Imaging ice only a few nanometers thick as it forms bulk ice was supposed to be impossible. A scanning tunneling microscope (STM) shouldn’t work with ice because STMs create images by relying on conducting current, which runs contrary to one of ice’s basic properties — insulation.
But no one told this to Sandia physicists Norm Bartelt and Konrad Thürmer (both 8756). Actually, everyone did, but they didn’t listen. Experience and a bit of stubbornness resulted in what Konrad describes as the “most important scientific accomplishment of my career.”
The images captured by Norm and Konrad show ice sheets growing one molecular layer at a time. These images also reveal some new truths about ice and solve a decades-old mystery about why ice grows in cubic form at very cold temperatures, as opposed to the expected hexagonal form that leads to the sixfold symmetric snowflake shape. By combining their talents and experience — Norm is the theorist and Konrad is the experimentalist — these two physicists made a discovery they hadn’t dreamed was possible.
Science always builds on past discoveries. Norm’s and Konrad’s research was inspired by Peter Feibelman’s (1130) research in water-solid interactions. In 2002, Peter achieved a major breakthrough in interpreting water-solid interactions (Jan. 25, 2002, Sandia Lab News). His research explained why an initial layer of water molecules lies flat on the precious metal ruthenium.
Peter says that Norm and Konrad’s work shows Sandia research at its best.
“A gifted experimentalist determined to solve a problem as hard — and important — as how ice grows on a metal crystal contrives to make a scanning tunneling microscope take pictures that, because ice is a good insulator, no one imagined could be taken. Then working closely with a theorist known for a deep understanding of the forces that drive film growth, he concludes that the structure of the imaged ice film is not what thermodynamics would favor, but instead reflects the growth process. So, pictures ‘impossible’ to take yielded a lesson no one had imagined,” Peter says.
Norm and Konrad published their work in a paper, “Growth of multilayer ice films and the formation of cubic ice imaged with STM,” that was published in Physical Review B in May. The paper has generated excitement in the physics world and accolades for the researchers.
“How water interacts with solids is extremely important,” says Norm. He points to the design of fuel cells and water purification systems as two areas that could benefit from new STM information. “Getting direct information is difficult, so imaging how small ice crystals grow on solid surfaces is an important advance. This is solid information that allows basic theories to be verified. This was our goal — to provide unambiguous information.”
Ice cubes or snowflakes?
The ice-growth images answer a fundamental mystery about ice: Snowflakes form in the classic six-sided symmetrical shape, but at low temperatures, ice grows in a cubic form. This phenomenon is something that has puzzled scientists for 60 years.
Norm and Konrad discovered that when an ice film is extremely thin, measuring an average of about 1 nanometer thick, the water molecules form little islands of crystalline ice. Once the thickness reaches 4 or 5 nanometers, the ice islands join together and start to form a continuous film.
In the Physical Review B paper, the researchers showed that cubic ice forms when the ice crystals merge. Because of a mismatch in the atomic step heights of the platinum substrate relative to ice, the coalescence often creates screw dislocations in the ice. Further growth occurs by water molecules attaching to the steps that spiral around screw dislocations, creating cubic ice in the process.
“In retrospect, this process might seem obvious, but it was not anticipated. The ability of microscopy to directly observe ice growth allowed us to solve this very old problem,” says Norm.
Pushing the boundaries of STM
The STM is a notoriously finicky piece of scientific equipment, and working with ice only increased the difficulty. An STM functions by positioning a narrow needle tip near the sample and then allowing a tiny electrical current to flow across the gap. As the tip of the STM is scanned across the sample surface, the voltage required to position the scanner is used to form an image of the sample.
“Typically, an STM only works if the substrate is conductive,” says Norm. “Through persistence and patience, Konrad learned that to image ice, one needs a very small current — three orders of magnitude smaller than what had previously been tried.”
It was Konrad’s intuitive decision to change the STM’s parameters, namely those for voltage and current, that made imaging ice crystals feasible. Basically, Konrad found the sweet spot where none was believed to have existed.
The STM was developed in 1981 and earned its inventors, Gerd Binnig and Heinrich Rohr, a Nobel Prize for physics in 1986. “The discovery caused a rebirth of surface science and completely changed the field, but until now, people had not been able to apply it to ice,” says Norm. “The fact that we can apply these same methods to ice is very exciting.”
STM requires artistry and intuition on the part of the scientist. Konrad jokes that he has been working with STM for his entire life (actually, it’s been 15 years).
“STM is an experiment that doesn’t always work. Because you are trying to get atomic resolution, a few atoms on the apex of the tip can completely throw off the experiment,” says Norm. “If you are not getting an image, you don’t know if your tip is bad or you are choosing the wrong parameters.”
This means that experiments have to be repeated over and over. “It’s like fishing in the dark,” says Konrad.
In fact, the two physicists never expected that they could image thick ice films; they were hoping for a few molecules. Konrad explains that even after he began imaging thicker ice films, he didn’t trust the results. Instead, he thought they were just very misleading artifacts of the measurement approach, which have fooled other scientists to the degree of publishing results based on such artifacts.
Because Konrad only expected to see films a few molecules thick, he had the STM tip set too close; it was shaving off the top of the films. “For about a month, we thought the films were not really as high as they seemed. We thought the insulating quality of ice made them appear to be higher,” he explains. “I increased the voltage, and the ice appeared to really pop out. Still, I thought it was just a really striking electronic effect.”
However, the researchers could not come up with another explanation for why the films appeared so high. Konrad then purposely grew very thick films and reversed the polarity on the STM, which resulted in an ice carving that proved the thickness was, in fact, real.
Norm and Konrad credit Peter’s work plus 10 years of basic energy science research at Sandia for laying the foundation for their breakthrough. “Sandia has been studying metal epitaxy for the last 10 years; we’ve gained a thorough understanding of the physics of the very early stages of crystal growth,” says Norm. “We’ve also developed a strong modeling capability, so we could immediately begin working with the images.”
The two Sandians are not resting on their initial success; in fact, they say they are working to build on their breakthrough. Future experiments include putting salts on an ice crystal to see how salts change the crystal’s growth and depositing molecules that react with water, such as atomic oxygen, to determine the exact point on the surface where water dissociates.
“Our ability to image these ice films opens the door to a multitude of exciting new experiments,” says Konrad.