Mapping chemicals with nanoscale resolution

Nanostructured compound materials are attracting increasing interest from scientists in fields as different as physics, materials science, chemistry, and biology. Many not-yet-understood phenomena could potentially have a remarkable impact on both our fundamental understanding of matter at the nanoscale and potential applications. Areas of research that would gain great momentum from nanostructuring compound materials include solid-state lasers, quantum computers, and biocompatible technologies.^1,2

The investigating of nanoscale phenomena still poses major challenges. However, we believe that if we can map chemical composition with nanoscale resolution, we will make progress, since the behavior of nano-composite materials depends largely on nanoscale chemical inhomogeneities.

A wealth of experimental techniques have been adapted to sense chemical information on the nanoscale. In some cases the information is obtained indirectly through strain measurements, made both by direct space imaging and reciprocal space mapping.^3,4 Promising results have been achieved by combining microscopic measurements with chemical-etching procedures.⁵ Nonetheless, we regard spectroscopic probing at the nanoscale as the most attractive approach because it allows us to gain direct elemental information with large flexibility. Tools for nanoscale spectroscopy include the transmission- or scanning-electron microscope combined with electron-energy-loss spectroscopy or energy dispersive x-ray analysis (TEM or SEM with EELS or EDX), the scanning Auger microscope (SAM), and the x-ray photoemission electron microscope (XPEEM). The latter is one of the best experimental techniques for accessing chemical gradients at the nanoscale because it is minimally intrusive and convenient to operate,⁶ except that it is usually used only for qualitative measurements.⁷ We recently developed an analytical framework for XPEEM images and spectra that allows us to quantify chemical gradients in nanostructured binary alloys.⁸

XPEEM basically operates the same way as x-ray photoelectron spectroscopy (XPS). The essential difference is in the collection of the photoelectron (PE) signal: the energy-filtered PE yield is integrated by XPS and laterally resolved by XPEEM. The spatial resolution of XPEEM images stems from the technique's ability to localize the origin of the photoelectron emissions from a surface element to within a few tens of nanometers. Unfortunately, the local density of x-rays exciting each surface element is hard to determine. That is why—unlike XPS—quantitative exploitation of XPEEM is usually cumbersome.

Recently we proposed an analytical procedure for measuring the local stoichiometry in nanostructured binary alloys from XPEEM images and spectra. Our approach relies on experimentally accessible variables (the integrated PE intensities). Quantitative information is retrieved by normalizing the PE peaks from each surface element (exposed to an undefined photon flux) by a factor defined as the expected PE yield from a surface element (under the same undefined illumination conditions) at a reference—not necessarily known—composition.⁸ Chemical maps are obtained by processing normalized XPEEM images related to specific core levels for both elements in the alloy.

We demonstrated the power of our approach by measuring the stoichiometry within individual self-organized semiconductor nanostructures.⁹ Experiments were performed at the Nanospectroscopy beamline located at the Elettra synchrotron light source.¹⁰ Germanium was deposited on Si(111) substrates by molecular beam epitaxy. Following the Stranski-Krastanov growth mode, 3D islands nucleated on a critically-thick wetting layer.¹¹ The compositional inhomogeneity inside the single nanostructures stemmed from the nature of the intermixing diffusion phenomena. Chemical maps were obtained by probing the Ge3d and Si2p core levels excited by synchrotron x-rays. The relevant XPEEM micrographs were processed by using the intensity from the wetting layer as a reference: the normalization images were calculated by simple extrapolation of the PE yield from the surroundings of the islands.⁸ We were able to correlate the average chemical properties of individual nanostructures to their geometrical size and growth temperature.¹² Moreover, we very recently reported obtaining stoichiometry maps from within single islands.⁹ These display Si-rich edges as opposed to Ge-rich centers, a valuable piece of information for understanding the interdiffusion dynamics in our system of interest.^5,13,14

Gaining insight into the chemical features of nanostructured compound materials at the nanoscale is not simple, but XPEEM represents an experimental technique that combines direct chemical sensitivity, operational convenience and minimal intrusiveness: fundamental requirements for a reliable elemental mapping. We recently demonstrated that quantitative XPEEM measurements are possible. So far, we have worked on the simple case of a binary alloy. In future, we will aim to generalize our analytical procedure to allow scrutiny of more complex systems. The subsequent evolution of our method will touch on the issue of mapping the strain gradients in nano-heterostructures using XPEEM. This represents another major hurdle for the semiconductor industry.