Article pubs.acs.org/JPCB
Mesoscale Phenomena in Ternary Solutions of Tertiary Butyl Alcohol, Water, and Propylene Oxide Deepa Subramanian,†,‡,¶ Jeffery B. Klauda,† Peter J. Collings,§ and Mikhail A. Anisimov*,†,‡ †
Department of Chemical and Biomolecular Engineering and ‡Institute for Physical Science and Technology, University of Maryland, College Park, Maryland 20742, United States § Department of Physics and Astronomy, Swarthmore College, Swarthmore, Pennsylvania 19081, United States ABSTRACT: The phase behavior and mesoscopic inhomogeneities in the ternary system of tertiary butyl alcohol (TBA), water, and propylene oxide (PO) have been studied by static and dynamic light scattering, gas chromatography, mass spectrometry, and molecular dynamics simulations. Mesoscale inhomogeneities are observed in this system in a broad range of PO concentrations, from 0.02 to about 65 mass %, and at certain TBA/water mass ratios varying from about 3/97 to about 30/ 70. This TBA/water composition domain corresponds to a region where short-lived micelle-like molecular clustering and thermodynamic anomalies are observed in the TBA−water binary system. At dilute PO concentrations (0.02 to about 1 mass %) the mesoscale inhomogeneities are Brownian diffusive droplets, with a size of the order of a hundred nanometers. We hypothesize that these droplets have a hydrophobic core enriched by oily impurities and oligomerized PO molecules. A hydrogen-bonded layer of TBA, water, and PO molecules surrounds this hydrophobic core. At high PO concentrations (beyond 50 mass %), the interfacial curvature of the mesoscopic inhomogeneities changes its sign and the exteriors of these inhomogeneities become hydrophobic. The inversion of the internal curvature may result in the formation of a spongelike bicontinuous mesoscale structure at intermediate PO concentrations. This mesostructure appears to be at a nonequilibrium state, although extremely long-lived.
1. INTRODUCTION
clusters have also been detected by neutron spin echo experiments.15 In addition to molecular-scale clustering, a large amount of experimental work on aqueous solutions of small amphiphilic molecules has shown the presence of mesoscale (order of a hundred nanometers) inhomogeneities.17−32 Mesoscale inhomogeneities have been observed in aqueous solutions of ionic and nonionic compounds, such as ammonium sulfate, tetrahydrofuran, tertiary butyl alcohol, and 2-butoxyethanol, as well as in aqueous solutions of some polymers, such as poly(ethylene oxide)33−42 and polyelectrolytes, such as poly(Llysine), poly(methacrylic acid), t-RNA, and DNA.43−49 The origin and nature of the mesoscale inhomogeneities have been discussed in the literature over the past decade and still remain highly debatable and controversial subjects (see, for example, the most recent Faraday Discussions on Mesostructures and Dynamics in Liquids and Solutions50). Debates on the existence of mesoscale inhomogeneities in aqueous solutions of small molecules have led to a variety of different explanations. Initially, the phenomenon was attributed to a “structural phase transition”.20 Another interpretation explains the origin of the inhomogeneities as due to the presence of impurities, thus rendering this effect a ternary system
Small amphiphilic molecules that increase the solubility of sparingly soluble compounds in water are known as hydrotropes.1 Knowledge of the interactions between hydrotrope and water molecules is fundamental to understanding the role of water in chemical and biological systems. Aqueous solutions of hydrotropes such as alcohols, amines, ethers, and oxides play a dominant role in various fields ranging from medicine to materials science. Aqueous solutions of certain nonionic hydrotropes, such as ethanol, n-propanol, isopropanol, tertiary butyl alcohol, 2butoxyethanol, 3-methylpyridine, and 1,4-dioxane, over a concentration range of 1 to about 10 mol %, show striking anomalies in their thermodynamic properties.2−9 These anomalies are enhanced as the temperature is lowered. In the same concentration range, molecular dynamics (MD) simulations have shown the presence of dynamic molecular clustering, which can be viewed as “micelle-like” structural fluctuations.9−16 The structural fluctuations seem to be responsible for the thermodynamic anomalies in these aqueous solutions.9,16 Recent work has also shown that the structural fluctuations are caused by the transient formation and breakage of hydrogen bonds between water and small nonionic amphiphilic molecules.16 These fluctuations are short-lived molecular clusters, with a lifetime of hundreds of picoseconds and short-ranged with a length scale of a nanometer.16 Such © 2014 American Chemical Society
Received: December 20, 2013 Revised: May 12, 2014 Published: May 13, 2014 5994
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
phenomenon rather than that of a binary system.25−27 The mesoscale inhomogeneities have also been associated with the formation of loose supramolecular structures formed due to strong solute−solvent interactions28−30 or electrostatic interactions44−46 in the case of ionic species. In the case of watersoluble polymers, such as poly(ethylene oxide), this effect has been attributed to the interactions between the chain-end groups of the polymer.41 The presence of clathrate-hydrate precursors,22−24 a possible precrystallization phenomenon,37 and the possibility of kinetically arrested gaseous nanobubbles31 have also been discussed and contended. Investigations of mesoscale inhomogeneities further indicate that very slow kinetics plays an important role in these phenomena.26 In some cases, especially for low-molecular-weight salts, such as sodium bromide, the inhomogeneities were shown to be at a nonequilibrium state, dissipating after a few hours of equilibration.18 The spectrum of the different systems, effects, and explanations suggests that one might be dealing with multiple mesoscale phenomena.50−52 Among the various aqueous solutions that exhibit mesoscale inhomogeneities, we have focused our work on elucidating the mesoscopic properties of aqueous solutions of tertiary butyl alcohol (TBA). TBA is a “perfect” hydrotrope, which symmetrically aligns itself at the oil−water interface, such that the methyl groups are fully in the oil-rich phase while the hydroxyl group is fully in the water-rich phase.53 TBA is the highest-molecular-weight alcohol to be completely miscible with water under ambient conditions.54 Aqueous solutions of TBA have faced their fair share of controversies regarding the presence, or absence, of mesoscopic inhomogeneities.20−27 Mesoscopic inhomogeneities have been commonly observed in aqueous solutions of commercial samples of TBA (even of relatively high purity, such as 99.8%) in the concentration range from 1 to 8 mol % (or from 3 to 30 mass %) TBA. The inhomogeneities become more pronounced below room temperature and practically disappear at high temperatures (at 40 or 50 °C). It has been proven that these mesoscopic inhomogeneities are Brownian diffusive droplets, with a size on the order of a hundred nanometers.26,27 Remarkably, the mesoscopic droplets occur in the same concentration range where short-lived “micelle-like” structural fluctuations and thermodynamic anomalies are observed in binary TBA−water solutions.9,16,26,27,55,56 The mesoscopic droplets are observed to be unusually long-lived, stable over a year or longer.55 In our earlier work,27 we showed that the mesoscopic droplets could be eliminated from aqueous TBA solutions (25 mass % TBA, sample 1 from Table 1) by multiple filtrations through a 20 nm pore filter at about 5−8 °C (a step referred to as “cold filtration”). Furthermore, we showed that adding trace amounts of propylene oxide (PO) (0.02 mass %) to the coldfiltered aqueous TBA solution (sample 2 from Table 1) led to the reemergence of mesoscopic droplets.27 We thus concluded that the mesoscale inhomogeneities in aqueous solutions of TBA are associated with a third, more hydrophobic compound, i.e., an impurity that could be present in commercial TBA samples. At that time we thought that propylene oxide, being a coproduct in the manufacture of TBA and exhibiting partial miscibility with water (hence being more hydrophobic than TBA), could represent such an impurity. Our further, more elaborate experiments16,56 with a truly hydrophobic compound, cyclohexane, unambiguously demonstrated that the addition of trace amounts of cyclohexane (0.1 mass %) to an aqueous solution of TBA (25 mass % TBA) indeed led to formation of
Table 1. Concentrations (in Mass Fractions) and Properties of Ternary Samples Studied TBA (mass %)
water (mass %)
PO (mass %)
refractive index
1
24.49
75.51
0.00
1.3630
2
24.48
75.50
0.02
1.3630
3
14.11
85.75
0.14
1.3630
4
19.30
67.80
12.90
1.3593
5
25.00
49.10
25.90
1.3649
6
27.15
72.85
0.00
1.3630
7
5.40
56.90
37.70
1.3601
8
3.90
43.50
52.70
1.3631
9 10
2.50 24.91
34.50 73.99
63.00 1.09
1.366a 1.3630
11 12
4.0 4.0
43.0 48.1
53.0 47.9
c 1.3631
viscosity (Pa·s) 6.30 (8.5 °C) 6.30 (8.5 °C) 6.30 (8.5 °C) 3.69 (10 °C) 3.32 (10 °C) 6.30 (8.5 °C) 2.13 (10 °C) 1.82 (10 °C) 3.12 (25 °C) 0.33b 6.30 (8.5 °C) c 1.82 (10 °C)
a For organic phase assumed as pure PO. b25 °C, for organic phase assumed as pure PO. cAssumed to be the same as that of sample 8.
long-lived mesoscopic droplets. Moreover, experiments have shown that the mesoscopic droplets observed in TBA−water− cyclohexane solutions have a hydrophilic shell, while simulations have demonstrated that the core is hydrophobic.16 Simulations have further shown that the shell consists of multilayer hydrogen-bonded network of TBA and water molecules.16 The structure of the hydrophilic shell is reminiscent of the dynamic micelle-like clustering found in TBA−water binary solutions, but now stabilized by a hydrophobic core. We call this effect “mesoscale solubilization”, a phenomenon intermediate between molecular solubility and macroscopic phase separation.56 Coming back to our initial experiments with PO27 an important question remained unanswered. PO, unlike cyclohexane and other hydrocarbons, is a relatively weak hydrophobic compound; it is soluble in water up to 40 mass %.57 How can one justify mesoscale aggregation of PO molecules when PO concentration is about 1000 times lower than the molecular solubility? Most recently, Sedlák and Rak have published a systematic study of mesoscale inhomogeneities in ternary solutions of TBA−water−PO.58 They did not observe mesoscale inhomogeneities in aqueous TBA solutions upon addition of small amounts of PO (0.02 mass %). Sedlák and Rak thus concluded that hydrophobic impurities present in PO, or in TBA, could be the cause of the mesoscale droplets observed in our earlier experiments.58 In view of the new findings16,56,58 in this important, yet controversial, field, we have decided to revisit the TBA−water− PO system by combining light scattering, gas chromatography, mass spectrometry, and MD simulations. We have observed mesoscale inhomogeneities in TBA−water−PO solutions in a broad range of PO concentrations, from 0.02 to about 65 mass %. However, the domain of mesoscale inhomogeneities is limited by the TBA/water ratio of about 3/97 to about 30/70 5995
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
scattering experiments were carried out in the macroscopic one-phase region close to the binodal curve. The point on the binodal curve closest to the sample that showed the maximum light scattering intensity was interpreted as the critical point. 2.3. Light Scattering. Static and dynamic light-scattering experiments were carried out across various concentrations of the TBA−water−PO ternary system to characterize the mesoscopic properties. Light-scattering experiments were performed at two different locations, University of Maryland, College Park, and Swarthmore College, Swarthmore, PA. Light-scattering experiments at the University of Maryland were performed with a PhotoCor Instruments setup, as described in ref 26. The temperature was controlled within ±0.1 °C. At Swarthmore College, the dynamic light-scattering measurements were carried out with a Brookhaven Instruments BI-9000AT setup equipped with a TurboCorr digital correlator. For a single-exponentially decaying relaxation process, the intensity autocorrelation function g2(t) (obtained in the homodyning mode) is given as60,61
on a mass basis (or from about 1/99 to 9/91 TBA/water on a molar basis). These “magic” ratios correspond to the domain where TBA−water binary solutions exhibit thermodynamic anomalies and demonstrate dynamic molecular clustering.9,16 We have also shown that the inhomogeneities that occur at large PO concentrations (above 50 mass % PO) have a hydrophobic exterior. At very low PO concentrations (between 0.02 and 1 mass % PO) the number of mesoscale inhomogeneities strongly depends on the source of TBA. This indeed supports the conclusion of Sedlák and Rak58 that PO alone, at least at small concentrations, cannot generate mesoscale inhomogeneities. However, based on the results of MD simulations, we conclude that PO, together with TBA and hydrophobic impurities, is actively involved in the formation of the mesoscale inhomogeneities, especially at higher PO concentrations. We hypothesize that the interaction between the mesoscale inhomogeneities at intermediate PO concentrations (between 1 and about 50 mass % PO) may lead to the formation of a bicontinuous spongelike structure with interpenetrating waterrich and water-poor domains.
⎡ t⎤ g2(t ) − 1 = A exp⎢ −2 ⎥ ⎣ τ⎦
2. EXPERIMENTAL AND COMPUTATIONAL METHODS 2.1. Materials. Experiments at the University of Maryland were performed with various sources of TBA. TBA with a labeled purity of 0.990+ (Sigma-Aldrich), 0.997+ (SigmaAldrich), and 0.998+ (Alfa Aesar) were used. Propylene oxide (PO) with a labeled purity of 0.995+ was purchased from Sigma-Aldrich. Deionized water was obtained from a Millipore setup. Experiments at Swarthmore College were performed with TBA of 0.998+ purity, PO of 0.995+ purity, and deionized water from a Millipore setup. Samples were prepared first by filtering the binary TBA− water solutions with 200 nm Nylon filters to remove dust particles. If the TBA−water solution showed the presence of mesoscopic droplets, an additional filtration with 20 nm Anopore filters was carried out under cold conditions (at about 5 °C). This filtration was repeated several times until the mesoscopic droplets were apparently eliminated from the binary solution (this was checked by monitoring the coldfiltered sample in the light scattering instrument for about 24 h). PO was then added to this “clean” solution. PO was used without any filtration because of its high volatility (boiling point ∼34 °C). All experimental measurements were performed after equilibrating the samples for about 24 h. The refractive index was measured with an Abbe refractometer. The viscosity of the samples was measured with an Ubbelohde viscometer in a temperature-controlled (±0.2 °C) water bath. The concentrations, along with refractive index and viscosity values of the various characteristic samples studied, are summarized in Table 1. 2.2. Phase Diagram. The macroscopic ternary phase diagram of the TBA−water−PO system was determined by the cloud-point method.59 The third component was added (in small steps) to binary solutions consisting of the other two components. At each step, the ternary mixture was manually shaken and then allowed to rest for about 3−5 min in a temperature-controlled (±0.1 °C) water bath. The sample was then visually observed to determine if the phase separation had occurred. If not, more of the third component was added and the above procedure was repeated. Macroscopic phase separation was determined at two temperatures: 25 and 10 °C. In order to estimate the location of the critical point, light-
(1)
where A is the amplitude of the relaxation process, t is the “lag” (or “delay”) time of photon correlations, and τ is the characteristic relaxation time. For a diffusive relaxation process, the decay time is related to the diffusion coefficient, D, as60,61
τ=
1 Dq2
(2)
where q is the difference in the wave vectors between the incident and scattered light, q = (4πn/λ) sin(θ/2), n is the refractive index of the solvent, λ is the wavelength of the incident light in vacuum (λ = 633 nm for a He−Ne laser), and θ is the scattering angle. For monodisperse, noninteracting, spherical Brownian particles, the hydrodynamic radius R can be calculated with the Stokes−Einstein relation60,61
R=
kBT 6πηD
(3)
where kB is Boltzmann’s constant, T is the temperature, and η is the viscosity of the medium. 2.4. Gas Chromatography. Gas chromatography (GC) experiments were performed on the Shimadzu GC-17A, version 3 instrument. The Restek RTX-5 GC column was used. Methanol, HPLC grade, of 0.999+ purity, was used as a solvent for GC. An aqueous TBA solution (sample 1 from Table 1) was mixed with methanol, and heated over a linear temperature ramp ranging from 25 to 200 °C for 15 min. The evaporated materials were detected with a flame ionization detector. It should be noted here that the GC column was not compatible with water; hence methanol had to be used as a solvent. Had we injected the aqueous TBA solution directly into the column, water would act as a solvent and this could have damaged the column. 2.5. Mass Spectrometry. Mass spectrometry (MS) measurements were carried out to determine whether PO polymerizes to higher molecular weight (MW) compounds. Two different MS instruments were used: (1) JEOL AccuTOF with electrospray ionization (ESI) to determine the molecular weights from 100 to about 5000 Da and (2) Shimadzu AximaCFR MALDI TOF to analyze molecular weights from about 5000 to about 100 000 Da. For measurements in the AccuTOF, 5996
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
Figure 1. (a) Phase diagram of the tertiary butyl alcohol (TBA)−water−propylene oxide (PO) system at T = 25 °C (circles) and at T = 10 °C (squares) determined by the cloud-point method. The asterisk indicates the approximate location of the critical point. (b) Phase diagram of the TBA−water−PO system at T = 25 °C (circles) and at T = 10 °C (squares). The asterisk determines the approximate location of the critical point. The dotted area in the figure corresponds to the region where mesoscale phenomena are observed. The dotted-dashed lines, bordering the dotted area, represent the region where the TBA/water ratio varies from 3/97 to 30/70 (mass basis), and where thermodynamic anomalies and structural fluctuations are observed in TBA−water binary systems. This dotted area is divided into three regions. Region 1 corresponds to mesoscale droplets, which are hypothesized to have a hydrophobic core, surrounded by a hydrophilic shell. The mesoscale inhomogeneities observed in region 2 are hypothesized to be interacting droplets akin to a spongelike structure (interconnected water-poor and water-rich domains). The mesoscale inhomogeneities observed in region 3 have a hydrophobic outer shell. (c) Phase diagram of the TBA−water−propylene oxide system at T = 10 °C. The asterisk determines the approximate location of the critical point. The dashed lines border an area where the TBA/water ratio varies from 3:97 to 30:70 (mass basis), and where thermodynamic anomalies and structural fluctuations are observed in TBA−water binary systems. The vertical crosses represent the composition of all the samples studied in this work. (All the phase diagrams are expressed in mass fractions.)
molecular level. The TIP4P/ICE water model was used for these simulations as it accurately predicts the normal freezing point of water and of solid hydrate phases in water62,63 and has been used in our previous study on TBA−water.9 Parameters for TBA and PO were taken from the CHARMM general force field.64 The van der Waals interactions were smoothly switched off between 8 and 10 Å by a potential-based switching function. Long-range electrostatic interactions were calculated by using the particle-mesh Ewald (PME) method.65 An interpolation order of 4 and a direct space tolerance of 10−6 were used for the PME method. Three systems were built using the Packmol package,66 which randomly packs all molecules in a simulation box. An infinite dilution PO model system of one PO molecule was
the sample was mixed with methanol (HPLC grade with 0.999+ purity, Sigma-Aldrich) and injected into the instrument by direct infusion (a LC pump was used for this purpose). For measurements in the MALDI TOF, 25 μL of the sample was mixed with 2 μL of potassium chloride solution (5 mg/mL in 9/1 ethanol/water mixture) and 25 μL of dithronol solution. This matrix was left to air-dry and then subjected to analysis in the instrument under vacuum conditions. We need to note that these instruments were not calibrated to a known standard (a step that could help in the determination of exact molecular weights), but were rather used to determine whether any large molecular weight species were present in solution. 2.6. Molecular Dynamics. MD simulations were performed to understand TBA−water−PO interactions at a 5997
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
simulated in a solution with 14.1 mass % TBA at 8 °C containing 2182 molecules (sample 3 from Table 1). Then, higher concentrations of PO were simulated at 10 °C with 2000 molecules: (1) 12.5 mass % PO and 19.6 mass % TBA (similar to experimental sample 4 from Table 1) and (2) 25.9 mass % PO and 25.0 mass % TBA (similar to experimental sample 5 from Table 1). The NAMD simulation program67 was used for all MD simulations with 2 fs time step. Langevin dynamics was used to maintain constant temperatures for each system, while the Nosé−Hoover Langevin-piston algorithm68,69 was used to maintain constant pressure at 1 bar. The length of the infinite dilution PO model system was 1200 ns, while the others were run for 500 ns. Analyses were performed on the last 500 ns of the MD simulations. The Visual Molecular Dynamics (VMD)70 program was used to create snapshots and to calculate the radial distribution functions (RDF).
Figure 2. Time-dependent intensity autocorrelation functions obtained by dynamic light scattering in aqueous solutions of TBA for the scattering angle θ = 45°. The solid lines are fits to the data in accordance with eq 1. Crosses represent the correlation function obtained for ∼8 mol % (26 mass %) TBA solution (0.997+ purity TBA) at T = 24 °C. This correlation function shows the presence of two relaxation modes: a fast mode with a relaxation time of ∼65 μs and a slow mode with a relaxation time of ∼22 ms. After this solution is filtered multiple times with a 20 nm Anopore filter at ∼5 °C, the slow mode is virtually eliminated (the correlation function at 8.5 °C and θ = 45° is represented by vertical crosses). This correlation function shows only the fast mode, with a relaxation time of about 28 μs. Adding trace amounts of propylene oxide (0.02 mass %) regenerates the slow mode. The resultant correlation function at 8.5 °C and θ = 45°, represented by circles, shows a strong slow mode. It has a relaxation time of about 70 ms.
3. RESULTS 3.1. Ternary-Solution Phase Behavior. TBA and water are completely miscible with each other,54 as well as are TBA and PO.55 However, PO and water exhibit a miscibility gap.71 The resultant ternary phase diagrams at 25 and 10 °C are shown in Figure 1a. The region inside the phase boundary is the two-phase region, while the region outside the boundary is the macroscopically homogeneous one-phase region. From Figure 1a, it is seen that the two-phase domain is relatively small. It also shows that the boundary curve is rather flat, making the determination of the exact location of the critical point difficult. Our measurements further indicate that the miscibility gap in this system does not significantly depend on the temperature. To characterize the behavior of mesoscopic inhomogeneities in the ternary system, various samples within the one-phase, macroscopically homogeneous, domain were analyzed by static and dynamic light scattering at a scattering angle of θ = 45° and T = 25 °C. Figure 1b shows the region within the macroscopically homogeneous domain, where mesoscale inhomogeneities are observed. The light-scattering intensity of the samples in this region is at least an order of magnitude higher than the intensity observed for the corresponding binary systems, namely TBA−water, TBA− PO, and the rest of the ternary system. This region coincides with the region where TBA/water molar ratio varies between 1/99 and 9/91 and where thermodynamic anomalies and micelle-like structural fluctuations are observed in TBA−water binary solutions. To better characterize the nature of the mesoscale phenomena in the TBA−water−PO ternary system, the domain with anomalous scattering has been demarcated into three regions, also shown in Figure 1b. The first region is the dilute PO region, where the concentration of PO is between 0.02 and 1 mass %. The second region is the intermediate region, where the PO concentration varies from about 1 to about 50 mass %. The third region is the concentrated PO region, where PO concentration is above 50 mass %. Although PO concentration varies in all these three regions, the molar ratio of TBA to water remains between 1/99 and 9/91. The location of the various samples studied in this work is shown in Figure 1c. 3.2. Mesoscopic Inhomogeneities Characterized by Light Scattering. To investigate the origin of the mesoscopic behavior of the TBA−water−PO ternary system, lightscattering experiments were initially carried out on TBA− water binary solutions. Figure 2 shows the time-dependent intensity autocorrelation function observed from an aqueous
TBA (0.997+ purity, Sigma-Aldrich) solution (sample 6 from Table 1). The correlation function shows the presence of two relaxation processesa fast mode with a relaxation time of 65 μs and a slow mode with a relaxation time of 22 ms. The fast mode corresponds to mutual molecular diffusion, with a diffusion coefficient of 1.5 × 10−6 cm2/s. In accordance with eq 3, this corresponds to an effective hydrodynamic radius of about 0.6 nm. The slower process corresponds to the diffusion of mesoscale droplets, with a size of about 300 nm. The intensity autocorrelation function obtained after filtering the aqueous TBA solutions at cold conditions, to eliminate the mesoscale droplets, is also shown in Figure 2. The resultant correlation function shows no mesoscale droplets, but only the contribution from molecular diffusion. The relaxation time of this fast mode is about 28 μs. Trace amounts of PO (0.02 mass %) were added to this cold-filtered aqueous TBA solution (sample 2 from Table 1, belonging to region 1). The addition of PO reignited the formation of the mesoscopic inhomogeneities, and the slow mode is observed again (also shown in Figure 2). The reemerged mesoscale inhomogeneities have a relaxation time of about 70 ms. We have earlier shown27 that these inhomogeneities are Brownian diffusive droplets, since the decay rate and inverse light-scattering intensity linearly depend on q2. These mesoscopic droplets have a hydrodynamic radius of about a hundred nanometers. Furthermore, it has been observed that the number of these droplets decreases at higher temperature (25 °C and above) and increases as the temperature is lowered (25 °C and below). All these observations are typical for TBA−water−PO samples within region 1. This behavior is also similar to other systems such as TBA−water−cyclohexane and aqueous solutions of a commercial TBA that are contaminated by unknown hydrophobic impurities. 5998
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
We have also carried out light-scattering experiments in pure TBA, pure PO, dust-free DI water, and binary solutions of TBA−PO and water−PO, and in solutions of TBA and cyclohexane with addition of small amounts of water. No anomalous light scattering and no “slow modes” were observed in any of these samples. Figure 3 shows the time-dependent intensity autocorrelation functions (acquired at a scattering angle of θ = 45° and T = 10
Figure 3. Time-dependent intensity autocorrelation functions obtained by dynamic light scattering for four characteristic TBA− water−propylene oxide samples at a scattering angle of θ = 45° and T = 10 °C. Squares, sample 4; crosses, sample 5; circles, sample 7; vertical crosses, sample 8. (Refer to Table 1 for sample designation.) The solid lines are fits to the data in accordance with eq 1.
Figure 4. Wave number dependence of the relaxation rate (a) and inverse intensity (b) at T = 10 °C (for sample 8, Table 1). Solid line in (a) is a fit to the data in accordance with eq 2. This sample is hypothesized to exhibit a spongelike mesophase.
°C) obtained for four characteristic samples (samples 4, 5, 7, and 8 from Table 1) from region 2, with intermediate PO concentrations. All the correlation functions show the presence of a slow mode, indicating the existence of mesoscopic inhomogeneities. The static light-scattering intensity, relaxation time, diffusion coefficient, and hydrodynamic radius obtained from analyzing the correlation functions of these four characteristic samples are given in Table 2. Table 2. Results Obtained from the Correlation Functions Acquired at a Scattering Angle θ = 45°, and Temperature T = 10 °C, in Accordance with Eqs 1 and 3 sample
intensity (counts/s)
relaxation time (ms)
4 5 7 8
436083 385976 643333 922840
29.7 38.3 15.5 21.0
diffusion coeff (m2/s) 3.2 2.4 6.0 4.4
× × × ×
10−13 10−13 10−13 10−13
hydrodynamic radius (nm)
Figure 5. Time-dependent intensity autocorrelation functions obtained by dynamic light scattering for a TBA−water−propylene oxide solution (sample 8, Table 1) at a scattering angle of θ = 45°. Circles, T = 10 °C; crosses, T = 25 °C. The solid lines are fits to the data in accordance with eq 1. The resultant relaxation times, diffusion coefficients, and hydrodynamic radii are given in Table 3
178 257 161 257
functions obtained at a scattering angle of θ = 45° and temperatures of 10 and 25 °C for a characteristic TBA−water− PO solution (sample 8 from Table 1, belonging to region 2). The static light-scattering intensity, relaxation time, diffusion coefficient, and hydrodynamic radius of this sample at the two temperatures studied are given in Table 3. From this figure and from Table 3, it is seen that, as the temperature is lowered, the scattering contribution from the mesoscopic inhomogeneities increases. This behavior is typical for the samples belonging to region 2 and is similar to what is observed for samples from region 1. Next, we characterize the region with high PO concentrations (region 3 in Figure 1b). Experiments were carried out in the two-phase region of the ternary system. A two-phase sample (sample 9 from Table 1) was prepared and each of the
Figure 4a,b shows the wavenumber dependence of the decay rate and the inverse intensity for one of the samples with intermediate PO concentration, from region 2 (sample 8 from Table 1). The decay rate vs q2 approximately follows a linear dependence, similar to what has been observed for TBA− water−PO samples in region 1. However, the inverse intensity clearly does not follow the Ornstein−Zernike linear q2 dependence, as would be expected for noninteracting Brownian droplets or for near-critical fluctuations.61 Such a nonmonotonic q2 dependence of the inverse intensity has been observed to be typical for TBA−water−PO samples belonging to region 2. To further characterize the region with intermediate PO concentrations, various samples from region 2 were studied at two different temperatures. Figure 5 shows the autocorrelation 5999
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
impurities present in TBA, which were not completely filtered out and removed by nanopore filtration, and triggered by addition of PO. 3.3. Role of Filtration. At this point, it is also worthwhile to discuss the role of filtration in the apparent removal of mesoscopic droplets in aqueous solutions. Does cold filtration indeed remove the hydrophobic impurities that are necessary for the mesoscopic droplets to form? Or could the mesoscopic droplets be broken down by filtration, coalescing back into mesostructures over long periods of time (over a period of months)? Effects of filtration on the size and stability of the mesoscopic droplets were investigated by Sedlák.72 These studies show that the role of filtration is rather complicated and cannot be simply reduced to the removal of impurities. We have tackled this issue by carrying out gas chromatography (GC) experiments with an aqueous TBA solution. Table 4, a
Table 3. Results Obtained from the Correlation Functions Acquired at a Scattering Angle θ = 45°, and Temperatures T = 8.5 and 25 °C, in Accordance with Eqs 1 and 3a
a
temp (°C)
intensity (counts/s)
relaxation time (ms)
diffusion coeff (m2/s)
hydrodynamic radius (nm)
8.5 25
1249416 497515
25.9 13.1
3.6 × 10−13 7.1 × 10−13
300 100
Data belong to sample 8 from Table 1.
phases was analyzed by DLS. Figure 6 shows the intensity− autocorrelation function obtained from the organic phase at a
Table 4. Details of the Results from Gas Chromatography peak no.
Figure 6. Time-dependent intensity autocorrelation function obtained by dynamic light scattering from the top layer (organic-rich phase) of a TBA−water−PO solution (sample 9 from Table 1). T = 25 °C and scattering angle θ = 45°. The autocorrelation function from the organic-rich phase shows the presence of mesoscale inhomogeneities (with a relaxation time of about 2.9 ms), while no measurable correlation function was detected from the aqueous-rich phase. The solid line is a fit to the data in accordance with eq 1.
retention time (s)
2 3 4
1.325 9.148 11.748
1 1 2
1.126 1.126 1.328
area under curve
height
(a) Before Any Filtrations 1380267 735582 6285 313 14008 987 (b) After Cold Filtration 3749051 1119877 3585470 1120272 1672445 816250
compd name TBA unknown unknown methanol methanol TBA
and b, shows the results, obtained in GC experiments, from an unfiltered and a cold-filtered aqueous TBA solution, respectively (TBA concentration at 25 mass %, sample 1 from Table 1). TBA of 0.990+ purity was used for this experiment so that the impurity levels in the unfiltered sample are within the resolution of the GC instrument. Table 4a shows that the TBA source is contaminated with at least two substances, whose molecular weights are higher than that of TBA. Table 4b shows that in the cold-filtered aqueous TBA solution the levels of these impurities become undetectable. Thus, it confirms that cold filtration removes, at least partially, the impurities from aqueous TBA solutions. Sédlak and Rak have further confirmed that these impurities are hydrophobic.58 These results validate our conclusion that the mesoscopic droplets seen in TBA−water solutions indeed originate from the presence of hydrophobic impurities. 3.4. Gaseous Nanobubbles? Another question that routinely arises during the discussion of the origin of mesoscale inhomogeneities in aqueous solutions is whether the inhomogeneities could simply be gaseous nanobubbles? In our earlier work, published in ref 27, we investigated the effect of injecting gas bubbles into aqueous TBA solutions. We injected methane gas (0.999+ purity, purchased form Airgas) into an aqueous TBA solution (25 mass % TBA), which was cold-filtered earlier to remove the mesoscopic inhomogeneities initially present in solution. The injection of methane did create mesoscopic bubbles; however, we showed that they behave very differently from the inhomogeneities created by the addition of PO. The inhomogeneities generated by PO were almost monodisperse and caused stable uniform light scattering. On the other hand, the DLS results from the injection of gas indicated the presence of highly polydisperse gas bubbles (sizes ranging from 50 to about 500 nm), while the light-scattering intensity was unstable and randomly fluctuated over an order of
scattering angle of θ = 45° and a temperature of 25 °C after about 24 h of preparation and equilibration. Analysis of this correlation function shows that the organic phase consists of mesoscale inhomogeneities, with a relaxation time of about 2.9 ms (which would correspond to a hydrodynamic radius of about 250 nm if the viscosity and refractive index were those of pure PO). No discernible correlation function was detected from the aqueous phase. Based on the virtually flat nature of the immiscibility gap of the ternary system, the aqueous phase seems like it is close to the water−PO binary coexistence. Hence, significant concentration fluctuations were not detected in the aqueous phase either. The study of TBA−water−PO system by Sedlák and Rak58 suggests that the source of the mesoscale inhomogeneities in this system is the presence of hydrophobic impurities in PO and TBA. To address this issue, we repeated light-scattering experiments with another source of TBA (0.998+ purity, AlfaAesar). Addition of 0.02 mass % PO (purchased from Sigma-Aldrich, 0.995+ purity; same purity as before but a different bottle) to a cold-filtered TBA−water solution (25 mass % TBA) did not lead to the formation of mesoscale droplets. Mesoscopic droplets were observed in this sample, only when the PO concentration was raised to about 1 mass % (sample 10 from Table 1). This shows that while the addition of PO indeed triggers the formation of mesoscopic droplets in a cold-filtered aqueous TBA solution, the amount of PO needed for this effect to appear could depend on the source and purity of TBA. This led us to conclude, in agreement with the findings reported by Sedlák and Rak,58 that our results reported in ref 27 could have been affected by trace amounts of some unknown 6000
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
magnitude. Furthermore, in contrast to the mesoscale inhomogeneities in the ternary TBA−water−PO system, the gaseous nanobubbles did not reappear when the sample was cooled after heating, indicating that the bubbles disappeared at higher temperatures. Another indication that the observed inhomogeneities in aqueous TBA solutions are not gaseous nanobubbles is the negative effect of ultrasonication. Our studies of TBA−water solutions55 indicated that, when an aqueous TBA solution showing mesoscopic inhomogeneities was subject to ultrasonication at 60 kHz for about 10 min, no change in the lightscattering intensity or in the dynamic autocorrelation function was observed. Other studies in this area, including vacuum degassing, have also dismissed the hypothesis of the nanobubble nature of the mesoscopic inhomogeneities in such solutions.32,73 3.5. Results from MD Simulations. In order to comprehend the molecular-level interactions between TBA, water, and PO, MD simulations have been carried out for a sample from region 1 (sample 3 from Table 1) and two samples from region 2 (samples 4 and 5 from Table 1). Figure 7 shows the RDFs obtained from the MD simulations. Figure 7a shows
the RDF between the cyclic carbon atom, located opposite to the branched side of the PO molecule and the central carbon atom of the TBA molecule. This figure shows the presence of a sharp peak at about 5.1 Å and a smaller peak at about 7.5 Å. This RDF corresponds to van der Waals interactions between the TBA and PO molecules. Figure 7b shows the RDF between the oxygen on PO and the oxygen on the water molecule. This shows the presence of a strong initial peak and a secondary peak at 2.5 and 4.5 Å, respectively. Figure 8a shows a snapshot for a ternary system with a single PO molecule. This figure shows that the PO molecule tends to hydrogen bond with a water molecule, while interacting strongly with TBA molecules via van der Waals interactions. A snapshot with higher PO concentration is shown in Figure 8b. This figure shows that TBA and PO molecules interact strongly to create water-poor and water-rich regions. It is important to note here that, although the experimentally observed mesoscale inhomogeneities are of the order of a hundred nanometers and are extremely long-lived, all-atom MD simulations cannot explore this length scale or time scale. The RDFs shown in Figure 7 explore a length scale of the order of 1 nm. The lifetime of the representative holes in the spongelike phase seen in Figure 8b is of the order of a few nanoseconds but larger simulations may stabilize this phase on the time scale observed from light scattering. The results obtained from MD simulations, however, can shed some light on the interactions that occur at the mesoscale. Results indicate that PO exhibits strong interactions with TBA molecules and weak hydrogen bonds with water molecules. This picture can be extrapolated to the mesoscale as well, and can possibly explain the formation of a spongelike bicontinuous phase over a certain concentration domain of TBA−water−PO system. 3.6. SAXS Experiments. We have also carried out smallangle X-ray scattering (SAXS) experiments to detect the possible presence of a short-range structure, which is suggested by MD simulations. These experiments, performed at the University of Pennsylvania, employed Cu Kα radiation from a Bruker-Nonius FR591 generator with mirror−monochromator optics and a multiwire detector. The compositions of the samples studied correspond to samples 4, 5, 7, and 8 from Table 1. Samples were loaded into thin-walled capillary tubes, and the scattering wave vector ranged from 0.06 to 23 nm−1 (length scale from 0.27 to 105 nm). No detectable scattering peaks were observed in these samples. This negative effect could be attributed to a poor contrast of the scattering objects.
4. DISCUSSION 4.1. Region 1. This region in the ternary phase diagram corresponds to dilute solutions of PO, where PO concentration varies from 0.02 to about 1 mass %. In our earlier work, we have shown that the mesoscale inhomogeneities in this region correspond to noninteracting Brownian diffusive droplets. The fact that the number of these droplets and the threshold concentration of PO, at which they emerge, depend on the source of TBA supports the conclusion made by Sedlák and Rak in their most recent paper.58 Residual hydrophobic impurities remaining in TBA, even after the cold filtration of the aqueous solution, and the hydrophobic impurities present in PO indeed seem to be responsible for the formation of mesoscopic droplets in this region. However, the role and origin of impurities in PO could be more complicated. An issue to be addressed in this respect is the chemical stability of PO. PO, which is highly volatile with a boiling point
Figure 7. (a) Paired radial distribution functions (RDFs) for the ringed carbon atom of PO on the opposite of the branched carbon and the central carbon atom of TBA. Solid line, sample 3; dashed line, sample 4; dashed-dotted line, sample 5. The RDFs are vertically displaced for better clarity. (b) Paired radial distribution function for the oxygen atom of PO and oxygen atom on water. Dotted line, sample 3; solid line, sample 4; dashed line, sample 5. The RDFs are vertically displaced for better clarity. 6001
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
Figure 8. (a) A snapshot of MD simulation (at 540 ns) for a single PO molecule with carbon in gray, oxygen in red, and hydrogen in white. Hydrogen bonds are shown with dashed red lines between water and PO. (b) 500 ns snapshots of sample 4 showing different views in the XYZ space. Only TBA and PO are shown with carbon as gray and oxygen as red in a surface representation. Interconnected holes are water-rich domains. TBA and PO form a spongelike structure.
of 34 °C, may also be chemically unstable. It is known to polymerize to poly(propylene oxide) in an aqueous alcoholic environment.57 So the question arises: is it possible that the polymerization of PO generates additional hydrophobic impurities, which could further contaminate the solution? In order to answer this question, we carried out mass spectrometry (MS) experiments. Figure 9 shows the results obtained from MS of a pure PO sample and of a ternary TBA− water−PO solution (sample 4 from Table 1). Figure 9b, the MS data of pure PO, shows that PO does polymerize, however only to form dimers, trimers, and some other oligomers. These oligomers are also present in the aqueous solution of TBA− water−PO. We investigated a broad range of molecular weights from 50 to 100 000 and did not observe any polymers with a molecular weight greater than 500. Oligomers of PO are too small to cause anomalous light scattering and a slow mode in DLS, which is typical for polymer solutions. However, along with hydrophobic impurities in TBA and with the help of the hydrogen-bonded network between TBA−water and PO− water molecules, such hydrophobic oligomers may play an important role in the formation of mesoscopic inhomogeneities in dilute solutions of PO. Another noteworthy factor that could have an effect on the mesoscopic droplets is the interaction between PO, and the micelle-like clusters formed between TBA and water. MD simulations on the dilute PO system show that PO tends to hydrogen bond with water, while strongly interacting with TBA molecules. We also have evidence from simulations that the presence of PO does not significantly alter the dynamic clustering between TBA and water molecules. Based on the results obtained by studying the dilute solutions of PO in the TBA−water−PO system and the TBA−water−
cyclohexane system,16 we hypothesize that the mesoscopic droplets from region 1 have a hydrophobic core, surrounded by a hydrophilic shell. We further hypothesize that the core of the mesoscopic droplets could consist of hydrophobic impurities, from either TBA or PO, as well as PO oligomers. In contrast, the exterior portion of the mesoscopic droplets could consist of a hydrogen-bonded network between TBA, water, and PO molecules. 4.2. Region 2. This area is characterized by PO concentrations ranging between 1 and about 50 mass % and includes the region around the critical point of miscibility. In this region, the light-scattering intensity is quite high (similar to what has been reported by Sedlák and Rak58 for this region), and, at the first glance, resembles critical opalescence. However, the main contribution to the scattering is neither the scattering from the critical concentration fluctuations nor from individual noninteracting droplets (as has been observed for the droplets in region 1). First, the q2 dependence of the inverse intensity for the samples belonging to this region is nonmonotonic (Figure 4b), indicating that the dominant scattering objects are different from the individual noninteracting droplets or from critical fluctuations.60,61 Furthermore, the inverse intensity exhibits a maximum at q2 = 6 × 10−14 m−2. This corresponds to a length scale of l = 2π/q ∼ 250 nm. Second, critical opalescence is characterized by the diverging correlation length of the concentration fluctuations as the system approaches the critical point.60,61 This effect is not observed simply because the scattering from the mesoscopic inhomogeneities is overwhelming. Rather, as seen from Table 2, the hydrodynamic radii obtained for the various samples from region 2 are virtually independent of the proximity to the critical point. However, we have clearly observed near-critical concentration 6002
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
Figure 9. Mass spectrometry data: (a) TBA−water−PO solution (sample 4 from Table 1). (b) Pure PO. The various peaks correspond to various ionic states of PO in monomer (m/z = 59), PO dimer (m/z = 100), TBA + PO dimer (m/z = 117), and PO or TBA + PO oligomers (m/z = 273, 331, 389). No detectable ions were found to have a molecular weight (m/z) over 500.
fluctuations in another system, TBA−water−isobutanol, where the critical miscibility point is located at a much higher TBA− water ratio.55,56 We interpret the results obtained in region 2 as follows. As the concentration of PO increases, the number of the mesoscopic droplets increase and they start interacting. As the PO concentration further increases, the system passes through the critical-composition region. The critical point divides the liquid−liquid coexistence into two branches, waterrich on the left side and water-poor on the right side. This concentration also divides the one-phase region with respect to hydrophobicity of the environment. Correspondingly, the mesoscopic droplets (containing TBA, water, PO, and hydrophobic impurities) on the water-rich side of the ternary phase diagram should have a hydrophilic exterior. As the PO concentration is increased, the role of water with respect to the mesoscopic inhomogeneities should be inverted. When the structure of the mesoscale inhomogeneities is inverted from a hydrophilic exterior to hydrophobic exterior, the curvature of the mesoscopic interface changes sign. One can interpret this effect as the formation of a bicontinuous spongelike mesophase that occurs in microemulsions.74,75 Another relevant analogy is the percolation phenomenon in microemulsions, when waterrich mesoscopic domains interpenetrate with water-poor domains.76,77 The change of the curvature sign must be accompanied by strong curvature fluctuations (around zero or very small curvature) that could cause strong light scattering. At
smaller PO concentrations (still in region 2) the structure is expected to be transient from increasingly interacting droplets to a sponge phase. The spongelike domains could also be characterized by the length scale, l ∼ 250 nm, estimated from static light scattering, which is of the same order as the size of the mesoscale inhomogeneities obtained by DLS. The formation of a sponge phase, although on a molecular scale, is also supported by MD simulations (Figure 8b). Additionally, it is also possible that the spongelike mesoscale structure in TBA−water−PO systems could be affected by a coupling with near-critical concentration fluctuations, which in our experiments are dominated by much larger inhomogeneities. Exploring this possibility requires further studies. Another challenging question on the behavior of the mesoscopic inhomogeneities in region 2 is whether the spongelike mesophase is a thermodynamically equilibrium state. In order to answer this question, we studied two TBA− water−PO samples (samples 11 and 12 from Table 1, belonging to region 2). Although all the results reported in section 3 were obtained after “equilibrating” the samples for around 24 h, we decided to monitor these two samples over a few months. The hydrodynamic radius and the light-scattering intensity as a function of time are shown in Figure 10. These figures indicate that, as time progresses, the “size” of the inhomogeneities increases and the intensity goes down. Thus, we conclude that the mesoscale inhomogeneities may not be a truly equilibrium phenomenon, but separated from an 6003
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
5. CONCLUSION Mesoscale phenomena in the TBA−water−PO ternary system have been studied by combining different experimental techniques and by MD simulations. Mesoscopic inhomogeneities are observed in this ternary system at PO concentrations varying from about 0.02 to about 65 mass %, and at the TBA/ water mass ratio in the range of 3/97 to 30/70. Remarkably, this TBA/water ratio corresponds to a region where micellelike molecular clustering and thermodynamic anomalies are observed in the TBA−water binary system. The mesoscale phenomena in the ternary system are divided onto three regions. In region 1, the mesoscale inhomogeneities are Brownian diffusive droplets, with a size of the order of a hundred nanometers. We hypothesize that these droplets have a hydrophobic core, enriched by hydrophobic impurities and oligomerized PO molecules, and are surrounded by a hydrophilic shell consisting of TBA, water, and PO molecules. In region 2, where PO concentration varies from about 1 to about 50 mass % (close to the critical concentration of PO), the mesoscale inhomogeneities strongly interact. At higher PO concentrations (region 3), the mesoscale inhomogeneities have a hydrophobic exterior. The inversion of the internal curvature around the critical concentration of PO, from the droplets having a hydrophobic core (region 1) toward the droplets having a hydrophobic exterior (region 3), may result in large internal curvature fluctuations.64,65 We hypothesize that in this region the interacting mesoscale inhomogeneities tend to form a spongelike mesoscale structure (region 2). Although this structure is probably not at a thermodynamically equilibrium state, it is extremely long-lived (order of months). Our findings, in many aspects, agree with the results recently reported by Sedlák and Rak;58 however, our interpretations of certain facts are different. We agree with Sedlák and Rak that hydrophobic impurities, which remain in TBA even after cold filtration and which could also be present in PO,58 play a crucial role in the formation of the mesoscale inhomogeneities. However, we disagree with them on the passive role of PO in this system. In particular, we underscore that PO plays an active role in the formation of the mesoscale inhomogeneities. PO molecules strongly interact with TBA and water molecules (as seen from MD simulations), thus participating in the formation of the hydrophilic shell of the inhomogeneities. Furthermore, PO oligomerizes to form dimers, trimers, etc. (as seen from mass spectrometry), the process which could add to more hydrophobic impurities in the system. The detection of the mesoscale inhomogeneities in the organic layer of the water-poor region of the two-phase sample studied by us and the results from MD simulations support our hypothesis of the formation of a bicontinuous structure containing TBA, water, PO, and hydrophobic impurities, in the intermediate range of PO concentrations (from about 1 to about 50 mass %). Verification of this hypothesis requires additional studies. Finally, we confirm the unique role of TBA−water molecular clustering in the stabilization of the mesoscale inhomogeneities that are observed in the TBA−water−PO system.
Figure 10. Evolution of two near-critical samples of TBA−water− propylene oxide solutions: (a) hydrodynamic radius of the mesoscale inhomogeneities vs time; (b) light-scattering intensity vs time. Circles, sample 11 at T = 10 °C. Crosses, sample 12 at T = 25 °C. The dashed lines are guides to the eye.
equilibrium state by a very high energy barrier. It is not yet clear whether the “ultimate” equilibrium state is a mesoscopically homogeneous state. Somewhat similar nonequilibrium behavior was observed in solutions of 3-methylpyridine, water, and sodium bromide.18 However, the important difference is that, in the system reported in ref 18, the mesoscale inhomogeneities relax to a homogeneous equilibrium state after about 4−6 h, while in the TBA−water−PO system the inhomogeneities persist for months. We hypothesize that the bicontinuous phase could be stabilized by the presence of the mesoscopic droplets. Herzig et al.78 have demonstrated that a bicontinuous interface can be stabilized with the help of colloidal particles, which can be wetted in both the phases and thus can limit the growth of the bicontinuous phase. In our case, the presence of the mesoscopic droplets could arrest the spongelike phase, leading to a formation of a highly stable, long-lived (over months) bicontinuous phase. 4.3. Region 3. In order to verify our hypothesis on the formation of a long-lived bicontinuous mesostructure in region 2, we look at the results obtained from studying a sample in the two-phase region. Figure 6 indicates that the mesoscale inhomogeneities formed in the concentrated PO solution (sample 9 from Table 1, belonging to region 3) prefer to go toward the organic phase as opposed to the aqueous phase. This confirms that these droplets have a hydrophobic exterior. This also implies that a transition between the droplets having a hydrophilic exterior (water-rich environment, region 1) and the droplets with a hydrophobic exterior (water-poor environment, region 3) naturally occurs in the intermediate range of PO concentrations (region 2).
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Present Address ¶
Department of Chemical and Environmental Engineering, Yale University, New Haven, CT 06511.
6004
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
Notes
(16) Subramanian, D.; Boughter, C. T.; Klauda, J. B.; Hammouda, B.; Anisimov, M. A. Mesoscale Inhomogeneities in Aqueous Solutions of Small Amphiphilic Molecules. Faraday Discuss. 2013, 167, 217−238. (17) Georgalis, Y.; Kierzek, A. M. W. Saenger. Cluster Formation in Aqueous Electrolyte Solutions observed by Dynamic Light Scattering. J. Phys. Chem. B 2000, 104, 3405−3406. (18) Kostko, A. F.; Anisimov, M. A.; Sengers, J. V. Criticality in Aqueous Solutions of 3-Methyl Pyridine and Sodium Bromide. Phys. Rev. E 2004, 70, 026118. (19) Li, V.; Cheng, H.; Li, J.; Hao, L.; Zhang, L.; Hammouda, B.; Han, C. C. Large-Scale Structures in Tetrahydrofuran-Water Mixture with a Trace Amount of Antioxidant Butylhydroxytoluene (BHT). J. Phys. Chem. B 2011, 115, 7887−7895. (20) Vuks, M. F.; Shurupova, L. V. The Scattering of Light and Phase Transition in Solutions of Tertiary Butyl Alcohol in Water. Opt. Commun. 1972, 5, 277−278. (21) Beer, C. W., Jr.; Jolly, D. J. Comments on “The Scattering of Light and Phase Transition in Solutions of Tertiary Butyl Alcohol in Water. Opt. Commun. 1974, 11, 150−151. (22) Euliss, G. W.; Sorensen, C. M. Dynamic Light Scattering Studies of Concentration Fluctuations in Aqueous t-Butyl Alcohol Solutions. J. Chem. Phys. 1984, 80, 4767−4773. (23) Iwasaki, K.; Fujiyama, T. Light Scattering Study of Clathrate Hydrate Formation in Binary Mixtures of tert-Butyl Alcohol and Water. J. Phys. Chem. 1977, 81, 1908−1912. (24) Iwasaki, K.; Fujiyama, T. Light-Scattering Study of Clathrate Hydrate Formation in Binary Mixtures of tert-Butyl Alcohol and Water: 2. Temperature Effect. J. Chem. Phys. 1979, 83, 463−468. (25) Bender, T. M.; Pecora, R. Dynamic Light Scattering Measurements of Mutual Diffusion Coefficients of Water-Rich 2Butoxyethanol/Water Systems. J. Phys. Chem. 1988, 92, 1675−1677. (26) Subramanian, D.; Ivanov, D. A.; Yudin, I. K.; Anisimov, M. A.; Sengers, J. V. Mesoscale Inhomogeneities in Aqueous Solutions of 3Methylpyridne and Tertiary Butyl Alcohol. J. Chem. Eng. Data 2011, 56, 1238−1248. (27) Subramanian, D.; Anisimov, M. A. Resolving the Mystery of Aqueous Solutions of Tertiary Butyl Alcohol. J. Phys. Chem. B 2011, 115, 9179−9183. (28) Sedlák, M. Large-Scale Supramolecular Structure in Solutions of Low Molar Mass Compounds and Mixtures of Liquids: I. Light Scattering Characterization. J. Phys. Chem. B 2006, 110, 4329−4338. (29) Sedlák, M. Large-Scale Supramolecular Structure in Solutions of Low Molar Mass Compounds and Mixtures of Liquids: II. Kinetics of the Formation and Long-Time Stability. J. Phys. Chem. B 2006, 110, 4339−4345. (30) Sedlák, M. Large-Scale Supramolecular Structure in Solutions of Low Molar Mass Compounds and Mixtures Of Liquids: III. Correlation with Molecular Properties and Interactions. J. Phys. Chem. B 2006, 110, 13976−13984. (31) Jin, F.; Ye, X.; Wu, C. Observation of Kinetic and Structural Scalings during Slow Coalescence of Nanobubbles in an Aqueous Solution. J. Phys. Chem. B 2007, 111, 13143−13146. (32) Häbich, A.; Ducker, V.; Dunstan, D. E.; Zhang, X. Do Stable Nanobubbles Exist in Mixtures of Organic Solvent and Water? J. Phys. Chem. B 2010, 114, 6962−6967. (33) Polik, W. F.; Burchard, W. Static Light Scattering from Aqueous Poly(Ethylene Oxide) Solutions in the Temperature Range 20−90 °C. Macromolecules 1983, 16, 978−982. (34) Duval, M. Monitoring of Cluster Formation and Elimination in Poly(Ethylene Oxide) Solutions. Macromolecules 2000, 33, 7862− 7867. (35) Devanand, K.; Selser, J. C. Poly(Ethylene Oxide) Does Not Necessarily Aggregate in Water. Nature (London) 1990, 343, 739−741. (36) Faraone, A.; Magazu, S.; Maisano, G.; Migliardo, P.; Tettamanti, E.; Villari, V. The Puzzle of Poly(Ethylene Oxide) Aggregation in Water: Experimental Findings. J. Chem. Phys. 1999, 110, 1801−1805. (37) Kinugasa, S.; Nakahara, H.; Fudagawa, N.; Koga, Y. Aggregation Behaviour of Polyethylene Oxide in Water and Methanol. Macromolecules 1994, 27, 6889−6892.
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We acknowledge useful discussions with D. Blair, B. Chu, J. Leys, V. Molinero, C. J. Peters, W. Schröer, and M. Sedlák. We also thank E. Altabet, E. V. Jouravleva, and S. Hayward for help with sample preparation. M.A.A. appreciates stimulating discussions with D. Ben-Amotz and B. Widom. Research of D.S. and M.A.A. was supported by the Division of Chemistry of the National Science Foundation under Grant No. 1012052. Research of J.B.K. (MD simulations) was supported by the National Science Foundation through XSEDE resources provided by National Institute for Computational Sciences under grant No. TG-MCB100139. Additional computational resources were used on the High Performance Computing Cluster at the University of Maryland.
■
REFERENCES
(1) Neuberg, C. Hydrotropische erscheinungen. Biochem. Z. 1916, 76, 107−176. (2) Nakanishi, K. Partial Molal Volumes of Butyl Alcohols and of Related Compounds in Aqueous Solution. Bull. Chem. Soc. Jpn. 1960, 33, 793−797. (3) Knight, W. S. Ph.D. Dissertation; Princeton University, Princeton, NJ, 1962. (4) De Visser, C.; Perron, G.; Desnoyers, J. E. The Heat Capacities, Volumes and Expansibilities of Tert-Butyl Alcohol−Water Mixtures from 6 to 65 °C. Can. J. Chem. 1977, 55, 856−862. (5) Ott, J. B.; Goates, J. R.; Waite, B. A. Solid+Liquid) Phase Equilibria and Solid-Hydrate Formation in Water + Methyl, + Ethyl, + Isopropyl, and + Tertiary Butyl Alcohols. J. Chem. Thermodyn. 1979, 11, 739−746. (6) Roux, G.; Roberts, D.; Perron, G.; Desnoyers, J. E. Microheterogeneity in Aqueous Organic Solutions: Heat Capacities, Volumes and Expansibilities of some Alcohols, Aminoalcohol and Tertiary Amines in Water. J. Solution Chem. 1980, 9, 629−647. (7) Koga, Y. Excess Partial Molar Enthalpies of Water in Water-TertButanol Mixtures. Can. J. Chem. 1988, 66, 3171−3175. (8) Koga, Y.; Siu, W. Y. U.; Wong, T. Y. H. Excess Partial Molar Free Energies and Entropies in Aqueous Tert-Butyl Alcohol Solutions. J. Phys. Chem. 1990, 94, 7700−7706. (9) Subramanian, D.; Klauda, J. B.; Leys, J.; Anisimov, M. A. Thermodynamic Anomalies and Structural Fluctuations in Aqueous Solutions of Tertiary Butyl Alcohol. Herald of St. Petersburg University 2013, 4, 140−153. (10) Nishi, N.; Takahashi, S.; Matsumoto, M.; Tanaka, A.; Muraya, K.; Takamuku, T.; Yamaguchi, T. Hydrogen Bonding Cluster Formation and Hydrophobic Solute Association in Aqueous Solution of Ethanol. J. Phys. Chem. 1995, 99, 462−468. (11) Guo, J.-H.; Luo, Y.; Augustsson, A.; Kashtanov, S.; Rubensson, J.-E.; Shuh, D. K.; Ågren, H.; Nordgren, J. Molecular Structure of Alcohol-Water Mixtures. Phys. Rev. Lett. 2003, 91, 157401−1−4. (12) Roney, A. B.; Space, B.; Castner, E. W.; Napoleon, R. L.; Moore, P. B. A Molecular Dynamics Study of the Aggregation Phenomena in Aqueous N- Propanol. J. Phys. Chem. B 2004, 108, 7389−7401. (13) Allison, S. K.; Fox, J. P.; Hargreaves, R.; Bates, S. P. Clustering and Microimmiscibility in Alcohol-Water Mixtures: Evidence from Molecular-Dynamics Simulations. Phys. Rev. B 2005, 71, 024201−1−5. (14) Franks, F.; Ives, D. G. The Structural Properties of AlcoholWater Mixtures. Q. Rev. Chem. Soc. 1966, 20, 1−44. (15) Yoshida, K.; Yamaguchi, T.; Otomo, T.; Nagao, M.; Seto, H.; Takeda, T. Concentration Fluctuations and Cluster Dynamics of 2Butoxyethanol-Water Mixtures by Small-Angle Neutron Scattering and Neutron Spin Echo Techniques. J. Mol. Liq. 2005, 119, 125−131. 6005
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006
The Journal of Physical Chemistry B
Article
(38) Polverari, M.; van de Ven, T. G. M. Dilute Aqueous Poly(Ethylene Oxide) Solutions: Clusters and Single Molecules in Thermodynamic Equilibrium. J. Chem. Phys. 1996, 100, 13687−13695. (39) Hammouda, B.; Ho, D.; Kline, S. SANS from Poly(Ethylene Oxide)/Water Systems. Macromolecules 2002, 35, 8578−8585. (40) Ho, D.; Hammouda, B.; Kline, S. Clustering in Poly(Ethylene Oxide) in Water Revisitied. J. Polym. Sci. Polym. Phys. Ed. 2003, 41, 135−138. (41) Hammouda, B.; Ho, D.; Kline, S. Insight into Clustering in Poly(Ethylene Oxide) Solutions. Macromolecules 2004, 37, 6932− 6937. (42) Linegar, K. L.; Adeniran, A. E.; Kostko, A. F.; Anisimov, M. A. Hydrodynamic Radius of Polyethylene Glycol in Solution obtained by Dynamic Light Scattering. Colloid J. 2010, 72, 279−281. (43) Lin, S. C.; Lee, W.; Schurr, J. M. Brownian Motion of Highly Charged Poly (L-Lysine). Effects of Salt and Polyion Concentration. Biopolymers 1978, 17, 1041−1064. (44) Sedlák, M.; Koňaḱ , C.; Štěpánek, P.; Jakeš, J. Semidilute Solutions of Poly (Methacrylic Acid) in the Absence of Salt: Dynamic Light-Scattering Study. Polymer 1987, 28, 873−880. (45) Sedlák, M.; Amis, E. J. Dynamics of Moderately Concentrated Salt-Free Polyelectrolyte Solutions: Molecular Weight Dependence. J. Chem. Phys. 1992, 96, 817−825. (46) Sedlák, M.; Amis, E. J. Concentration and Molecular Weight Regime Diagram of Salt-Free Polyelectrolyte Solutions as Studied by Light Scattering. J. Chem. Phys. 1992, 96, 826−834. (47) Li, X.; Reed, W. F. Polyelectrolyte Properties of Proteoglycan Monomers. J. Chem. Phys. 1991, 94, 4568−4580. (48) Reed, W. F.; Ghosh, S.; Medjahdi, G.; Francois, J. Dependence of Polyelectrolyte Apparent Persistence Lengths, Viscosity, and Diffusion On Ionic Strength and Linear Charge Density. Macromolecules 1991, 24, 6189−6198. (49) Schmitz, K. S. An Overview of Polyelectrolytes. Biopolymers 1993, 33, 923−932. (50) Mesostructures and Dynamics in Liquids and Solutions. In Faraday Discussions; RSC Publishing: London, 2013; Vol. 167. (51) Kežić, B.; Perera, A. Fluctuations and Microheterogeneity in Mixtures of Complex Fluids. Faraday Discus. 2013, 167, 145−158. (52) Wilcox, D. S.; Rankin, B. M.; Ben-Amotz, D. Distinguishing Aggregation from Random Mixing in Aqueous t-Butyl Alcohol Solutions. Faraday Discuss. 2013, 167, 177−190. (53) Fiore, A.; Venkateshwaran, V.; Garde, S. Trimethylamine NOxide (TMAO) and tert-Butyl Alcohol (TBA) at Hydrophobic Interfaces: Insights from Molecular Dynamics Simulations. Langmuir 2013, 29, 8017−8024. (54) Kasraian, K.; DeLuca, P. P. Thermal Analysis of the Tertiary Butyl Alcohol- Water System and its Implications on Freeze-Drying. Pharm. Res. 1995, 12, 484−490. (55) Subramanian, D. Ph.D. Dissertation, University of Maryland, College Park, MD, 2012. (56) Subramanian, D.; Anisimov, M. A. Mesoscale Solubilization in Aqueous Solutions of Hydrotropes. Fluid Phase Equilib. 2014, 362, 170−176. (57) Trent, D. L. Propylene Oxide. In Kirk-Othmer Encyclopedia of Chemical Technology; John Wiley and Sons, Inc.: New York, published online 4 June 2001. (58) Sedlák, M.; Rak, D. On the Origin of Mesoscale Structures in Aqueous Solutions of Tertiary Butyl Alcohol: The Mystery Resolved. J. Chem. Phys. B 2014, 118, 2726−2737. (59) Othmer, D. F.; White, R. E.; Trueger, E. Liquid-Liquid Extraction Data. Ind. Eng. Chem. 1941, 33, 1240−1248. (60) Chu, B. Laser Light Scattering: Basic Principles and Practice; Academic Press: Boston, 1991. (61) Berne, B. J.; Pecora, R. Dynamic Light Scattering: With Applications to Chemistry, Biology, and Physics; Dover Publications: Mineola, NY, 2000. (62) Abascal, J. L. F.; Sanz, E.; Fernandez, R. G.; Vega, C. A Potential Model for the Study of Ices and Amorphous Water: TIP4P/Ice. J. Chem. Phys. 2005, 122, 234511−234519.
(63) Walsh, M. R.; Koh, C. A.; Sloan, E. D.; Sum, A. K.; Wu, D. T. Microsecond Simulations of Spontaneous Methane Hydrate Nucleation and Growth. Science 2009, 326, 1095−1098. (64) Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.; Zhong, S.; Shim, J.; Darian, E.; Guvench, O.; Lopes, P.; Vorobyov, I.; Mackerell, A. D. CHARMM General Force Field: A Force Field for Drug-Like Molecules Compatible with the CHARMM All-Atom Additive Biological Force Fields. J. Comput. Chem. 2010, 31, 671−690. (65) Darden, T.; York, D.; Pedersen, L. Particle Mesh Ewald - an NLog(N) Method for Ewald Sums in Large Systems. J. Chem. Phys. 1993, 98, 10089−10092. (66) Martínez, L.; Andrade, R.; Birgin, E. G.; Martínez, J. M. PACKMOL: A Package for Building Initial Configurations for Molecular Dynamics Simulations. J. Comput. Chem. 1993, 30, 2157− 2164. (67) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable Molecular Dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781− 1802. (68) Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R. Constant Pressure Molecular Dynamics Simulation: The Langevin Piston Method. J. Chem. Phys. 1995, 103, 4613−4621. (69) Martyna, G. J.; Tobias, D. J.; Klein, M. L. Constant Pressure Molecular Dynamics Algorithms. J. Chem. Phys. 1994, 101, 4177− 4189. (70) Humphrey, W.; Dalke, A.; Schulten, K. VMD: Visual Molecular Dynamics. J. Mol. Graphics 1996, 14, 33−38. (71) Wickert, J. N.; Tamplin, W. S.; Shank, R. L. Phase Equilibria in the System Propylene Oxide-Water. Chem. Eng. Prog. Symp. Ser. No. 2 1958, 48, 92−96. (72) Sedlák, M. Mechanical Properties and Stability of Multimacroion Domains in Polyelectrolyte Solutions. J. Chem. Phys. 2002, 116, 5236−5247. (73) Sedlák, M.; Rak, D. Large-Scale Inhomogeneities in Solutions of Low Molar Mass Compounds and Mixtures of Liquids: Supramolecular Structures or Nanobubbles? J. Chem. Phys. B 2013, 117, 2499−2505. (74) Gompper, G.; Schick, M. Self-Assembling Amphiphilic Systems. In Phase Transitions and Critical Phenomena; Academic Press: London, 1994; Vol. 16. (75) Freyssingeas, É.; Nallet, F.; Roux, D. Measurement of the Membrane Flexibility in Lamellar and “Sponge” Phases of the C12E5/ Hexanol/Water System. Langmuir 1996, 12, 6028−6035. (76) Grest, G. S.; Webman, I.; Safran, S. A.; Bug, L. R. Dynamic Percolation in Microemulsions. Phys. Rev. A (R) 1986, 33, 2842−2845. (77) Gu, G.; Wang, W.; Yan, H. Electric Percolation of Water-in-Oil Microemulsions: The Application of Effective Medium Theory to System Sodium Dodecylbenzenesulfonate (DDBS)/n-Pentanol/nHeptane/Water. J. Colloid Interface Sci. 1996, 178, 358−360. (78) Herzig, E. M.; White, K. A.; Schofield, W. C.; Poon, W. C. K.; Clegg, P. S. Bicontinuous emulsions stabilized solely by colloidal particles. Nat. Mater. 2007, 6, 966−971.
6006
dx.doi.org/10.1021/jp4125183 | J. Phys. Chem. B 2014, 118, 5994−6006