Gold membrane facilitates surface analysis

Using a special wafer-thin gold membrane, researchers have made it much easier to examine surfaces using Raman spectroscopy. They can now measure surface properties more easily and more precisely.

An interdisciplinary team of materials scientists and electrical engineers led by Lukas Novotny, Professor of Photonics at ETH Zurich, together with colleagues from Humboldt-Universität zu Berlin, have developed a method that will make it much easier to characterize surfaces in the future. They recently published the results of their work, which is based on the use of a wafer-thin gold membrane, in the scientific journal Nature Communications.

"The surface was made by the devil" - this sentence is attributed to the theoretical physicist Wolfang Pauli, who taught at ETH Zurich for many years and was awarded the Nobel Prize in 1945 for his work on quantum mechanics. Researchers do indeed have their difficulties with surfaces. On the one hand, they are very important in both animate and inanimate nature, but on the other hand, it is sometimes fiendishly difficult to study them using conventional detection methods.

"Whether it's catalysts, solar cells or batteries - surfaces are always extremely relevant for their functionality," says Roman Wyss, former PhD student in materials science and first author of the study, who is now conducting research at ETH start-up Enantios. This relevance stems from the fact that the important processes usually take place at interfaces. In the case of catalysts, it is about the chemical reactions that are accelerated on their surfaces. For batteries, on the other hand, the surface properties of the electrodes are decisive for their efficiency and long-term behavior.

For the non-destructive investigation of material properties - i.e. without damaging the material - researchers have been using Raman spectroscopy for many years. This involves directing a laser beam at the material and analyzing the reflected light. The properties of the reflected light, whose frequency spectrum has been altered by the vibrations of the molecules in the material, can be used to draw conclusions about the chemical composition of the object under investigation - known as a chemical fingerprint - as well as to detect mechanical effects such as stresses.

"This is a very powerful method, but it can only be applied to surfaces to a limited extent," says Sebastian Heeg, who was involved in the experiments as a postdoctoral researcher with Lukas Novotny and now heads a junior research group at Humboldt-Universität. As the laser light penetrates a few micrometers deep into the material during Raman spectroscopy, the frequency spectrum is mainly influenced by the interior of the material and only to a very small extent by the surface, which is only a few atomic layers thick.

In order to make Raman spectroscopy usable for surfaces as well, the ETH researchers developed a special gold membrane that is only 20 nanometers thick and contains elongated pores about a hundred nanometers in size. If such a membrane is applied to a surface to be examined, two things happen: firstly, the membrane prevents the laser beam from penetrating into the interior of the material. On the other hand, the laser light is concentrated where the pores are located in the gold membrane and is emitted only a few nanometers deep into the surface.

"The pores act as so-called plasmonic antennas - very similar to the antenna in a cell phone," says Heeg. The antenna effect amplifies the Raman signal of the material surface up to a thousand times compared to the signal of conventional Raman spectroscopy without a membrane. Heeg and his colleagues were able to demonstrate this impressively using the example of strained silicon and the perovskite crystal lanthanum nickel oxide (LaNiO3).

Strained silicon is important for applications in quantum technologies, but until now the strain could not be investigated using Raman spectroscopy because the signal generated by the surface was drowned out by the background noise of the measurement. After the gold membrane was applied, the strain signal was selectively enhanced to such an extent that it could be clearly distinguished from the other Raman signals of the material.

The metallic perovskite lanthanum nickel oxide is in turn an important material for the production of electrodes. "The strong coupling between its crystal structure and electrical conductivity makes it possible to control the conductivity by changing the electrode thickness in the nanometer range. It is assumed that the surface structure plays an essential role in this," says Mads Weber, former postdoc at ETH Zurich and now assistant professor at the University of Le Mans, who researches this class of materials and was also involved in the study. Thanks to the new gold membrane method, the researchers have now been able to gain an insight into the surface structure of lanthanum nickel oxide for the first time.

"Our approach is also interesting in terms of sustainability, as it gives existing Raman devices completely new capabilities without a great deal of effort," says Heeg. In future, the researchers want to further improve their method and adapt it to the needs of users. For example, the pores in the gold membrane currently vary in size and are arranged irregularly. By producing membranes with parallel pores of the same size, the method could be optimized for certain materials and the Raman signal strength could be increased a hundredfold. (OM-8/24)