other, showing a stacking mode of π−π conjugation. A view of the two probes at 45 ns is presented to show the π−π interactions between the two aromatic molecules in Figure 9c. As is known, one of the most important properties of pyrene is its ability to form excimers and exciplexes which yield a structureless emission 4936

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

Figure 10. Views of the configurations at different times of the simulation. All other details are identical to those in Figure 3.

band within the visible range15,39 when the local concentration of pyrene is high. In the present work, formation of the excimer was successfully observed, that is, a pyrene molecule can bind to another pyrene molecule to form an excited excimer with a π−π stacking mode. Since the π−π stacking interaction is very weak, the existence of the excimer is a dynamic process, alternatively showing the separation and association of the two pyrene molecules with time evolution. In Figure 7c, the location of the excimer is presented using a cartoon model. For Systems IIIa, III2, and IIIa2, some results including the views of the simulation with time evolution, the distance between pyrene and micelle COM or another pyrene, and the distribution probability of pyrene with respect to the micelle COM are provided in the Supporting Information, Figures S5− S7. The observations from System V are provided in Figures S8 and S9 in the Supporting Information. These results are very similar to those derived from model III. In summary, both pyrene molecules can be spontaneously solubilized into the micelle core region. In the interior of the micelle, two probes can form excimers with a separation of 0.42 nm, showing a stacking mode of π−π conjugation. The two pyrene molecules mainly distribute in the palisade layer between the core region and the micelle surface. 3.4. Pyrene in Premicelle Systems. In previous research measuring the fluorescence spectra of pyrene in surfactant aqueous solutions as a function of surfactant concentration, a sudden jump of Im3/Im1 could be observed as soon as the concentration reached the cmc value, which is thought to be related to the rearrangements of the pyrene taking place in the premicelle.6 As the concentration reaches cmc, the micelle began to form and the probe could be solubilized into it. Then the florescence spectra of the first to third vibronic ratio of the florescence band of pyrene can be observed. We constructed systems of probe molecules with nonaggregated surfactants, i.e., Systems IV and IVa, to verify the results from the above systems. The surfactants were randomly placed in the box, which corresponds to the concentration that just reached the

cmc value. Here, we only provide the results of two pyrene molecules in the premicellar system (System IVa). In Figure 10, views of the simulation at different periods are shown. Initially, the surfactants are in a homogeneous mixture along with the probes. As soon as part of the surfactants aggregate, the probe molecule is partitioned from the aqueous phase to the hydrophobic environment (at 5 ns as shown in Figure 10). After all 60 monomers aggregate to form a micelle, rearrangement of probes in the hydration region takes place in the micelle. The π−π conjugation structure of the two pyrene molecules can also be observed in this simulation model. The location of the probes in the inner micelle is found to be similar to those of Systems III and IIIa (see Supporting Information, Figure S10). Thus, we conclude that when the local concentration of surfactants or probes is constant, solubilization of two probe molecules in the micelle is nearly identical. The results of one pyrene molecules in the premicellar system (System IV) are provided in Figure S11 in the Supporting Information. We note that the equilibrated structure is also very similar to that of model II; the only difference is that the location of the pyrene is deeper in the core region, as shown in Figure S11(b), Supporting Information. In this case, the probe is distributed in the range of 0−0.5 nm, which corresponds to the core cavity of the micelle. It is mainly due to the progressive aggregation of the surfactants. As a portion of surfactants aggregates and forms a hydrophobic microenvironment in aqueous solution, the pyrene molecule tends to be solubilized in this environment. In contrast to the process in which the pyrene molecule solubilized into a preassembled micelle, there was more freedom for the probe molecule’s mobility in the aggregation structure, that is, after the aggregation number of surfactants increases to 60 forming a much more compact packing in the interior of micelle, it is more difficult for the probe to enter into the core cavity. 4937

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938

Langmuir

Article

(5) Lianos, P.; Viriot, M. L.; Zana, R. J. Phys. Chem. 1984, 88, 1098. (6) Szajdzinska-Pietek, E.; Sulak, K.; Dragutan, I.; Schlick, S. J. Colloid Interface Sci. 2007, 312, 405. (7) Honda, C.; Itagaki, M.; Takeda, R.; Endo, K. Langmuir 2002, 18, 1999. (8) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer/Plenum: New York, 1999. (9) Liu, J.; Fang, Y.; Chen, C. L. Langmuir 2008, 24, 1853. (10) Goodpaster, J. V.; Harrison, J. F.; McGuffin, V. L. J. Phys. Chem. A 2002, 106, 10645. (11) Kalyanasundaram, K.; Thomas, J. K. J. Am. Chem. Soc. 1977, 99, 2039. (12) Khatua, D.; Dey, J. J. Phys. Chem. B 2007, 111, 124. (13) Nakajima, A. Bull. Chem. Soc. Jpn. 1971, 44, 3272. (14) Turro, N. J.; Kuo, P.-L. J. Phys. Chem. 1986, 90, 420. (15) Vekshin, N. L. Anal. Chim. Acta 1989, 227, 267. (16) Pownall, H. J.; Smith., L. C. J. Am. Chem. Soc. 1973, 95, 3136. (17) Zana, R. J. Phys. Chem. B 1999, 103, 9117. (18) Matzinger, S.; Hussy, D. M.; Fayer, M. D. J. Phys. Chem. B 1998, 102, 7216. (19) Khan, A. M.; Shah, S. S. J. Dispersion Sci. Technol. 2008, 29, 1401. (20) Lebedeva, N.; Bales, B. L . J. Phys. Chem. B 2006, 110, 9791. (21) Lebedeva, N.; Ranganathan, R.; Bales, B. L . J. Phys. Chem. B 2007, 111, 5781. (22) Hoff, B.; Strandberg, E.; Ulrich, A. S.; Tieleman, D. P.; Posten, C. Biophys. J. 2005, 88, 1818. (23) Č urdová, J.; Č apková, P.; Plásě k, J.; Repáková, J.; Vattulainen, I. J. Phys. Chem. B 2007, 111, 3640. (24) Schuettelkopf, A. W.; van Aalten, D. M. F. Acta Crystallogr., Sect D: Biol. Crystallogr. 2004, 60, 1355. (25) Bales, B. L.; Messina, L.; Vidal, A.; Peric, M.; Nascimento, O. R. J. Phys. Chem. B 1998, 102, 10347. (26) Alargova, R. G.; Kochijashky, I. I.; Sierra, M. L.; Zana, R. Langmuir 1998, 14, 5412. (27) Bruce, C. D.; Berkowitz, M. L.; Perera, L.; Forbes, M. D. E. J. Phys. Chem. B 2002, 106, 3788. (28) Shang, B. Z.; Wang, Z. W.; Larson, R. G. J. Phys. Chem. B 2008, 112, 2888. (29) Huang, J. B.; Zhao, G. X. Colloid Polym. Sci. 1996, 274, 747. (30) Spoel, D. V.; van Buuren, A. R.; Apol, E.; Meulenhoff, P. J.; Tieleman, D. P.; Sijbers, A.; Feenstra, K. Gromacs User Manual, version 4.0; Gromacs: Groningen, The Netherlands, 2009; www.Gromacs.org. (31) Schuler, L. D.; Daura, X.; van Gunsteren, W. F. J. Comput. Chem. 2001, 22, 1205. (32) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.; Hermans, J. Intermolecular Forces; Reidel: Dordrecht, The Netherlands, 1981; p 331. (33) Berendsen, H. J. C.; Postma, J. P. M; van Gunsteren, W. F.; DiNola, A.; Haak, J. R. J. Chem. Phys. 1984, 81, 3684. (34) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. J. Comput. Chem. 1997, 18, 1463. (35) Essman, U.; Perela, L.; Berkowitz, M. L.; Darden, T.; Lee, H.; Pedersen, L. G. J. Chem. Phys. 1995, 103, 8577. (36) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graphics 1996, 14, 33. (37) Itri, R.; Amaral, L. Q. J. Phys. Chem. 1991, 95, 423. (38) Yan, H.; Yuan, S. L.; Xu, G. Y.; Liu, C. B. Langmuir 2010, 26, 10448. (39) Kim, J.-H.; Domach, M. M.; Tilton, R. D. Langmuir 2000, 16, 10037.

4. CONCLUSIONS We employed a molecular dynamics simulation to study the solubilization process of pyrene within a SDS micelle solution. With the MD simulation at NPT ensemble the macroscopic density of the SDS aqueous solution has been reproduced. Micellar geometrical parameters such as the total radius of the micellar aggregate and the ratios of the principle moments of inertia of the micelle were found to be in fairly good agreement with the experimental and the previous simulation values. MD simulations reveal that one pyrene molecule can be spontaneously solubilized within the interior of the micelle from the aqueous phase. When the pyrene is adsorbed onto the micelle surface, the molecular geometric plane of the pyrene is vertical to the micellar interface. Then the probe is located at the hydrophobic core region of the micelle. As there are two probe molecules in the SDS solution which represents an increase of the local concentration of pyrene, we find that the two probes are distributed in the palisade layer of the micelle, which is defined as the region between the surface of the micelle and the micelle core. In this case, the configuration of the excited excimer is found, formed by one pyrene binding to the other one with the closest separation of 0.42 nm and a π−π conjugation stacking mode. The observed behavior of the two pyrene molecules suggests that the two probe molecules alternately separate and associate with each other through the simulation. It is mainly due to the equilibrium between the weak π−π stacking interaction and the hydrophobic interaction of probe molecules with the surfactant tails. To investigate the behavior of probes in the premicellar solution, two simulations of one or two pyrene and surfactant monomers mixture were performed, which yielded similar results to the solubilization simulation.

■

ASSOCIATED CONTENT

* Supporting Information S

Flowchart sketch of the simulation systems; LJ potential between SDS micelle and pyrene molecule; results obtained from System III series; distribution of the two pyrene molecules in the micelle for System IVa; results obtained from System IV. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by the PetroChina Innovation Foundation, the National Science Foundation (21173128), and the National Basic Research program (2009CB930104) of China. Thanks to Dr. Edward C. Mignot, Shandong University, for linguistic advice.

■

REFERENCES

(1) Hamley, I. W. Introduction to Soft Matter (rev. ed.); J. Wiley: Chichester, 2007. (2) Kogan, A.; Garti, N. Adv. Colloid Interface Sci. 2006, 123, 369. (3) Eriksson, J. C.; Gilberg, G. Acta Chem. Scand. 1966, 20, 2019. (4) Zachariasse, K. A.; van Phuc, N.; Kozankiowicz, B. J. Phys. Chem. 1981, 85, 2676. 4938

dx.doi.org/10.1021/la300146s | Langmuir 2012, 28, 4931−4938