The classic explanation for why oil and water don’t mix centers on the molecules’ charge distributions. The combination of water’s polarity, which promotes hydrogen bonding among water molecules, and oil’s lack of polarity leads the two liquids to layer one on top of the other at the macroscopic scale.
For the charge-distribution hypothesis to hold up, however, physical chemists will need to explain the numerous experiments that show that pure oil can form droplets. Other experiments show that those droplets move under the influence of an applied electric field. The movement always takes place toward the anode, suggesting that the oil droplets have an unbounded negative charge.
Over the past several decades, physical chemists have vigorously debated what mechanism could produce oil’s negative surface charge while still allowing oil and water to separate. The leading hypothesis has been that hydroxide ions, the only naturally negative component in pristine water, must adsorb to an oil droplet’s surface to negatively charge it.
Sylvie Roke of EPFL in Switzerland and her collaborators now offer a different explanation. In their December paper in Science, they provide evidence that a charge transfer can explain oil’s intriguing properties. Studying the surfaces of oil droplets, they observed shifts in the vibrational frequencies of the carbon–hydrogen and oxygen–deuterium bonds of interfacial oil and water. The shifts correlate to the negative charge on an oil droplet’s surface. The evidence is supported by simulations that demonstrate how oxygen atoms transfer a bit of negative charge to the C–H bonds on the oil.
Hydroxide’s charge
The negative charge on the surface of an oil droplet dispersed in water can be more generally interpreted as the zeta potential ζ, or the electric potential at a slipping plane. The slipping plane is a mathematical convention that separates mobile fluid molecules from those that remain in a layer adsorbed to a surface.
A 1938 study and subsequent research have measured negative ζ values for oil droplets in water. Some of those studies found that ζ becomes even more negative as the pH of the solution increases. Those observations led many researchers to conclude that the accumulating negative charge on an oil droplet’s surface was due to the adsorption of OH−. The ion’s concentration is regulated by pH and is thermodynamically favored to adsorb to oil droplets dispersed in water.
But in 2011 Roke and her colleagues performed experiments and simulations and found no evidence of the ion at the oil–water interface. More recently Roke decided to investigate that interface further and teamed with Ali Hassanali, a physical chemistry theorist at the International Center for Theoretical Physics in Trieste, Italy.
An unexpected spectrum
To analyze the surface interfaces, Roke turned to sum frequency generation (SFG) spectroscopy, a method that relies on the light frequency whose sum is equal to two input laser-beam frequencies. It does have limitations (see the article by Gabor Somorjai and Jeong Young Park, Physics Today, October 2007, page 48). For example, the technique requires an extended planar surface to reflect the input laser light. To get around that constraint, Roke measured the sum frequency light that scattered from oil droplets dispersed in water. She had developed the technique in 2003 for probing the molecular surface structure of particles and droplets in solutions.
One major technical issue, however, complicated the measurement of water. In scattering experiments, the optical laser beams travel through the oil-droplet dispersion. That means that the IR light’s frequency is in resonance with not only the interfacial water droplets’ vibrational frequency but also that of the bulk water. The bulk water absorbs the IR light, which distorts the vibrational spectra. The result is a spectrum whose peaks don’t necessarily correspond to the expected vibrational modes of interfacial oil and water.
“For a very long time, I thought that the experiment wasn’t possible because of the absorption,” Roke says. “But then I calculated how much interface would be probed even with absorption and found it still compares favorably to a planar SFG experiment. I decided to give it a try.” Starting with a theoretical model of the light–matter interaction, the researchers looked at how the light signal’s measured sum frequency intensity compared with the theoretical nonlinear scattering response at the oil–water interface.
The water’s absorption of IR light, the efficiency of the optics in collecting the light, and other parameters had to be experimentally determined. All told, the measurements and the critical steps of properly verifying each parameter and correction took three years. “I spent as much time as I possibly could to verify things,” Roke says.
The figure below shows the spectra of interfacial water (D2O) with dispersed dodecane oil droplets (blue circles) and of an air–water interface (red circles). (Using heavy water stabilized the experimental conditions and reduced the run time by about 40%.) A water molecule’s vibrational frequency shift depends in part on the strength of the hydrogen bonds with neighboring molecules. The vertical dashed lines at around 2400 cm−1 and 2500 cm−1 mark the frequency boundary at which hydrogen bonding affects water’s oxygen–deuterium bond. Both interfaces have roughly the same peak frequency range.
The vibrational frequencies of the water molecules that lack strong hydrogen bonds at the oil–water interface are centered at about 2650 cm−1, a full 100 cm−1 less than that of the air–water interface. That observation was the researchers’ first clue that the oil and water may have had an unexpected interaction with each other.
If such an interaction exists, water’s 100 cm−1 vibrational frequency decrease (left-pointing arrow in the figure), or redshift, should be accompanied by a concomitant blueshift in the vibrational modes of the oil droplets. And indeed, in the spectra of the C–H bond of hexadecane droplets of oil (not shown), the researchers found a blueshift in the vibrations of the bonds that were in direct contact with D2O.
Thousands of atoms
To understand the apparent charge-transfer interaction between oil and water, Hassanali and his collaborators performed molecular dynamics simulations using density functional theory. Despite the successes made possible by DFT (see the article by Andrew Zangwill, Physics Today, July 2015, page 34), some problems have remained particularly stubborn, including the interactions of hydrogen bonds. To better understand hydrophobic electronic effects, mesoscopic-scale phenomena need to be modeled. Hassanali accomplished that by pushing the boundaries of electronic structure calculations, using thousands of atoms in the simulations rather than the more typical hundreds of atoms.
The computational requirements for most DFT simulations typically scale with the cube of the number of electrons. But for insulating systems characterized by large band gaps, DFT techniques that scale only linearly can be applied to study the electronic structure. Using that approach, Hassanali and his colleagues simulated at the nanometer length scale the electronic effects of hydrophobic interfaces composed of thousands of atoms. The image below shows the simulated oil–water interface.
To Hassanali’s surprise, the improved simulations showed that a negative charge developed on the dodecane oil molecules. Lone oxygen atoms of interfacial water molecules donated some of their electron density to C–H bonds in the oil, which led to the formation of weak hydrogen bonds between oil and water molecules. The charge-transfer effect in the simulations was apparent for water as far as five angstroms from the oil molecules. In that vicinity, water’s oxygen atoms oriented themselves toward the CH2 groups of the oil.
Hassanali says that the charge-transfer mechanism results from fluctuations in the hydrogen bonding network of water. The fluctuations create deviations from ideal water structure—topological defects—that create asymmetries in the local environment. The asymmetries, in turn, determine which lone pair of electrons delocalize along specific hydrogen bonds.
“Originally, we thought this is something that can only happen between pairs of water molecules,” says Hassanali. “We were convinced that it was an artifact of the DFT. It turns out that same sort of effect can take place between a water molecule and the boring oily molecule as well. And it’s that subtle quantum mechanical effect of the charge transfer that then results in the frequency shifts of the C–H bonds, which is what [Roke] measures.”
Some researchers in the field are skeptical of the Roke and Hassanali team’s results. “The measurements of Roke and coworkers are not wrong,” says Richard Zare of Stanford University. But he thinks that the bicarbonate anion (HCO3−) or trace amounts of dissolved carbon dioxide could negatively charge an oil droplet’s surface. However, Roke says that if the bicarbonate ions were “present in the quantity needed to generate the necessary charge, they would have been detected.”
If hydroxide or bicarbonate were present and contributing a negative charge to oil, that would help explain another curious finding: that increasing pH raises the mobility of oil droplets in water. Roke is investigating how to square her results with the pH-dependent behavior. “I have a very interesting explanation for this,” she says, “but that’s for the future.”
