Effect of the Hydrophobic Alcohol Chain Length on the Hydrogen

Jun 18, 2014 - indicates that alcohols do not have a substantial effect on the structure of the ... network of water, with the detailed effect dependi...
0 downloads 0 Views 306KB Size
Subscriber access provided by The University Library | University of Sheffield

Article

Effect of the Hydrophobic Alcohol Chain Length on the Hydrogen-Bond Network of Water Iina Juurinen, Tuomas Pylkkänen, Christoph J. Sahle, Laura Simonelli, Keijo Hamalainen, Simo Huotari, and Mikko Oskari Hakala J. Phys. Chem. B, Just Accepted Manuscript • Publication Date (Web): 18 Jun 2014 Downloaded from http://pubs.acs.org on July 8, 2014

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

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

Page 1 of 22

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

The Journal of Physical Chemistry

Effect of the Hydrophobic Alcohol Chain Length on the Hydrogen-Bond Network of Water Iina Juurinen,∗,† Tuomas Pylkkänen,† Christoph J. Sahle,† Laura Simonelli,‡ Keijo Hämäläinen,† Simo Huotari,† and Mikko Hakala∗,† Department of Physics, P.O.B. 64, FI-00014 University of Helsinki, Finland, and ESRF - The European Synchrotron, CS 40220, 38043 Grenoble Cedex 9, France E-mail: [email protected]; [email protected]

To whom correspondence should be addressed of Helsinki ‡ European Synchrotron Radiation Facility ∗

† University

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

Abstract The microscopic structure of the hydrogen-bond network of water-alcohol mixtures was studied using x-ray Raman scattering (XRS). In order to systematically examine how the hydrogen-bond network of water is affected by an increasing size of the hydrophobic group, small linear alcohols (methanol, ethanol, and propanol) in constant mole fractions were studied. The oxygen K-edge spectra were not altered upon hydration of the alcohols beyond a simple superposition of signals from alcohol and water. The experiment thus indicates that alcohols do not have substantial effect on the structure of the hydrogen-bond network of water. In particular, no apparent breaking or forming of the hydrogen bonds is observed to take place in the overall structure. In addition, there is no indication of changes in the tetrahedrality of the hydrogen-bond network of water in the vicinity of alcohol molecules.

Keywords: inelastic x-ray scattering, aqueous solution, mixtures, x-ray Raman scattering, oxygen K-edge, amphiphilicity

2 ACS Paragon Plus Environment

Page 2 of 22

Page 3 of 22

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

The Journal of Physical Chemistry

Introduction Water is essential in biological processes and it is a component in most solutions. 1 For example, the interplay between DNA and water eventually determines its structure and dynamics. 2 Polymers, such as DNA, are typically amphiphilic, including both hydrophobic and hydrophilic groups. In order to understand how these parts contribute to the structure of the polymer, detailed studies on hydrophilic and hydrophobic domains are needed. To reduce the complexity of studying a whole polymer, studies on the interaction with water are typically performed with small amphiphilic molecules to ensure their solubility. How the hydrophobes affect the structure of water has been debated since 1945, when Frank and Evans first proposed iceberg formation of water molecules around hydrophobic solutes as an explanation for the observed negative changes in enthalpy and excess entropy. 3 However, the subsequent alternative explanations have not reached full consensus on what takes place in the vicinity of hydrophobes. Moreover, hydrophilic groups are incorporated on the hydrogen-bond network of water, with the detailed effect depending on the particular molecule. In the majority of studies on hydrophobicity, no change in the structure of the hydrogen-bond network of water near hydrophobic domains compared to bulk water has been observed. 4–13 However, some studies report observations of increased structure 14–16 or decreased structure 17–19 for water in contact with hydrophobic groups. Also, the size of the hydrophobe has been found to affect its ability to change the structure of the hydrogen-bond network of water. 20–23 However, as most studies on the water structure and dynamics are done with amphiphilic molecules, it is also debated whether the changes are caused by the whole molecule, or the hydrophobic 24,25 or hydrophilic 8,10,11,26 group of the molecule. In order to systematically study the hydrophobic and hydrophilic interactions we chose linear alcohols, namely methanol, ethanol and 1-propanol. They all have the same hydrophilic O-H group, but the hydrophobic hydrocarbyl chain length increases from one to three carbon atoms. When alcohols are solvated in water in the same mole fraction their hydrophilic interactions should be similar while the hydrophobic interactions should increase as the hydrophobic group 3 ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

extends. This behaviour has been previously studied in sodium propanoate and sodium butanoate with broadband dielectric relaxation spectroscopy. 27 We use x-ray Raman scattering 28 (XRS) to obtain information on the structural changes of water-alcohol mixtures. XRS is an element sensitive non-resonant inelastic x-ray scattering technique where scattering from core electrons is observed. Previous studies on the oxygen K-edge for water and water mixtures have proven to be effective on describing the changes in the water structure. 29–43 This method enables us to identify if amphiphilic molecules cause notable changes in the hydrogen bond network of water.

Methods The experiments were performed at ID16 beamline∗ of the European Synchrotron Radiation Facility (ESRF). 44 Radiation from three undulators was focused to a 100 × 30µ m2 spot with a toroidal mirror after it was monochromatized with two Si double-crystal monochromators. The scattering was detected with nine spherically bent Si (660) analyzer crystals and a Medipix2 pixel detector. XRS spectra were measured by keeping the analyzer energy fixed at 9.68 keV while scanning the incident photon energy. The total energy resolution of the setup was 0.57 eV (full width at half maximum (fwhm)). The average scattering angle of the multianalyzer setup corresponded to a momentum transfer of 3.2±0.8 Å−1 . Methanol, ethanol and 1-propanol (ACS reagent grade) were obtained from Sigma Aldrich. Ultrapure milli-Q water (resistivity > 18.2 MΩ cm) was used in the experiments. Water and alcohols were mixed to alcohol mole fractions of 5% and 15%. Additionally pure water and pure alcohols, and a mixture of 25% of ethanol were measured. The uncertainty in concentrations was 1%. In order to eliminate radiation damage, the samples were contained in a peristaltic-pumpbased closed-loop setup, which produced a stable vertical liquid column (diameter 2 mm) from a stainless steel nozzle. Real-space imaging property of the spectrometer 45 was used to exclude the ∗ ID16

was closed for operation in 2012 and the inelastic x-ray scattering spectroscopy research of electronic excitations continues at the new beamline ID20 as a part of the ESRF Upgrade Programme.

4 ACS Paragon Plus Environment

Page 4 of 22

Page 5 of 22

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

The Journal of Physical Chemistry

background of scattering from air. The measurements were performed at room temperature. In the data analysis the spectra were processed separately for each analyzer. Measurements of the quasielastic line for all samples allowed accurate definition of the energy loss scale. A constant background below 532 eV was removed and the spectra were normalized to an equal area in the range 530-552 eV. The normalized spectra of all the analyzers were averaged to produce the final result for each concentration. Similar experimental setup and data analysis has recently been applied to water, 33 alcohols, 46 aqueous LiCl 34 and with a different sample cell to high-pressure ice. 37 In order to obtain molecular structures for computational spectra, classical molecular dynamics simulations were performed. The Gromacs software 47 was utilized, with TIP4P 48 and OPLSAA 49 force fields. The leap-frog integrator and the velocity-rescale thermostat and ParrinelloRahman methods were used for simulations in 300 K and 1 bar. For the subsequent x-ray Raman spectral calculations ERKALE code 50,51 was used utilizing the transition-potential approximation 52 within density functional theory. The revised Perdew-Burke-Ernzerhof functional 53–55 was used with the IGLO-III basis set 56 for the excited oxygen atom and Dunning’s augmented correlation consistent polarized valence double zeta basis set 57 for all other atoms. As a model of the fingerprints of different molecular structures on the spectral features, small example cases were calculated. For these calculations clusters of 3 to 5 molecules with desired topologies were selected from the simulations. The transitions in the example cases were broadened with Gaussians with fwhm 0.5 eV. Additionally, to obtain a reference spectrum, a more extensive computational approach was applied on pure water. For the calculations 100 clusters with a radius of 5 Å were randomly selected from the simulations. The transitions obtained from calculations were broadened with linearly increasing Gaussians 30,58,59 with a fwhm of 0.7 eV below 535 eV and 2.0 eV above 542 eV. Between the energies the fwhm was linearly increased from 0.7 to 2.0 eV. In the visualisation of the molecular orbitals Avogadro program was used and the isosurface was set to 0.04.

5 ACS Paragon Plus Environment

The Journal of Physical Chemistry

Results and discussion

0.2

15% propanol 0.15

Intensity (arb. units)

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

15% ethanol 15% methanol 0.1

5% propanol 5% ethanol 5% methanol

0.05 Water

A 0

535

B

C 540

545

550

Energy transfer (eV)

Figure 1: X-ray Raman scattering spectra of water and 5% and 15% aqueous solution of methanol, ethanol and propanol. The spectra are offset for clarity. Repeated black thin line represents the smoothed spectrum of pure water. A, B, and C denote pre-, main-, and post-edges. In order to observe the changes in the hydrogen-bond network of water induced by the alcohol molecules with increasing lengths of hydrocarbyl groups, the XRS oxygen K-edge spectra were measured for aqueous methanol, ethanol, and propanol in constant alcohol mole fractions. The resulting spectra for two different mole fractions (5% and 15%) are shown in Figure 1. The XRS spectral features are typically divided in three regions, namely the pre-edge (E≈535 eV), the mainedge (E≈538 eV) and the post-edge (E≈540 − 542 eV) (as illustrated in Figure 1). The overall changes in the experimental spectra are small upon addition of alcohols. However, there is an increase in the intensities of the spectra between the main-edge and post-edge for the 15% solutions. The increase is most pronounced in the mixture with propanol, which has the longest hydrocarbyl group. Above the post-edge the spectra seem to be unchanged, and in the pre-edge area there is little or no change. 6 ACS Paragon Plus Environment

Page 6 of 22

Page 7 of 22

0.12 Pure propanol Pure ethanol Pure methanol Water

Intensity (arb. units)

0.1

0.08

0.06

0.04

0.02

0

534

536

538

540

542

544

546

548

550

552

Energy transfer (eV)

Figure 2: X-ray Raman scattering spectra of pure methanol, ethanol and propanol. Black dashed line represents the spectrum of pure water.

0.2

15% propanol 0.15

Intensity (arb. units)

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

The Journal of Physical Chemistry

5% propanol

25% ethanol 0.1

15% ethanol

5% ethanol

15% methanol 0.05 5% methanol

0

534

536

538

540

542

544

546

548

550

552

Energy transfer (eV)

Figure 3: Superpositions (thin black lines) of spectra of pure alcohols and water in different concentrations with the spectra from mixtures (colorful, thicker lines). From top to bottom: aqueous propanol (15% and 5%), ethanol (25%, 15%, and 5%), and methanol (15% and 5%). Depending on the concentration, 5% or 15% of the signal comes from the oxygen in alcohols. Thus the results in Figure 1 do not purely represent the changes in water. The spectra for different pure alcohols are distinct (Figure 2), as the shape of the oxygen K-edge spectrum is sensitive to

7 ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

Page 8 of 22

the covalent environment of the excited oxygen. However, the spectra of propanol and ethanol are quite similar. They are relatively featureless compared to the spectrum of water: the pre-edge is combined to the main-edge and there is no post-edge shoulder as seen in the water spectrum. The intensity in the main-edge is higher for propanol than for ethanol. Methanol, on the other hand, has a visible pre-edge although it is not as pronounced as for water. In the post-edge area the methanol spectrum coincides with the spectra of ethanol and propanol. The pure alcohol results are similar to those reported previously. 38,46,59,60 It is surprising how little the oxygen K-edge spectra are changed upon addition of alcohols (Figure 1). For example, in the case of aqueous ethanol, already at a mole fraction of 5% of ethanol the volume fraction is ∼15 vol-% and rises to ∼37 vol-% at a mole fraction of 15%. Thus a substantial fraction of the solvent is interacting with the solute, while at the same time the majority of the signal comes from the oxygen in water. In order to examine the detailed behaviour of the spectra we used a superposition analysis in which the alcohol spectrum Ia (E) and water spectrum Iw (E) were summed weighting by the measured mole fraction x to get the superposition spectra

I(E) = x × Ia (E) + (1 − x) × Iw (E).

(1)

As shown in Figure 3, the superpositions match the experimental spectra of the solution closely. If the structure of the hydrogen-bond network of water were changed notably upon the addition of alcohols, the experimental spectra would differ from the superposition spectra. However, within the statistical accuracy no differences can be observed between the superposition and the experimental mixture spectra. This is the key experimental finding in this work. In the following, with the help of example calculations and knowledge based on previous studies, we discuss what this finding implicates. The oxygen K-edge spectrum depends on the molecular structure of the scattering system, and therefore the correlations between spectral features and molecular structures have been studied extensively. 29,30,35,37,46,61–66 Since the whole spectrum is a complex composition of different 8 ACS Paragon Plus Environment

Page 9 of 22

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

The Journal of Physical Chemistry

Figure 4: Spectra of six examples of local molecular geometries (snapshots with the first unoccupied molecular orbitals on the right hand side): chain of three water molecules (A), with excited atom in the middle (indicated by arrow); ethanol molecule donating hydrogen bond in a threemolecule chain (B); ethanol molecule accepting hydrogen bond in a three-molecule chain (C); three-molecule chain near hydrophobic group of ethanol (D); fully hydrated water molecule (E); ethanol as the molecule with the excited atom in a three-molecule chain (F). The figures include computational transitions (red vertical lines), transitions broadened with Gaussians with constant fwhm (dark blue spectra), and computational spectra for pure water (light blue spectra). The first transitions are marked with arrows. transitions characteristic to the local geometries, we approach it computationally through simple models which allows us to visualize the changes in a straightforward way. Six representative local geometries extracted from molecular dynamics simulations are shown in Figure 4, where also the computational transitions and broadened spectra of the excited oxygen are compared with the

9 ACS Paragon Plus Environment

The Journal of Physical Chemistry

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

Page 10 of 22

computational spectrum of pure water (see Methods). The studied model systems cover the different possible positions of an alcohol molecule in a three-molecule chain (water only in Figure 4 A), where the molecule with the excited oxygen donates one and accepts one hydrogen bond. Changing the donating molecule to an alcohol molecule (Figure 4 B) does not change the spectrum notably, while changing the accepting molecule (Figure 4 C) induces small spectral changes. Additionally, having the three-molecule water chain near the hydrophobic part of the alcohol molecule (Figure 4 D) changes the spectrum only slightly. For reference, the spectra of fully coordinated water (Figure 4 E) and ethanol as the excited molecule (Figure 4 F) differ notably from the spectra of the three-molecule water chain (Figure 4 A). Overall, the computations suggest that most notable changes in the oxygen K-edge spectra arise from the number of hydrogen bonds and of the covalent bonds of the excited atom, i.e. whether the molecule is water or ethanol (Figure 4 E and F). Smaller changes are seen with the changes in the unoccupied molecular orbitals arising from the alcohol molecules (Figure 4 B-C). However, these changes do not shift the overall spectral weights. The computational findings suggest that bonding between water and alcohol does not result in notable spectral changes. Thus any change beyond a single superposition would be related to the structural changes in the hydrogen bond network of water. These observations from small model systems are in line with previous interpretations of the composition of oxygen K-edge spectra of water based on experiments and more extensive calculations. For example, the pre-edge has been found to be sensitive to the changes in hydrogen bonds 29,35,61 that also affect the post-edge. The post-edge is very pronounced in cases where the tetrahedrality of the hydrogen bonds is extensive, for example, in hexagonal ice. 30,35,37 Changes in the main-edge have been previously attributed to density increases, where the hydrogen bonding does not necessarily change, for example, in high density ices. 46,61,63 This is similar to what we observe when the excited atom is in the alcohol molecule (Figure 4 F): the presence of the carbon atoms near the excited oxygen transfers the intensity from the pre-edge towards the main-edge. As was stated earlier, the superposition spectra (Figure 3) indicate that the water contribution

10 ACS Paragon Plus Environment

Page 11 of 22

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

The Journal of Physical Chemistry

to the oxygen K-edge spectra is not altered when water is mixed with alcohols. The observed small changes in the experimental spectra in Figure 1 mainly arise from the addition of the signal from oxygen in the alcohol molecule, i.e. oxygen atoms covalently bonded to the hydrocarbyl group. This is supported by the pure alcohol spectra (Figure 2) and the computational examples, where more marked changes are seen when the excited molecule is ethanol (Figure 4 F) compared to the cases where water interacts with ethanol either via hydrogen bonds or hydrophobically (Figure 4 B,C, and D). In fact, the results of this work rule out the possibility that hydrophobic interaction of small alcohols would either break hydrogen bonds or create them. It has been shown that the oxygen Kedge spectral features change notably especially in the pre-edge area when the hydrogen bonding changes. 29,33–35,41,61,62,64 Dixit et al. 11 previously demonstrated that the total number of hydrogen bonds does not change upon mixing water and methanol, and we can expand this interpretation to be valid for all small alcohols. Additionally, our results are in line with the findings on how the size of the hydrophobe affects its ability to break hydrogen bonds: the hydrogen-bond network stays intact near small hydrophobic molecules. 21–23 Our findings are also in agreement with a recent study of methanol-water mixtures by Nagasaka et al. They observed with x-ray absorption that the methanol molecules become part of the hydrogen-bond network of water for low molar fractions of methanol (